Television Subcarrier Phase Correction for Color Field Sequencing

VIDEO FRAME STORAGE RECORDING AND REPRODUCING APPARATUS, Ser. No. 763,371, filed Jan. 28, 1977, by Joachim P. Diermann and Thomas W. Ritchey, Jr.

PLAYBACK APPARATUS ASSIGNMENT MEANS, Ser. No. 763,462, filed Jan. 28, 1977, by Howard W. Knight and Edwin W. Engberg, now abandoned.

TELEVISION SIGNAL DISC DRIVE RECORDER, Ser. No. 763,795, filed Jan. 28, 1977, by Howard W. Knight and Edwin W. Engberg.

DISC DRIVE RECORDING PROTECTION APPARATUS, Ser. No. 763,761, filed Jan. 28, 1977, by Edwin W. Engberg.

METHOD AND APPARATUS FOR PROVIDING DC RESTORATION Ser. No. 763,461, filed Jan. 28, 1977, by Luigi C. Gallo.

METHOD AND APPARATUS FOR INSERTING SYNCHRONIZING WORDS IN DIGITIZED TELEVISION SIGNAL DATA STREAM, Ser. No. 763,463, filed Jan. 28, 1977, by Luigi C. Gallo.

PRECISION PHASE CONTROLLED CLOCK FOR SAMPLING TELEVISION SIGNALS, Ser. No. 763,453, filed Jan. 28, 1977, by Daniel A. Beaulier, Luigi C. Gallo.

DIGITAL TELEVISION SIGNAL PROCESSING SYSTEM, Ser. No. 763,941, filed Jan. 28, 1977, by Luigi C. Gallo.

CLOCK SIGNAL GENERATOR PROVIDING NONSYMMETRICAL ALTERNATING PHASE INTERVALS, Ser. No. 763,792, filed Jan. 28, 1977, by Daniel A. Beaulier and Luigi C. Gallo.

PHASE LOCK LOOP FOR DATA DECODER CLOCK GENERATOR, Ser. No. 763,793, filed Jan. 28, 1977, by Kenneth Louth and Luigi C. Gallo.

A CIRCUIT FOR DIGITALLY ENCODING AN ANALOG TELEVISION SIGNAL, Ser. No. 762,901, filed Jan. 26, 1977, by Daniel A. Beaulier.

HIGH BIT RATE DATA ENCODER FOR DATA TRANSMISSION SYSTEM, Ser. No. 763,763, filed Jan. 28, 1977, by Luigi C. Gallo.

DATA RATE AND TIME BASE CORRECTOR, Ser. No. 763,794, filed Jan. 28, 1977, by Luigi C. Gallo, now abandoned.

A DIGITAL CHROMINANCE SEPARATING AND PROCESSING SYSTEM AND METHOD, Ser. No. 763,251, filed Jan. 26, 1977, by Robert P. MacKenzie, abandoned in favor of continuation application, Ser. No. 765,563, filed Feb. 4, 1977.

FREQUENCY RESPONSE EQUALIZER, Ser. No. 762,902, filed Jan. 26, 1977, by Jerry W. Miller and Luigi C. Gallo.

A CIRCUIT FOR GENERATING A DIGITAL DELETED DATA, BLINKING CROSS SIGNAL WHICH IS STORED IN A DELETED TRACK AND SELECTIVELY DISPLAYED FOR DETECTION, Ser. No. 762,903, filed Jan. 26, 1977, by Luigi C. Gallo and Junaid Sheikh abandoned in favor of continuation application, Ser. No. 765,564, filed Feb. 4, 1977.

The present invention generally relates to recording and reproducing apparatus and, more particularly, to apparatus that is adapted to record and reproduce television signals, using digital techniques.

The continued advances in technology have resulted in many changes in the equipment that is currently being used in television broadcast stations. One of the more recent changes that has evolved is the shift away from photographic techniques toward the use of magnetic media in many phases of the operation of the commercial broadcast television station. For example, feature films being broadcast often originate from magnetic tape rather than film and television station news departments are increasingly converting to videotape recording systems rather than using film cameras to provide the visual coverage of the news stories. Moreover, many systems utilize travelling transmitters that can either broadcast on location coverage or transmit such coverage to the station which can either be broadcast "live" or videotaped, edited and broadcast at a later time. Some of the many benefits of these techniques are the ease of handling, flexibility and speed of processing compared to the use of photographic film, coupled with the ability to reuse the magnetic tape when the information that is recorded on them is no longer needed.

One of the last remaining film domains in the present day commercial television broadcasting station is the Telecine island which uses 35 millimeter film transparencies. The Telecine island is used to provide video still images that are used during programming, commercials, news and the like, i.e., wherever a still image may be used during operation. Their use is extensive as is evidenced by the fact that the average commercial broadcast television station maintains a total file on the order of about 2000 to 5000 35 millimeter transparency slides. The maintenance of the total file represents a laborious operation which requires introduction of new slides, the discarding of obsolete slides and the maintenance of an accurate index so that they can be readily obtained when needed. When slide program sequences are to be assembled, they must be manually carried to the Telecine island, cleaned and manually loaded. Even with the cleaning operation, dust particles and scratches and the like may easily result in an unsatisfactory end product even when the projectionist is careful. Morever, following their use during broadcasting, the slides must be removed and returned to the file. The entire assembling, use and refiling of the slides represent a substantial labor investment because of the many manual operations that are required. The Telecine operation is considered to be one of the most antiquated operations in many modern broadcast stations and is basically incompatible with a fully automated station operation.

In contrast to the Telecine island or the use of opaque graphic material as the source for generating video still images, the present invention described herein facilitates the use of a recording and playback apparatus that will record and reproduce still images, with the still image video information being stored on magnetic media. The magnetic recording and playback apparatus utilizes generally standard computer disc drives (though modified in some respects as will be described) as the magnetic storage media and thereby eliminates the many problems that are associated with slide transparencies. Since the still images are recorded on magnetic media, the problems of physical degradation during use, e.g., dust particles and scratches, are not experienced. Moreover, since the recorded information can be easily accessed, the same still image may be used by operators at different locations almost simultaneously.

As will be described in detail, the apparatus disclosed herein effectively records and reproduces still frame color video images on disc drive disc packs, and produces a full four field NTSC color code sequence, i.e., color frame, utilizing only a picture frame or a two field television signal sequence of recorded information. During the recording process, the analog color video information signal is sampled at a sampling rate of three times the frequency of the unmodulated subcarrier and the samples are converted into a plurality of digital data streams which are further processed and recorded. Only two fields of the four field NTSC color code sequence of a still color video image are recorded. The horizontal synchronization pulses of the analog video information signal are removed and a redefined digital synchronization word is inserted in the horizontal interval on alternate lines, with the synchronization words being inserted synchronously with the color subcarrier frequency of the color video information signal.

During reproducing, the recorded two field picture frame of digitalized color video information is reproduced at least two times to enable the generation of a full four field NTSC color frame. If the still video image contained in the recorded two fields is, for example, to be displayed on a video monitor, the two fields are repetitively reproduced during the interval that the still image is displayed. During the second, and thereafter every other reproduction of the recorded two fields, the synchronization word is effectively misplaced by 1/2 cycle of the three times subcarrier signal. This occurs by virtue of the reversal in phase of the color subcarrier (and necessarily the three times subcarrier signal component of the digitalized video) of the first field during its second reproduction relative to the proper color subcarrier phase for a four field NTSC color code sequence and the use of a phase continuous clock signal to time the processing of the repetitively reproduced two field sequence. A visual jittering, or horizontal motion, of the image being reproduced results, because of the mispositioning of the horizontal synchronization word on alternate frames relative to subcarrier. The present invention identifies the second reproduction of the two fields and adjusts for the misposition so that the jittering is eliminated.

Accordingly, it is an object of the present invention to provide an improved apparatus that enables a full four field NTSC color code sequence of color video information in a manner whereby a nonjittering video image can be displayed when reproducing less than a full four field color code sequence of recorded information.

Other objects and advantages will become apparent upon reading the following detailed description in conjunction with the drawings.

FIG. 1 is a perspective view of the apparatus embodying the present invention, illustrating its overall appearance, including the internal access station and two disc drive units;

FIG. 2 is an enlarged perspective view illustrating a representative remote access station that an operator can use to control the operation of the apparatus of the present invention;

FIG. 3 is an enlarged top view of a portion of the internal access station keyboard shown in FIG. 1 particularly illustrating the various keys and bars that an operator uses during operation;

FIG. 4 is a broad functional and simplified block diagram of the entire apparatus of the present invention;

FIG. 5A illustrates a portion of a typical television signal illustrating the vertical interval thereof;

FIG. 5B illustrates a portion of a color television signal, particularly illustrating the horizontal synchronization pulse and color burst signal;

FIG. 6 is a functional block diagram broadly illustrating the signal flow path through the apparatus during a record operation;

FIG. 7 is a functional block diagram broadly illustrating the signal flow path through the apparatus during a playback operation;

FIGS. 8A and 8B together comprise a block diagram illustrating the signal system for the apparatus of the present invention, including control interconnections between the various blocks;

FIG. 8C is a timing diagram illustrating sampling of a television signal and phase relationships that occur at different locations of the signal system;

FIG. 9 is a functional block diagram of the video input circuitry (substantially similar to the reference input circuitry) which is a portion of the signal system shown in FIG. 8A;

FIG. 10A is a functional block diagram of the reference logic circuitry which is a portion of the signal system shown in FIG. 8A;

FIG. 10B is a timing diagram for the PALE Flag generator included in the reference logic circuitry shown in FIG. 10A;

FIG. 11A is a functional block diagram of the reference clock generator circuitry which is a portion of the signal system shown in FIG. 8A;

FIGS. 11B and 11 are timing diagrams illustrating the operation of portions of the reference clock generator shown in FIG. 11A;

FIG. 12 is a functional block diagram of the encoder and sync word insertion circuitry which is a portion of the signal system shown in FIG. 8A;

FIG. 13A is a functional block diagram of the data rate and time base corrector circuitry which is a portion of the signal system shown in FIG. 8A;

FIGS. 13B and 13C are timing diagrams for the data rate and time base corrector circuitry shown in FIG. 13A;

FIG. 14 is a functional block diagram of the data transfer circuitry which is a portion of the signal system shown in FIG. 8A;

FIG. 15 is a block diagram of one embodiment of the chroma separator and processing circuitry of the signal system shown in FIG. 8A wherein the chrominance inverter portion is a digital transversal filter with odd symmetry;

FIG. 16 is a more detailed block diagram of the chroma inverter portion of the circuitry shown in the block diagram of FIG. 15;

FIGS. 17 and 18 are block diagrams of alternative embodiments of the chroma separator and processing circuitry of the signal system shown in FIG. 8A;

FIG. 19 is a block diagram of an alternative embodiment of the circuitry employed to reconstitute four fields of color television signals from a single stored field;

FIG. 20 is a functional block diagram of the blanking insertion and bit muting circuitry which is a portion of the signal system shown in FIG. 8A;

FIG. 21 is a functional block diagram of the digital-to-analog converting and burst and sync insertion circuitry which is a portion of the signal system shown in FIG. 8A;

FIG. 22 is a block diagram of a playback circuit which includes the equalization circuit of the signal system;

FIG. 23 is a block diagram of one embodiment of the equalization circuit shown in FIG. 22;

FIG. 24 is a block diagram of another embodiment of the equalization circuit shown in FIG. 22;

FIG. 25 is a graph showing the playback response of a conventional reproduce head and preamplifier combination circuit;

FIG. 26 is a graph showing an equalization curve provided by the equalization circuit shown in FIG. 22 which curve compensates for the curve shown in FIG. 25;

FIG. 27 is a functional block diagram of the disc drive servo phase lock circuitry of the disc drive portion of the apparatus;

FIGS. 28A and 28B are electrical schematic circuit diagrams of the record play control circuitry for the disc drive portion of the apparatus;

FIGS. 29A and 29B are electrical schematic circuit diagrams of the record timing circuitry for the disc drive portion of the apparatus;

FIG. 30 is an electrical schematic circuit diagram of the timing generator circuitry for the disc drive portion of the apparatus;

FIGS. 31A and 31B are electrical schematic circuit diagrams of the error check circuitry for the disc drive portion of the apparatus;

FIGS. 32A and 32B together comprise an electrical schematic diagram of the disc phase lock control circuitry for the disc drive portion of the apparatus shown in the block diagram of FIG. 27;

FIGS. 33A, 33B, 33C and 33D together comprise an electrical schematic diagram of the input circuitry of the signal system shown in the block diagram of FIG. 9;

FIGS. 34A, 34B, 34C and 34D together comprise an electrical schematic diagram of the reference logic circuitry of the signal system shown in the block diagram of FIGS. 10A and B;

FIGS. 35A, 35B, 35C and 35D together comprise electrical schematic diagrams of the reference clock generator of the signal system shown in the block diagram of FIG. 11A;

FIGS. 36A, 36B, 36C and 36D together comprise an electrical schematic diagram of the encoder and sync word inserter circuitry of the signal system shown in the block diagram of FIG. 12;

FIG. 36E is a timing diagram illustrating the operation of the data encoder circuitry shown in FIGS. 36A, 36B, 36C and 36D;

FIGS. 37A, 37B, 37C and 37D together comprise an electrical schematic diagram of the data decoder and the data rate and time base corrector circuitry of the signal system shown in the block diagram of FIG. 13A;

FIG. 37E is a timing diagram illustrating the operation of the data decoder circuitry shown in FIGS. 37A and 37B;

FIGS. 38A, 38B and 38C together comprise electrical schematic diagrams of the chroma separator of the chroma portion of the signal system shown in FIG. 15;

FIGS. 39A and 39B together comprise electrical schematic diagrams of the chroma inverter circuitry for use in the chroma portion embodiment illustrated by the block diagram of FIG. 16 and timing control therefor;

FIG. 39C is a functional block diagram of the timing control portion of the chroma inverter circuitry of the signal system schematically illustrated in FIGS. 38A, 38B, 38C and 38D;

FIG. 39D is a timing diagram illustrating the operation of the timing control portion of the chroma inverter shown in FIG. 39C;

FIGS. 39E and 39F together comprise electrical schematic diagrams of the chroma inverter circuitry for use in the chroma portion embodiment illustrated by the block diagram of FIG. 18 and timing control therefor;

FIGS. 40A and 40B together comprise electrical schematic diagrams of the chroma band pass filter circuitry of the chroma portion of the signal system shown in the block diagram of FIG. 15;

FIGS. 41A and 41B together comprise an electrical schematic diagram of the blanking and bit muting circuitry of the signal system shown in the block diagram of FIG. 20;

FIGS. 42A, 42B, 42C and 42D together comprise an electrical schematic diagram of the digital-to-analog converter and burst and sync insertion circuitry of the signal system shown in the block diagram of FIG. 21;

FIGS. 43A and 43B together comprise an electrical schematic diagram of the equalizer circuits of the signal system shown in the block diagram of FIG. 22;

FIGS. 44A and 44B together comprise an electrical schematic diagram of the preamplifier circuits employed in the playback circuit shown in the block diagram of FIG. 22;

FIGS. 45A and 45B together comprise an electrical schematic diagram of the disc predriver portion of the disc drive portion of the apparatus shown in the block diagram of FIG. 27; and

FIGS. 45A and 45B together comprise an electrical schematic diagram of the disc predriver portion of the disc drive portion of the apparatus shown in the block diagram of FIG. 27; and

FIGS. 46A and 46B together comprise an electrical schematic diagram of the data interface portion of the apparatus.

Referring to FIGS. 1-3, a recording and reproducing apparatus is illustrated, indicated generally at 70 in FIG. 1 which includes two bays 71 and 72 containing electrical circuitry associated with the apparatus, together with the various monitoring and control hardward shown specifically in the upper portion of the bay 72. The system also includes a pair of disc drives 73 located adjacent the rightward bay 72 with each of the disc drives 73 having a disc pack 75 mounted thereon. While two disc drive units are specifically illustrated in FIG. 1, it should be understood that there may be additional disc drives used with the system to increase the on-line storage capacity of the apparatus. It should also be appreciated that a single disc drive may be used. Operational control of the apparatus is performed by one or more operators using either one of many remote access stations, such as the remote access station 76 shown in FIG. 2, or an internal access station 78 which is located in the bay 72. If desired, a video monitor 79, vector and "A" oscilloscopes 80 may be provided as shown in bay 72. Phase control switches 81 are provided above the internal access station 78.

The apparatus is controlled by an operator using either the internal access station 78 or a remote access station 76, both types of which have a keyboard with numerical and function keys and bars, a 32 character display 82, which provides a readout of information that is needed to carry out functional operations during use, as well as to display the information concerning the identity of certain stills being addressed and other information. It should be understood that the remote access station 76 shown in FIG. 2 is representative of each of the remote access stations and that in the preferred embodiment, up to seven remote access stations can be used to control the apparatus 70. The internal access station keyboard indicated generally at 83 in FIG. 1, as shown in the enlarged fragmentary view in FIG. 3, has more expanded operational capability than the remote access stations, whose keyboards have fewer function keys. As will be explained in detail hereinafter, the keyboard contains a large cluster of keys indicated generally at 84 and a smaller cluster of function keys 85 located on the left side of the keyboard. Additionally, a turn key controlled switch 86 may be provided to switch between normal and delete operations to safeguard against the possibility of inadvertent or unauthorized erasure of actively used stills.

Referring to the very simplified block diagram shown in FIG. 4, the apparatus receives a video input signal which is processed by record signal processing circuitry 88 and is then applied to record signal interface circuitry 89 which directs the signal to all of the disc drives 73. Gating circuitry located within a selected disc drive 73 is enabled to allow the signal to be recorded on a selected drive. More than one disc drive 73 can be simultaneously selected for recording the video signal provided by the record signal interface circuitry 89. Switcher circuitry can be substituted for the signal interface and associated gate circuitry so that the signal provided by the record signal processing circuitry 88 is coupled only to selected disc drives having the disc packs 75 upon which the signal is to be recorded. During playback, a signal originating from one of the disc drives is applied to the playback switching circuitry 90 which directs it to one of the playback channels 91, each of which provides a video output channel. A computer control system 92 is interfaced with the record processing circuitry, signal interfacing and switching circuitry and disc drives for controlling the overall operation of the various components of the apparatus and also interfaces the remote access stations and internal access station. The circuit details of the computer control system 92 and of the access stations 76 and 78 for controlling the recording and reproducing apparatus 70 are described in the above-identified related application, Ser. No. 763,371. An operator can select a particular disc in which to store a still, provided that the disc pack is on-line, i.e., it is physically loaded on one of the disc drives 73. In this regard, it should be understood that the apparatus addresses disc packs rather than disc drives for the reason that the apparatus is adapted to identify up to 64 separate disc packs, only one of which can be located on a disc drive at any one time. Thus, in the event the apparatus has two disc drives, only two disc packs can be on-line at one time. The operator can use an access station keyboard 83 to enter the address of a disc pack upon which he wishes to record a still and, through the interaction of the computer with the disc drive on which the selected disc pack is loaded, can carry out the recording operation on the selected on-line disc pack. Similarly, an operator can play back a still frame from the disc pack on one of the disc drives and can define the playback channel that he wishes the still frame to be played through.

The apparatus has four major operating modes or conditions, i.e., (1) record/delete, (2) playback or reproduce, (3) sequence assembly and (4) sequence play. The record and play operations will be initially described, while referring to FIGS. 6 and 7 which illustrate somewhat simplified block diagrams of the signal flow paths during recording and playback, respectively, with respect to one of the disc drives 73.

Turning first to the record signal flow block diagram of FIG. 6, the composite video input signal is applied to the input stage circuitry 93 where clamping of the signal takes place and the synchronization and subcarrier components are stripped from the composite video signal. The input stage also regenerates the synchronization (hereafter often referred to merely as "sync") and subcarrier signals for later use during reproduction and, accordingly, the regenerated sync and subcarrier signals are directed to a clock generator 94 which also generates reference signals that are used by the downstream elements during operation. The clamped analog video signal with the color burst component is then applied to an analog-to-digital converter (A/D) 95 which provides an output signal at a sample rate of 10.7 megasamples per second, with each of the samples comprising 8 bits of information. The digital video signal is a nonreturn to zero code (NRZ) which means that it is a binary code defining a ONE as a high level and a ZERO as an equivalent low level. The digitized video signal appears on 8 parallel lines, i.e., one bit per line, which is applied to an encoder and sync word inserter 96 which converts the digitized video into a special recording code (referred to herein as a Miller code or a Miller squared code) that is particularly suitable for digital magnetic recording in that it minimizes DC content of a data stream. The circuitry also inserts a synchronizing word on alternate television lines with respect to a particular phase angle of the color subcarrier as represented by the color burst sync component. The sync word is used as a reference for correcting time base and skewing errors that occur during playback among the eight parallel bits of data that must be combined to define the value represented by each sample. The digital video information in the eight parallel lines is then applied to a recording amplifier circuitry 153 and head switch circuitry 97 associated with the selected disc drive 73 which switches between two groups of eight recording heads for recording the digitized video signal by the disc drive. The disc drive is servo controlled so that its spindle rotational speed is locked to vertical sync, with the rotational disc speed being 3600 revolutions per minute. By locking the spindle drive to vertical sync, the apparatus records one television field per revolution of the disc pack and simultaneously records the eight data streams on eight disc surfaces. At the completion of recording one field, the recording amplifier circuitry 153 and head switch circuitry 97 is commanded to activate another set of heads for simultaneoulsy recording the second field of a television frame on another set of eight disc surfaces so that a picture frame, i.e., two interlaced television fields is recorded on two revolutions of the disc drive, using 16 heads. Each disc pack located on a disc drive preferably contains 815 cylinders, each of which has 19 recording surfaces and can therefore store 815 digital television frames. There is one read/write head for each of the 19 disc recording surfaces of a disc pack and all heads are mounted vertically aligned on a common carriage whose position is controlled by a linear motor. It should be understood that a cylinder is defined to comprise all recording surfaces that are located on the same radius of a disc pack. However, the term track, rather than cylinder, is preferred herein and, accordingly, a track is meant to include all recording surfaces on a same radius, i.e., all surfaces on a cylinder. Thus, an addressed track for recording or playing back a still actually refers to the 19 individual surfaces on the cylinder available at that radius. Of the 19 surfaces that are available for recording, one is used to record the address and other housekeeping information, rather than active video information, and it is specifically referred to as the "data track". Two of the 19 surfaces are available for recording a parity bit and 16 surfaces are used to record the picture frame of video data as will be explained further hereinbelow. Also one of the heads, generally referred to as the servo head, travels on the 20th disc pack surface that contains only servo track information prerecorded by the pack manufacturer. The servo tracks carry out two functions, i.e., following a seek command the head stack traverses servo tracks that are counted to determine the instantaneous location of the heads and, after completion of a seek phase, the servo head generates an error signal that is used to control the linear motor position to hold the head carriage centered on the appropriate servo track. By using such a feedback system, it is possible to achieve a radial packing density of about 400 tracks per inch or a total of 815 tracks per disc pack.

Since the present apparatus does not record analog video signals because of frequency response limitations of disc pack memories, the video signal is digitized for recording. Because the digitized signal is recorded, the video signal to noise ratio of the system is primarily determined by quantization noise rather than recording media and preamplifier noise as is the case with conventional videotape recorders. Thus, the present apparatus delivers a signal to noise ratio of about 58 dB and effects such as moire and residual time base error do not exist, the digital random error of the storage channels being typically low enough to make occasional transmission errors virtually invisible.

By recording a digital data stream at a rate of 10.7 megabits per second on each of the eight disc surfaces, the linear packing density of the apparatus is about 6000 bits per inch which is about 60% greater than is used in conventional disc drive usage in data processing.

During playback and referring to FIG. 7, the heads read, i.e., reproduce the digital video information from the eight surfaces per field and obtain the recorded channel encoded digital video signal from the two fields forming each picture frame. The reproduced signal is applied to a playback amplifier circuitry 155 and head switch circuitry 97 associated with the selected disc drive 73 which amplifies the data streams of digital video information carried by the eight data bit lines and applies the same to equalizer and data detector circuits 99. The equalizer compensates for phase and amplitude distortion introduced to the signal by the band limiting effects of the record and reproduce processes and insures that the zero crossings of the reproduced signal are distinct and accurately positioned. Following equalization, the channel encoded signals in each data bit line are processed as described hereinbelow for transmission to the playback circuitry of the signal system over a twisted pair line. The processed channel encoded signals are in the form of a pulse for each zero crossing or signal state transition of the channel encoded signal. The twisted pair lines for the eight data bits of the digital video information apply the processed channel encoded signals to the decoder and time base corrector circuitry 100 of one or more of the playback channels 91 of the apparatus. The decoder and time base corrector circuitry 100 reprocesses the received signals to place them in the channel encoded format, decodes the signal to the non-return to zero digital form and time base corrects the digital signal with respect to station reference to remove inter-data bit line time displacement errors (commonly referred to as skew errors) and timing distortion within each of the data streams carried by the data bit lines. To facilitate processing of the reproduced signals, phase continuous clock signals are used to time the operation of the decoder and time base corrector 100 and following circuitry. As will be explained in more detail hereinbelow, this prevents the time base corrector portion of the circuitry 100 from correctly positioning the synchronization word in alternate reproductions of the picture frame. Thus, the time base corrector portion of the circuitry 100 serves to align the eight bits defining a single sample and remove timing distortion in each of the data bit lines relative to station reference. However, the aforementioned mispositioning of the synchronization word would lead to horizontal displacement of the picture frame upon alternate reproductions and resulting visible jitter in the displayed video image. It should be realized that each playback channel 91 is provided with decoder and time base corrector circuitry 100 and within each playback channel each of the eight data bit streams travels through a separate decoder and time base corrector. The output of the circuitry 100 is then applied to a comb filter and chroma inverter circuitry 101 which separates the chroma information and selectively inverts and recombines the signal for reconstruction of a four field NTSC sequence. This reconstructed digital signal is applied to circuitry 127, which in accordance with the present invention, adjusts for the mispositioning of the synchronizing word in alternate reproductions of the recorded two fields of the video information and applies the adjusted video signal to a digital-to-analog converter 102 which provides an analog video signal. The new sync and burst are then added by a process amplifier 103 to produce a composite video analog output signal of the playback channel 91 as is desired.

While the signal flow paths for both the recording and playback operations have been briefly and broadly described, the signal processing system for the composite television signal is much more detailed than is shown by the signal flow diagrams contained in FIGS. 6 and 7. The video signal system will now be described in greater detail in conjunction with the block diagram illustrated by FIGS. 8A and 8B which contains additional blocks than previously identified. However, the reference numbers previously identified will remain where corresponding functions are performed. The block diagram of FIGS. 8A and 8B also includes wider lines representing the video data flow through the signal system as well as other interconnecting lines that are necessary for controlling the timing and synchronization of the circuitry represented by the various blocks. The input and output lines from the various blocks in FIGS. 8A and 8B which have an asterisk adjacent to them are lines which extend to the computer control system 92.

It should also be understood that the apparatus of the present invention will be described herein with respect to use in an NTSC system which has a television field comprised of 525 lines, horizontal synchronizing pulses occurring at a rate of about 15,734 Hz (often referred to herein as "H sync") which means that the period between successive H pulses is approximately 63.5 microseconds. Moreover, the vertical blanking rate in the NTSC system occurs at a 60 Hz frequency and the chrominance information is modulated on a subcarrier signal having a frequency of about 3.58 megahertz (MHz). Because of the relationship of the color subcarrier phase with respect to horizontal sync, NTSC color signals have a four field sequence, which is commonly referred to as a color frame. The subcarrier frequency of 3.58 MHz will often be referred to herein simply as SC which means 1 times the subcarrier frequency and, similarly, other commonly used clocking frequencies in the described apparatus include 1/2SC, 3SC and 6SC. The 3 times subcarrier frequency (3SC) often occurs for the reason that during sampling of the analog composite video signal for digitizing the signal, a sampling rate of 3 times the subcarrier frequency, i.e., 10.7 MHz is used. The composite video signal of an NTSC system is illustrated in FIGS. 5A and 5B.

Referring again to FIG. 8A, but before discussing the functions of each of the blocks shown therein, some broad general considerations should be understood with respect to the overall operation of the illustrated signal system. Firstly, the video input signal that is fed to the video input circuitry 93A is an analog signal which is processed and applied to an analog-to-digital converter 95. The output of the converter contains the video information in digital format and the digitized data is further processed and recorded on a disc pack in a digital format. Similarly, it is played back from the disc pack, time base corrected and chroma separated and processed using digital techniques and is not converted to an analog signal until one of the final steps where the digital-to-analog converter and sync and burst insertion circuitry 102, 103 provides the analog composite video output as shown.

In the analog-to-digital converter 95, the analog composite video signal is sampled three times per nominal subcarrier cycle, or at a sampling rate of 3SC (10.7 MHz), and each sample is digitally quantized into an 8 bit digital word. A sampling clock having a frequency of three times or any odd multiple of the NTSC subcarrier frequency is necessarily an odd multiple of one-half of the horizontal line frequency. If such a sampling clock is phase continuous from line to line, its phase at the start of consecutive lines changes. Using such line to line phase continuous sampling clocks will result in the instantaneous amplitude of the analog signal being sampled during consecutive lines at different times relative to the start of the consecutive lines. Because of this, the quantized samples are not in vertical alignment from line to line. Vertical alignment of the samples from line to line is desired to facilitate the use of a digital comb filter to obtain a separated chrominance component of a television signal by combining quantized samples from three consecutive (all odd or all even fields) television lines of a television field, which may be designated T (for top), M (for middle), and B (for bottom) in proportion to the formulae

it should be appreciated that if the samples of the NTSC television signal are taken at an even multiple of the subcarrier frequency, the comb filtering technique would be ideal because of the phase of the sampling clock would not change from line to line. Hence, the digital code words or quantized samples would describe the instantaneous amplitudes of each line of the analog signal at the same times relative to the start of each line and all of the samples in the consecutive lines would be aligned vertically from top to middle to bottom.

The lack of vertical alignment of the samples of consecutive lines when using a 3SC, line to line phase continuous sampling clock can be more readily appreciated with reference to FIG. 8C(1) which shows a number of cycles of subcarrier in television line 1 that are sampled by the positive transition of a 3SC sample clock (FIG. 8C(3)) wherein the upward transition has an arrow depicting an "X" sample point that is also placed on the subcarrier for television line 1 at every sample point (FIG. 8C(1)). As shown, there ae three samples for each cycle of the subcarrier. However, during television line 2, i.e., the next consecutive line, the subcarrier has a reversed phase as shown in FIG. 8C(2) and similarly, the sampling clock 3SC is of opposite phase (FIG. 8C(4)) relative to its phase in line 1 (FIG. 8C(3)) so that during television line 2 the samples are taken where shown by the X's of the television line 2 subcarrier (FIG. 8C(2)) on the upward transitions and it is seen that the X samples from line 1 to line 2 are misplaced by 60.degree. with reference to SC, which detrimentally affects the response of the comb filter, which utilizes the instantaneous amplitude of the analog signal in the above mentioned equations for properly deriving the chrominance information. It should be appreciated that the samples taken on all odd lines will be vertically aligned and that the samples taken on all even lines will be vertically aligned but that the samples taken on even lines will be displaced 60.degree. with reference to SC relative to those samples on the odd lines.

To avoid the problem created by sampling at an odd multiple of subcarrier frequency, i.e., 3SC in the present apparatus described herein, vertical alignments of samples in all lines can be achieved by changing the phase of the sampling clock on alternate lines. In the examples shown in FIG. 8C, reference is made to FIG. 8C(5) which illustrates the 3SC sampling clock for television line 2 which has its phase reversed relative to what it would have been for television line 2, which is shown in FIG. 8C(4). By sampling on the upward transitions at the "0" sample points, samples marked by the "0" on the subcarrier for line 2 result as shown in FIG. 8C(2). Thus, the sample points in the subcarrier for television line 1 ("X's") are vertically aligned relative to the sample points ("0's") that are sampled using the alternated phase sample clock shown in FIG. 8C(5) rather than what would have normally occurred as shown by FIG. 8C(4). This technique is commonly referred to as phase alternate line encoding or PALE and the terms PALEd, PALEing and the like will commonly be referred to throughout the description of the apparatus described herein.

While the apparatus described herein utilizes comb filtering techniques together with a sampling rate of 3SC or 10.7 MHz and requires the use of a PALE sampling clock, it should be appreciated that a 4SC sampling frequency would eliminate the need for PALEing. The use of a 4SC sampling frequency is within the contemplation of the apparatus described herein in the event that the frequency response of the recording media, i.e., the disc packs on the disc drives is sufficient to permit operation at the 4SC, 14.3 MHz frequency. In this regard, it is to be appreciated that standard disc drives used in data processing applications typically operate in the range of about 61/2 megabits and the recording at a rate of 10.7 megabits represents a significant increase in the packing density of the disc packs themselves.

Another important aspect of the operation of the present apparatus that is a result of the use of PALEing will also be described with respect to FIG. 8C. By changing the phase of the sampling clock on every consecutive line, a phase discontinuity necessarily occurs with respect to SC. It is more convenient during the channel encoding of the signal for use in subsequent recording to channel encode the digitally quantized samples with respect to a continuous phase clock, i.e., no phase discontinuities from line to line. For this reason during recording, the PALEd data that results at the output of the analog-to-digital converter 95 is clocked out of the channel encoder 96 with a clock that has a continuous (i.e., no discontinuities) 3SC phase from line to line. However, clocking the encoder with a line to line continuous phase clock shifts the data in time on alternate lines by 1/2 cycle of 3SC, which disturbs the line to line sample time alignment created by sampling with a PALE clock. Since during playback the chroma processing circuitry requires the samples of data to be vertically aligned from line to line, which was the reason that a PALE sample clock was used in the analog-to-digital converter in the first place, it is necessary to retime or reclock the data from the continuous phase clock back to the PALE clock so that the sample time disturbance is removed and the chroma processing comb filter can process the data without error. Succinctly stated, the A/D converter 95 samples the analog signal using a PALE clock having line to line phase discontinuities. For recording, the channel encoder 96 encodes the PALE data with a line to line continuous phase clock, which requires, during playback and after decoding, the retiming of the NRZ information to a PALE clock for use by the chroma processing circuitry. However, the latter retiming from a continuous to a PALE clock is not performed during transfer modes of operation when the video data recorded on one disc drive memory is played back to be transferred and recorded on another disc drive memory. In such cases, the line to line continuous phase data clocking of the played back video data is retained and the data is rerecorded without disturbing the data clocking.

The above considerations will now be described in conjunction with FIG. 8C where the PALE data for lines 1 and 2 are shown in FIGS. 8C(6) and 8C(7), respectively. The bits A1 through E1 are consecutive bit cells that represent the instantaneous samples of the analog video signal that occur in line 1 corresponding to the X's shown in FIG. 8C(1), with each bit cell lasting a full clock cycle of the 3SC clock shown in FIG. 8C(3). Similarly, the line 2 bit cells A2 through E2 represent data that is derived by the sampling at the "O's" in FIG. 8C(2) using the PALE sample clock, which for television line 2 is shown in FIG. 8C(5). To clock the PALE data with a line to line continuous phase 3SC clock, arrows beneath the bit cells shown in FIGS. 8C(6) and 8C(7) depict the clocking points of the line to line continuous phase clock that produce the bit cells that are shifted and are in the relation shown in FIGS. 8C(8) and 8C(9). The start of each bit cell occurs at the clocking point and the level of the cell will be continuous through the bit cell interval so that the bit cells maintain their identity during the clocking.

To retime the data from the line to line continuous phase clock back to PALE clock so that the bit cells (samples) are vertically aligned as they should be, i.e., A2 is vertically aligned with A1, B2 with B1, etc., the retiming from the continuous phase clock to the PALE clock must be correctly done or misalignment of the bit cells will result. In this regard, the retiming or reclocking must be complementary, i.e., a bit cell that was clocked in the right portion thereof in a PALE-to-continuous reclocking must be left clocked in the continuous-to-PALE reclocking to insure proper playback. Thus, give the line to line continuous phase clocked data shown in FIGS. 8C(8) and 8C(9), the solid arrows illustrate the proper complementary clocking for the two television lines and produce the retiming of the data to the PALE clock having the A1 and A2 bits vertically aligned as shown in FIGS. 8C(10) and 8C(11). It should be noted that where bit cells that were right clocked going from PALE-to-continuous reclocking, are left clocked in the opposite conversion as is evident from viewing any of the bit cells, e.g., Al, with their associated clocking arrows in FIGS. 8C(6) and 8C(8). In the event that complementary clocking is not performed, then the bits will not be properly aligned as is shown by the dotted clocking arrows in FIGS. 8C(8) and 8C(9) which produce the relationship shown in FIGS. 8C(12) and 8C(13). The reclocking from either PALE to continuous or the converse is performed at various locations as will be evident from the ensuing description.

It should also be realized that the NTSC television signal does not have any specified, defined relationship between the horizontal sync pulse occurring at each line and the phase angle of the subcarrier signal with the exception that the phase of the subcarrier changes 180.degree. from line to line. In other words, the phase angle of the subcarrier signal relative to the H sync signal can vary from one video source to another and this variance makes the H sync an undesirable signal to control the operation of the apparatus. Accordingly, the apparatus herein uses the input signal's subcarrier as represented by the color burst sync component as the basic timing reference for the system and defines a new H sync related signal that is used for timing purposes instead of the signal's H sync. The new H sync related signal is chosen to be at a frequency of 1/2 of the nominal horizontal line frequency because it represents a whole number of cycles of the subcarrier frequency, i.e., two complete horizontal lines of subcarrier frequency or 455 cycles. Moreover, the H sync related signal is given a definite relation to the subcarrier, i.e., it is synchronized with respect to the phase angle of the subcarrier. In the record portion of the signal system a synchronizing word is inserted in the video signal on alternate television lines at a location corresponding approximately to that of the video signal's H sync and phase coherent with respect to a particular phase angle of SC generated from the video signal color burst subcarrier synchronizing component. The location of the new H sync related signal is defined at the beginning of each picture frame and is maintained for the duration of the picture frame to provide the video signal with an H sync related signal accurately and consistently defined with respect to the phase of the video signal's subcarrier. For the playback portion of the signal system, an H sync related signal designated H/2 is provided that is redefined to be coherent with respect to a particular phase angle of he reference input subcarrier, which phase angle is selectable through the playback system phase control.

The redefined H sync related signal, H/2, is used as a basic timing reference signal for the system during playback operations.

By using the redefined H sync related signal as the horizontal sync reference for the system, processing signals for recording, playback and other operations of the system is facilitated because a consistent time relationship is established between the video signal's subcarrier and redefined H sync related signal.

Additionally, the use of internal horizontal and subcarrier reference signals that can be varied in time relative to the television station reference sync, permits timing control that will enable the television signal to reach a remote location at the proper time after having experienced the usual propagation delays that occur.

Referring again to the block diagram of FIGS. 8A and 8B, the analog video input is applied to the input of input circuitry 93A where several operations occur in the processing of the analog video signal before it is applied to the analog-to-digital converter 95. More specifically, the input circuitry 93A amplifies the analog video signal, provides DC restoration, separates the sync components contained in the video signal for use in generating timing signals for the signal system, detects the level of the tip of the H sync and thereafter clips the same. Moreover, the H sync is separated using a precision sync circuit for use in producing a regenerated sync. The circuit also produces a regenerated SC signal that is derived from the burst of the video input or, in the absence of burst, from an H/2 reference signal that is generated and is derived from the video input H sync.

It should be understood that the video input circuitry 93A and the reference input circuitry 93B shown in the lower left of FIG. 8A perform similar functions, the video input circuitry primarily for the signal recording portion of the signal system and reference input circuitry primarily for the playback portion of the signal system. Therefore, for convenience of manufacturing and service, identical circuitry is used. However, the input circuits are connected in the apparatus to receive only the input signals required to perform their respective functions and while the same signals are produced by each circuit, they are not all utilized from each circuit. The reference input to the reference input circuitry is the station reference color black video signal which contains all components of a color television signal except that the active video portion of it is at a black level. Thus, the burst, H sync and the like are present at the reference input circuitry 93B as they are at the video input circuitry 93A. In addition, the reference input circuitry 93B uses an H phase position adjusting circuit that receives H position control signals from an operator controlled thumb wheel switch or the like, such as phase control switches 81, for adjusting H phase position of the regenerated H sync used in the playback portion of the signal system.

As shown, many of the output signals provided by the input circuits 93A and 93B are applied to the reference logic circuits 125A and 125B associated with the respective input circuits. The reference logic circuit 125A during the record mode of operation uses the inputs from the video input circuitry 93A, the analog-to-digital converter 95 and the computer control system 92 and through precision phase lock loop circuitry, generates a number of recording clocks at frequencies of 6 SC, 3SC, 1/8SC and a PALE flag signal. The PALE flag and 3SC signals are used by the reference logic circuit 125A to generate a 3SC PALE sampling clock signal whose phase is set for each line of the video signal by the PALE flag, which is at a frequency of H/2 . The PALE flag signal changes state at that rate although it does so asymmetrically, i.e., the two stages of the PALE flag signal are of unequal time intervals. It is made asymmetrical so that the sampling clock phase for the color burst portion of the video signal is constant with the phase of the subcarrier and only the portion of the television line thereafter has a sampling phase which is alternated on consecutive lines. This PALE clock is coupled to the analog-to-digital converter 95 and is the sampling clock signal for deriving the samples at 3SC or 10.7 MHz.

The reference logic circuit 125B uses inputs from the reference input circuitry 93B and the computer control system 92 and generates a clock reference signal at a frequency of SC and various other timing control signals. These signals are used in the operation of the apparatus in modes other than that of recording input video signals.

During the record and playback modes of operation, the reference logic circuits also generate servo sync signals for each of the disc drives for properly operating the disc drives at the proper phase.

During playback and other modes of operation other than that of recording input video signals, a reference clock generator 98 generates various clocks and additional timing control signals required by the various parts of the signal system used in such modes. The reference clock generator uses the inputs from reference input circuitry 93B, reference logic 125B, the playback portion of the signal system, an operator's control switch and generates clock signals at frequencies of 6SC, 3SC, SC and 1/2SC and various other timing control signals. The reference logic circuitry 125A and 125B and the reference clock generator circuitry 98 together comprise the signal system's clock generator 94 that provides the system timing control signals.

The clamped and H sync stripped analog video signal from the video input board is applied to the analog-to-digital converter 95 which converts the signal to an 8 bit binary coded signal in a PALEd NRZ (non-return to zero) format which is applied to the encoder switch 126. The analog-to-digital converter 95 is not shown in detail herein as it is identical in its design and operation to the one incorporated in the Ampex Corporation digital time base corrector No. TBC-800. More specifically, the schematic diagrams of the analog-to-digital converter 95 are shown in the catalog No. 7896382-02 issued October 1975. The specific circuitry for the analog-to-digital converter is shown in schematic drawing No. 1374256 appearing on page 3-31/32 of the catalog and in schematic drawing No. 1374259 appearing on page 3-37/38 of the catalog. These schematics are incorporated by reference herein.

The output from the analog-to-digital converter is then fed to an encoder switch 126 which comprises switching circuitry that ordinarily receives either the 8 bit digitized video data from the converter or from data transfer circuitry 129. The data transfer circuitry 129 enables the video information to be transferred from one disc drive to another disc drive. During the transfer mode of operation, the digitized information is read off of the disc drive, decoded to the NRZ digital format time base corrected and is then applied to the encoder switch which can select either source of digitized video information for the encoder 96. Because the channel encoded data recorded on the disc drives 73 has been clocked with a continuous phase clock, the NRZ data received by the data transfer circuitry 129 also is timed with respect to the continuous phase clock. Ordinarily, the data transfer circuitry 129 is provided with a PALE flag signal that is used to effect retiming of the NRZ digital data with respect to a PALE clock signal so that the data provided to the chroma separator and processing circuitry 101 is in the correct PALEd format. During the transfer mode of operation, this retiming is not necessary. The encoder switch 126 has circuitry for interrupting the coupling of the PALE flag signal to the data transfer circuitry 129 and thereby preventing the retiming of the NRZ data with respect to the PALE clock during the data transfer mode.

The encoder switch 126 is controlled by the computer control system 92 to gate the video data from either the input video or data transfer paths. It also switches between video and reference 6SC and 1/2SC timing signals since the reference timing signals are used during the data transfer mode and video timing signals during the record mode. The encoder switch is also adapted to generate a signal that will produce a blinking cross through the TV image which is a visual indication that the still location or address for a still is unoccupied and therefore available for recording and also to provide signals for performing diagnostic functions. With respect to the sync word inserter, the encoder switch 126 couples the 8 bit digital video data from the analog-to-digital converter 95 and the timing signals derived from the input video signal to the encoder 96.

The 8 bit data from the encoder switch 126 is then applied to the encoder 96 which initially generates a parity bit and then encodes the PALEd data into a Miller squared channel code format, which is a self-clocking, DC free, non-return to zero type of code. While PALEd data is applied to the encoder, the output of the encoder is a 9 bit data stream (if parity is included) that has a phase continuity with respect to 3SC. The continuous phase clocked data is easier to process, particularly, during the decoding operations. The DC free code avoids any possible DC component that could occur due to a preponderance of one logical state over a period of time which could have an effect of disturbing the data in the playback process. Reference is made to the U.S. Pat. by Jerry Wayne Miller No. 4,027,335, entitled "DC Free Eencoding For Data Transmission System"

As is comprehensively described therein, the coded format can be characterized as a DC free, self-clocking, nonreturn to zero format. It provides for transmitting binary data over an information channel of limited bandwidth and signal to noise, where the data is transmitted in selfclocking format that is DC free.

In limited bandwidth information channels which do not transmit DC, binary waveforms suffer distortions of zero crossing location which cannot be removed by means of linear response compensation networks. These distortions are commonly referred to as base line wander and act to reduce the effective signal to noise ratio and modify the zero crossings of the signals and thus degrade the bit reliability of the decoded signals. A common transmission format or channel data code that is utilized in recording and reproducing systems is disclosed in Miller U.S. Pat. No. 3,108,261. In the Miller code, logical 1's are represented by signal transitions at a particular location, i.e., at mid-cell, an logical 0's are represented by signal transitions at a particular earlier location, i.e., near the leading edge of the bit cell. The Miller format involves the suppression of any transition occurring at the beginning of 1 bit interval following an interval containing a transition at its center. Asymmetry of the waveform generated by these rules can introduce DC into the coded signal and the so-called Miller "squared" code used in the present apparatus effectively eliminates the DC content of the original Miller format and does so without requiring either large memory or the necessity of a rate change in the encoding and decoding.

The encoder circuitry 96 also generates a unique sync word in the form of a 7 digit binary number and inserts the sync word on alternate lines in a precise location determined by the 6SC and 1/2SC clock signals. In the record mode of operation, clock signals generated from the synchronizing components of the input video signal by the reference logic circuitry 125A are provided to the encoder circuitry 96 by the encoder switch 126 and result in the sync word being inserted at a location that approximately corresponds to where the video signal's horizontal sync pulse was previously located. In other modes of operation, the 6SC and 1/2SC clock signals are generated from the synchronizing components of the station reference color black video signal by the cooperative action of the reference logic circuitry 125B and reference clock generator 98. The encoder gates the H sync related sync word into the data stream on alternate television lines at the proper time relative to the regenerated subcarrier phase.

Data track information to be recorded on the data track of the disc drives 73 is also encoded by the encoder 96 prior to recording. The data track information is provided by the computer control system 92.

With reference to FIG. 8B, the ten data streams of encoded digital data appearing at the output of the encoder 96 is applied to an electronics data interface 89 which is merely signal splitting and buffering circuitry which couples the encoded data to the three disc drives 73 for selective recording on a disc pack 75. Each disc drive includes a disc drive interface 151 adapted to receive the encoded digital data from the electronics data interface 89 and send it to the record amplifier circuitry 153 and head switch circuitry 97 for recording on an associated disc pack 75 as well as to receive reproduced or detected data from the playback amplifier circuitry 155 and head switch circuitry 97 and send it to the data select switch 128. In addition, the disc drive data interface 151 receives the multiplex servo reference signal through the record signal splitter and sends it to the timing generator (FIG. 30) of the disc drive control circuitry. This signal is selected by the computer control system 92 from either reference logic circuitry 125A or 125B. The timing generator employs the multiplex servo reference signal to time the operation of the disc drive system so that record and playback operations and the rotational position of the disc pack 75 within the disc drive 73 are synchronized to the appropriate signal system timing reference.

The disc drive control circuitry returns prerecord timing and data timing signals through the disc drive data interface 151 to the electronics data interface 89 of the signal system. In the particular embodiment of the apparatus described herein, only two fields of the four field NTSC color television signal color code sequence are recorded, with each of the two fields recorded during separate revolutions of the disc pack 75. Immediately prior to the recording of the two fields of video data, the pre-record timing signal is generated and coupled to the electronics data interface 89. The interface sends the pre-record timing signal to the encoder 96 to cause the generation for an interval equivalent to two fields data equivalent to color black, which is digitally defined by logical 0's in the apparatus described herein. The two field interval of color black data is returned through the interfaces for recording on the disc pack at the track location selected for recording video data and its associated data track information. The recording of the two fields of color black data occurs during two revolutions of the disc pack 75 immediately preceding the two revolutions during which the two fields of video data are to be recorded. This conditions the track location for the subsequent over recording of the video and data track data. Because over recording previously recorded digital data with new digital data can be conducted to obliterate the previously recorded digital data and leave a recorded signal of sufficient quality to provide an acceptable signal of noise ratio upon playback, the pre-record cycle of operation could be eliminated from the apparatus and the recording of the two fields of video data and associated data track data accomplished in only two revolutions of the disc pack 75.

The data timing signal is returned to the electronics data interface 89 to time the generation and recording of the data track information during the second or last field of the two fields of video data. The signal is a pulse which begins after the vertical sync occurring between the two fields of video data and terminates at the end of the second field. It is during this interval that the data track information is recorded on the data track of the disc pack 75. The record signal splitter 89 couples the returned data timing signal to the computer control system 92 for identifying the data track recording interval to the system. In response, the computer control system 92 performs functions incident to the recording of data track information, including the provision to the signal system of the data track information associated with recording video data on a specified track of a specified disc pack. The encoder 96 receives the data track information and processes it as described herein for sending to the disc drive 73 and recording simultaneously with the last field of video data.

The record annd playback amplifier circuitry 153 and 155, the head switch circuitry 97, and the disc drive control circuitry of the apparatus described herein are arranged together so that the playback amplifier circuitry 155 and head switch circuitry 97 are activated to reproduce data from the associated disc pack 75 at all times except when a record operation is being performed. Hence, except during record operations, reproduced data is always being received by the disc drive interface 151, which in turn always provides the reproduced data to the data select switch 128. To record data, a record command provided by the disc drive control circuitry is coupled to the record playback amplifier circuitry 153 and 155 to activate the record amplifier circuitry 153 and disable the playback amplifier circuitry 155. The disc drive control circuitry also provides a 30 Hz head switch signal to the head switch circuitry 97 during record operations to cause the head switch circuitry to couple the data streams to one set of heads during the first field of two consecutive fields of data to be recorded and to the second set of heads during the second field. The 30 Hz head switch signal is continuously available and is similarly employed during playback operations to control the head switch circuitry 97 to switch the playback amplifier circuitry 155 between the two sets of heads for the reproduction of both fields of a desired video data signal.

Returning to FIG. 8A, during playback operations, the reference input circuitry 93B together with the reference logic 125B produces the regenerated subcarrier frequency for application to the reference clock generator 98 and the reference clock generator has outputs of 6SC, 1/2SC. H/2 and other timing signals providing the basic timing for playback operations. The clock and timing signals, including the reference H/2 signal, are synchronized to the reference color subcarrier to facilitate processing of the reproduced video signals. The reference H/2 signal is defined with respect to a particular phase of the reference color subcarrier in the first line of alternate fields of the reference color black video signal. The reference clock generator outputs are applied to the data decoder and time base corrector 100, data transfer circuitry129 and the chroma separator and processor 101 in addition to a blanking insertion and bit muting circuit 127 that inserts blanking, performs selective bit muting, and provides a selected picture frame video signal for output by the signal systems when the heads associated with a disc drive coupled to the playback channel are moved between track locations. Because of the use of the redefined reference H/2 signal in the data decoder and time base corrector 100, the synchronizing word contained in alternate reproductions of the two field video signal is mispositioned relative to the station reference H sync. This would introduce a jitter in the displayed video image if not corrected. The 8 bits of digital information are then applied to the digital-to-analog converter and sync and burst insertion circuitry 102 and 103. The aforementioned mispositioning of the synchronizing word is corrected in the blanking insertion and bit muting circuitry 127 preceding the digital-to-analog converter 102 by appropriately inserting a corrective delay in the signal path upon alternate reproductions of the two field video signal. The reference clock generator 98 identifies which reproduction of the two field video signal sequence requires the delay by examination of a color frame rate signal, H drive signal and field index signal, all provided by the reference logic circuitry 125B, and the reference color subcarrier signal. In response to the identification, the reference clock generator 98 generates a frame delay switch signal that is coupled to the blanking insertion and bit muting circuitry 127 for controlling the insertion of the corrective delay. Moreover, during the transfer and diagnostic modes of operation, the reference clock generator 98 supplies the basic timing clocks for the encoder 96 through the encoder switch 126 as shown.

During playback, the 10 bit parallel data stream comprising 8 bits of video data, the parity bit and data from the data track reproduced from a disc pack is amplified, equalized and detected by circuitry shown and described herein with reference to FIGS. 22 through 26, 43 and 44 and is then applied through the disc drive data interface circuitry 151 to a data select switch 128 which can switch any of the outputs of the three disc drives onto one or more of three channels. Thus, the data select switch can switch the information from disc drive No. 1 into channel A, or to two channels while simultaneously applying a data stream from another disc drive onto another channel. While information from two drives can not be simultaneously applied to a single channel, the converse is possible. The data select switch 128 comprises conventional switching circuits which are not set forth in detail herein.

Each of the detected nine bit streams of video data are parity data from the data select switch 128 is then applied to nine individual data decoders and time base correctors 100 which decode the data and then independently time base correct each of the nine data streams with respect to a common H/2 reference, which is defined with respect to the phase of the regenerated reference subcarrier, to remove any timing errors that may be present among the nine lines of data, i.e., it aligns all sync words so that each 9 bit parallel byte comprises the correct 9 bits of data. The other bit stream from the data track is coupled by the data select switch 128 to only the decoder portion of the decoder and time base corrector circuitry 100 and the decoded data track information is coupled to the CPU 106. The time base corrector does its correction using a continuous phase clock. However, the data is again retimed with respect to a PALE clock by the data transfer circuitry 129, i.e., the phase of the signal is alternated by reclocking it at every horizontal line, so that the 8 bit data stream that comes from the data transfer circuitry is a true PALEd signal. The data transfer circuitry 129 also performs a pariety check of the off disc data and performs error masking of individual byte errors when they occur by substituting what is likely to be the most similar previously appearing byte for the byte that was detected as being in error. In this regard, the byte that is substituted is the third previous byte, which is the most recent sample that was taken with the same phase relation to SC.

The output of the data transfer circuitry is applied to the chroma separator and processing circuitry 101 in the event that the video information is desired for viewing, as opposed to being recorded on another disc drive (transfer), in which case the data from the data transfer circuitry 129 is coupled to the encoder switch 126. The chroma separation and processing circuitry 101 works in the digital domain and separates the chroma information from the luminance using comb filter techniques and inverts the chroma information on alternate frames to form a four field composite NTSC signal that is then applied to the blanking insertion and bit muting circuitry 127 which inserts a reference black level during the blanking period, inserts grey level signals during the interval between the playback of consecutive stills, and performs bit muting operations if desired. The bit muting effectively mutes any bit or bits of an 8 bit television signal by shutting down that data bit stream and by so doing, achieves unusual visual effects in the resulting television signal such as producing exaggerated tones, ghostlike images and the like. The output from the blanking insertion and bit muting circuitry 127 is then applied to the following digital-to-analog converter 102. The digital-to-analog converter receives clock signals from the blanking insertion and bit muting circuitry 127 and converts the data to its analog form and also inserts the sync and burst components of the signal to produce a full composite analog television signal.

While the foregoing generally describes the overall operation of the signal system in a general manner, a more specific description of each of the blocks that are contained in FIGS. 8A and 8B as need for an understanding of the present invention will be given either with respect to the separate function block diagrams or the specific electrical schematic diagrams for the circuits themselves. Also where functional block diagrams are used to described the operation of the individual blocks of FIGS. 8A and 8B, the electrical schematic diagrams corresponding to those more detailed block diagrams are also included herein.

The video input and reference input circuitry 93A and 93B broadly described with respect to the block diagram of FIG. 8A contain substantially similar circuitry in both locations, although different inputs are received by each and all of the outputs that are available from each are not used. During record operations, the composite video input signal to be recorded is applied to the video input circuitry 93A which is used to obtain a regenerated subcarrier signal, and various vertical and horizontal sync rate ralated signals that are used by the apparatus in the performance of the record operations. The video input circuitry also provides an amplified and filtered video signal suitable for feeding the A/D converter 95. During playback operations, a reference color black video signal is applied to the reference input circuitry 93B which provides similar signals for use by the apparatus in the performance of the playback operations.

Referring more specifically to the block diagram of the video and reference input circuits shown in FIG. 9, the video signal is applied on line 200 into a video amplifier 201 which amplifies the signal and restores the DC component through a clamp 202. The clamp 202 samples the output of the amplifier on line 203 and produces a DC component on line 204 that extends to the amplifier 201. The DC restored video signal on line 203 is then passed through a low pass filter 205, the output of which appears on line 206 extending to a video gain control amplifier 207. The amplifier 207 is connected to another video amplifier 208 where a second clamp circuit 209 assures that the blanking level of the signal is at ground level by the application of a DC control signal via the line 210 to the video amplifier 208. The output of the video amplifier appears on line 211 and is coupled by one of the lines 218 extending therefrom to the sampling input of the clamp 209. Line 211 also extends to a gated sync clipping circuit 212 as well as to a precision sync separator 213. A tip of sync detector 214 detects the level of the tip of sync and provides a corresponding signal level on line 215 that extends to a comparator 216 as well as to the precision sync separator 213. In the video input circuitry 93A, a remote video gain control signal on line 217 is also applied to the comparator 216 for controlling the gain control amplifier 207 from a remote location. In the reference input circuitry 93B, the gain of amplifier 207 is not controlled from a remote location. The output of the tip of sync detector 214, which may contain alternating current ripple, is applied to one input of the precision H sync separator 213 while the other input to the separator is provided by one of lines 218 that extends from the output of the video amplifier 208. The two inputs to the precision sync separator 213 will both have AC ripple thereon if present in the signal and, accordingly, they are common moded so that the separator produces an AC ripple free precision separated sync on line 220 that is applied to miscellaneous sync circuits 221 and to an input of a horizontal sync phase detector 222. Another of the lines 218 from the output of the video amplifier 208 extends to a less precise sync separator 219 that produces a generally less precise separated sync signal which is applied to a gate pulse generator 223, outputs of which appear on lines 224 that extend to both clamps 202 and 209 as well as to the tip of sync detector 214. When the horizontal sync signal is detected and separated, a gate is produced by the pulse generator 223 which closes the clamps as well as the sync tip detector at the appropriate time during horizontal blanking.

The clamp 209 is closed during burst time for a whole, integral number of cycles, rather than arbitrary period, so that the blanking level of the video signal can be accurately obtained using integration techniques as will now be described in detail. The burst appears on line 225 which is applied to a burst limiter circuit 226 that is in turn connected to an amplifier 227 providing complementary outputs of the limited burst input. The output of the limiter circuit 226 is also connected to a burst presence detector circuit 228 having an output on line 229 that extends to a precision gate generator 230 as well as an output on line 260 that extends to a phase detector 231. When the presence of burst is detected, the precision gate generator 230 generates a precision burst gate signal that is coupled to enable the amplifier 227 and permit it to pass the middle three cycles of burst to apply them to the phase detector 231. The phase detector responsively provides an error signal to a voltage controlled oscillator 232 that reflects the difference in phase between the output of the oscillator and the phase of the burst cycles from the amplifier 227. The effect of the phase detector circuit controlling the oscillator 232 is to correct for longer term changes and not short term changes in the phase of the three cycles of burst that are used on every line as the subcarrier reference. The output of the oscillator 232 appears on line 233 after having been buffered by a buffer 234. The output of the oscillator is a continuous regenerated subcarrier signal SC (3.58 MHz) that is phase locked to the color burst when burst is present. However, in the event that the burst detector circuit 228 fails to detect burst, then the phase detector 231 compares the phase of an H/2 signal with the regenerated subcarrier output of the oscillator 232, the H/2 signal being produced by a sync generator 235 from an oscillator 236 that is controlled by the horizontal sync phase detector 222. This continuously regenerated subcarrier signal SC is coupled to the reference logic circuit 125A and, as will be described in detail hereinbelow, is employed in the apparatus described herein to generate the 3SC PALE clock used by the A/D converter 95 to effect digitization of the video signal.

A horizontal phase position control, indicated generally at 237, is provided for use in the reference input circuitry 93B to adjust the horizontal positioning of the regenerated sync. An 8 bit binary number is loaded into latches 238 by an operator controlled thumb wheel switch or the like for example, control switches 81 located by the internal access station 78 (FIG. 1), to preset a counter 239 which is clocked by a 400H clock derived from the oscillator 236. When the counter reaches its terminal count, it triggers a ramp generator 240 having an output 241 which extends to a second input of the H sync phase detector 222. Thus, by adjusting the latches, up to plus or minus 20 microseconds can be inserted in the feedback loop on line 241 and the phase of the regenerated sync signal can be adjusted for horizontal positioning of the picture during playback. Since a delay in the feedback loop means that the regenerated sync will be advanced, the horizontal position control can effectively advance the picture to compensate for propagation delays during transmission of a signal through cabling in a television station. As will be explained hereinafter in the detailed description of the reference clock generator circuitry 98, this horizontal phase position control is operated in conjunction with a subcarrier phase control operatively associated with the reference clock generator 98 whereby the amount of delay can be controlled in small increments, which in the embodiment of the apparatus described herein is about .+-. 0.8 nsec.

The output of the oscillator 236 also is used by the sync generator 235, which is of conventional design for television signal processing equipment to generate the various vertical and horizontal sync rate related signals indicated in FIG. 9. These sync rate related signals are generated with respect to the phase of the precisely regenerated H sync as provided by the phase detector 222 and, therefore, will always have a phase related to the input signal.

An important aspect of the circuitry shown in FIG. 9 is that the H sync of the video signal is clipped at precisely 1/2 its value and the level of the blanking is precisely clamped to ground. The regenerated subcarrier is phase locked with the burst and a precision horizontal sync signal is regenerated utilizing the precision sync separator. This signal is used by the sync generator 235 to provide a reset pulse (30 Hz field index pulse) for resetting a line identification or sync word inserter that will be hereinafter described. Since the clamp circuitry 209 examines for a zero average level of video at burst time using a clamping pulse which lasts precisely a whole number of cycles of burst, there is no need for low pass filtering the video and rejecting the burst before clamping is performed. This is due to the fact that resulting integration of the burst is equal to zero and there is no H/2 ripple introduced by integrating a signal that does not contain complete cycles of burst.

The block diagram shown in FIG. 9 describes the functional operation of the input circuitry and specific circuitry which can be used to carry out the operation thereof is shown in FIGS. 33A through 33D which together comprise a single circuit diagram for the video input processing circuitry.

With respect to the operation of the clamp 209 (see FIG. 33C), the voltage at the output of the amplifier 208 appears on lines 211 and 218, one of the latter of which extends downwardly to the base of an emitter follower transistor 244 that provides a voltage drop. Under equilibrium conditions, the blanking level of the video signal appearing on line 218 will be at ground potential. This signal is shifted by about 0.7V toward the negative as a result of the voltage drop through the emitter follower 244. A matching emitter follower transistor 245 with its emitter connected to the negative input of a differential amplifier 246 by line 247 shifts the comparison level (ground) toward the negative as does transistors 244. The emitter of the transistor 244 is connected to the positive input of the differential amplifier 246 when a transmission gate or switch 248 is closed during and for a whole number of cycles of burst by a signal on the line 224 that is produced by the redefined gate pulse generator 223 shown in FIG. 33D. Thus, during the burst time, switch 248 is closed charging a capacitor 249 to the average level of the burst. The switch is closed for an integral number of cycles of the subcarrier. This eliminates the need for low pass filtering the video to remove the burst before the clamping is performed, which is ordinarily done in the prior art in order to eliminate H/2 modulation of the clamping level. The charge on the capacitor 249 reflects exactly the average value of the burst and the differential amplifier 246 output represents an error that is applied to the video amplifier 208 through line 251, transistor 252 and line 210 which is connected to the emitter of the transistor 252. The blanking level of the signal on line 211 is thus held very close to ground due to the high DC gain of the differential amplifier 246. The operation of the clamp 202 is substantially similar to the operation of the clamp 209 and is shown in FIGS. 33A and 33B.

Referring again to FIG. 33C, the closing of the switch 248 gates burst through the switch into capacitor 249 and onto line 225 which extends leftwardly to FIG. 33A which is connected to the emitter of a transistor 254 and the burst therefore appears on the collector and on line 255 that extends to the burst limiter circuit 226. When burst is present, the burst presence detector circuit 228 provides a limited burst signal on its output line 229 that clocks the precision gate generator 230. A counter is employed as the precision gate generator and counts cycles of the limited burst signal and produces a precision burst gate during the middle three cycles of the nine to eleven cycle burst interval that is coupled by line 256 to enable the amplifier 227. Therefore, except for the middle three cycles of burst, the amplifier 227 is disabled by the output of the precision gate generator 230. When burst is present, the diode detector 257 and following latch circuit 258 of the detector circuit 228 provides a more negative level on line 260 extending to a switching transistor 259 (FIG. 33B) of the phase detector 231. When burst is present, switching transistor 259 is shut off and another switching transistor 261 of the detector 231 is turned on. When transistor 261 is on, the three cycles of burst from the amplifier 227 is applied by the driver 277 to a transformer 262 of the detector 231. The driver is in turn connected to the phase comparator 231a for comparing the phase of the burst with the phase of the output of the 3.58 MHz (SC) oscillator 232 that is present on line 233. When burst is not detected by the detector circuit 228, transistor 259 is switched on, which applies the signal H/2 to the other input of the driver 277 that is also connected to the transformer 262 and the phase of the oscillator output on line 233 is compared with the phase of the H/2 signal.

Turning now to the detailed circuitry for performing the precision H sync separation and referring to FIG. 33C, the sync is taken from the amplifier 208 on the line 218 extending to a low pass filter 264 whose output is coupled to the base of a transistor 265. The emitter of transistor 265 is connected to a transmission gate or switch 266 that is closed during the presence of sync by control line 224. The level of the sync is determined by charging a following capacitor 267 (FIG. 33D), which is buffered by a unity gain amplifier 268, and 1/2 of the DC level of the tip of sync together with the full level of AC ripple present in the signal is then applied via line 215 to one input of sync separator 213, the other of which is supplied by line 269 that comes from the emitter follower transistor 265. In the embodiment of the input circuitry 93A and 93B illustrated in FIGS. 33A-D, the precision H sync separator 213 is a comparator. In this manner, the output on line 220 is a separated sync whose timing is not affected by AC ripple on the video, because any AC ripple will appear on both inputs of the comparator 213 and will be prevented from appearing in the output of the comparator because of common mode rejection. The sync appearing on line 220 is a precision sync that is used by other parts of the signal system to generate horizontal line related synchronizing signals redefined in relation to a particular phase angle of the subcarrier signal which serve as timing references in the signal system for processing the video signals. Also, the horizontal line related synchronizing signal used in the system is at a rate of 1/2H sync because there are a whole number of subcarrier cycles for every two horizontal lines (227.5 .times.2 =455) and this consideration becomes important in the operation of apparatus described herein as will be evident from the ensuing description.

A less precise separated sync is also developed by taking the sync from the low pass filter 264 via line 270 to the imprecise sync separator 219, the output of which appears on line 271 that is applied to the gate pulse generator 223 which includes a one shot serving as a sync presence detector 276. The upper circuit, indicated generally at 272, generates a gate for use by the switch 266 to close the switch during the presence of sync, a circuit 273 produces a backporch sample and a circuit 274 redefines with respect to SC phase a burst gate signal. With respect to the generator 223, it should be appreciated that if no sync is present and therefore does not appear on line 271 from the imprecise sync detector 219, the sync presence detector 276 will through circuit 274 close the switch 248 in the clamp circuit 209 as well as a similar switch 275 in the clamp 202 so that all clamps operate on a DC feedback loop rather than permitting them to remain open. Thus, if sync is not present, the level on line 224 is placed high until sync returns and is detected. In addition, as a precautionary measure in the event the precision gate generator 230 does not receive the necessary number of burst cycles to clock it to its terminal state or count after its count cycle has been initiated, the detector 276 is coupled through circuit 274 to provide the burst gate signal to the precision gate generator 230 to assure termination of its count cycle and provision of the precision burst gate signal. This assures that the precision gate generator 230 will always properly respond to every input burst signal.

Because of the desirability of having a field index signal in the encoder switch 126 that is accurately related in phase to the input video signal's vertical sync, the output of the precision H sync separator 213 and an output of a vertical sync detector 278 (FIG. 33B) are provided to a NOR gate 279 (FIG. 33D) which provides the desired field index signal.

The reference logic circuitry 125A and 125B shown in the block diagram of FIG. 8A receive various signals from the input circuitry 93A or 93B relating to horizontal and vertical sync signals, regenerated subcarrier and the like and respectively generate a number of clock and timing control signals used in the operation of the apparatus. In addition, the computer control system 92 provides control signals to both logic circuitry 125A and 125B which cause the generation of servo sync signals which control the operating phase of the disc drives in accordance with the operation, viz, record, playback, transfer and the like, being performed by the apparatus. The reference logic circuitry is essentially duplicated so that one reference logic circuit is provided for use with the video input circuitry 93A and another for the reference input circuitry 93B, with the function of the reference logic circuitry being somewhat different during different operations of the apparatus such as recording, playback, transfer and the like. Because the logic circuitry 125A and 125B perform different functions, different inputs are received by each and all outputs that are available from each are not used.

The operation of the reference logic circuitry will now be explained in further detail with reference to a functional block diagram shown in FIG. 10A that has a dotted line extending horizontally in approximately the middle of the drawing. As is shown thereon, the upper portion of the circuitry is used only during a recording operation, whereas the lower portion is used during recording, playback and other operations performed by the signal system. The function of the upper portion of the circuitry is to generate various phase locked clock signals for recording operations using the regenerated subcarrier that was produced by the video input circuitry 93A from the color burst as has been previously described. The circuitry also generates a nonsymmetrical PALE flag signal at a rate of H/2 which is used within the circuitry to alternate the phase of the analog-to-digital converter sampling clock on consecutive horizontal lines for the reasons that have been hereinbefore described. The PALE flag is also available as an output from the reference logic circuitry 125B for use by other parts of the signal system, primarily those used in processing playback signals. The circuitry also generates a drive synchronization signal for operation of the servo control of the disc drive motors, providing a set of three pulses at a rate of 15 Hz which is multiplexed with H sync for use in controlling the disc drive servo. Other timing control signals are provided by the reference logic circuitry 125B as will be described in the following detailed description.

Referring to the upper portion of FIG. 10A, the subcarrier signal (SC) from either the video input circuitry 93A for the reference logic circuitry 125A or reference input circuitry 93B for the reference logic circuitry 125B is applied on line 300 and it is extended to a phase comparator 302, the output of which appears on line 303 to a summer 304 that has a second input on line 305 provided by an integrator 306. A precision digital burst phase decoder 307 receives the actual digitized video data taken from the output of the analog-to-digital converter 95 on line 308 and decodes whether the samples were taken at the proper phase of burst and produces a plus or minus error signal to the integrator 306 via line 309 for use in adjusting the phase of the sample clock so that the video signal is always correctly sampled. The output of the summer 304 appears on line 310 which is applied to a loop amplifier and filter 311 that is connected to a voltage controlled oscillator 312 by line 313 which also extends to one of two trouble lamp drivers 314. The output of the oscillator 312 appears on line 315 at a frequency of 6SC which is applied to a divide by 6 counter 316 as well as to a divide by 2 counter 317 which produces a PALE clock output at a frequency of 3SC on line 318. The divide by 6 counter has an output on line 319 at a frequency of SC which is applied to a divide by 2 counter 320 as well as to the other input of the phase comparator 302. The output of the divide by 2 counter 320 is a 1/2SC signal on line 321 which also extends to a pulse former 322 that is used to set and reset the divide by 2 counter 317 on alternate lines, the control being supplied through line 323 at an H/2 rate that is supplied by a PALE flag generator 324 as will be discussed hereinafter.

The operation of the upper portion of the circuit is to generate a 6SC frequency signal at the output of the voltage controlled oscillator 312 that is precisely controlled so that sampling that is performed by the analog-to-digital converter 95 is done precisely at the same phase of the color burst synchronizing signal at all times. This is important when it is considered that the phase of the video that is sampled will ultimately determine the color that is produced by the apparatus. Thus, the phase comparator 302 having one input supplied by the divided output of the VCO 312 through line 319 provides a phase lock loop that will lock the phase of the output relatively close to the video or reference subcarrier synchronizing signal phase appearing on line 300 supplied to the other input of the comparator 302. The divided output of the VCO 312 through the phase lock loop produces an SC signal that is generally within approximately 10.degree.. However, the digitized video output from the analog-to-digital converter 95 is also applied through line 308 to the precision digital burst phase decoder 307 which is enabled by the precision burst sampling gate signal received from the video input circuitry 93A over line 307a to generate an error signal derived during the burst interval of the video that is integrated by integrator 306 to provide an average value that is applied to the summer 304. This causes the voltage level out of the loop amplifier 311 controlling the VCO 312 to be adjusted to correct variations in the sampling times of the video signal as reflected in the burst samples provided to the decoder 307. The burst samples will represent the same quantity values for all lines if no variation in sampling times occur. By examining the sampled data actually appearing at the output of the analog-to-digital converter 95, it can be precisely determined whether the samples were taken at the proper phase and in this manner, the VCO output on line 315 which is applied to the divide by 2 counter 317 produces a PALE 3SC clock on line 318 which controls the analog-to-digital converter 95 for keeping the sampling at the proper phase. The precision digital burst phase decoder 307 effectively corrects any errors that may be produced due to temperature drifting and the like which can be on the order of 5.degree. to 10.degree.. In this regard, the phase of the video (or reference) subcarrier synchronizing signal on line 300 provides the basic lockup for the VCO 312 and the precision correction that appears on line 305 in the reference logic circuitry 125B is arranged to change the phase by a few degrees, i.e., up to about 20.degree..

With respect to the lower portion of the block diagram of FIG. 10A, the PALE flag generator 324 produces a PALE flag signal at the H/2 rate for switching a switch 325 which steers 1/2SC pulses into the set or reset terminals of the divide by 2 counter 317 that produces the PALE clock on the output line 318. The PALE flag changes state every line as will be described herein with respect to FIG. 10B. The PALE flag signal is nonsymmetrical so that the phase of the 3SC PALE clock is never reversed during the burst interval of the video signal even though it is reversed during the active video of alternate lines. Thus, the net effect is that only the portion of the line after burst is sampled with a clock signal whose phase is reversed on alternate lines, i.e., a nonsymmetrical signal. As is shown in FIG. 10A, the PALE flag generator 324 has inputs from the video (or reference) input circuitry 93A or 93B of H drive applied on line 326, a field index pulse on line 327 and a burst flag on line 328. The burst flag keeps the PALE flag generator from producing the PALE flag signal on output 323 until after burst has occurred, since the sampling phase of burst must not be altered for the operation of the burst phase decoder 307 in the upper portion of FIG. 10A. The PALE flag generator 324 provides an H/2 rate transfer reset pulse which is sent over line 324a to the encoder switch 126 which employs it during data transfer operations to generate a signal that is used by the encoder 96 to reset its sync word inserter.

The H drive and field index signals are also applied to a drive servo sync generator 330 which has an output extending to a drive sync switcher 331 through line 332 and it provides the basic drive sync signals on line 334 for each of the disc drives 73 when commanded by the control line 333 from the computer control system 92. The sync signals are required for all operations in which the information is transferred between a disc pack 75 and the signal system. The computer system 92 differentiates whether a record or playback operation is desired. The sync information is in the form of a multiplex sync signal that appears on lines 334 that extend to the disc drive units and includes a set of three consecutive wide pulses to indicate the first field being recorded or played back at a 15 Hz set rate as well as horizontal sync pulses (at H rate) and is used for control of the spindle servo motor. Color frame and related sync signals also are provided for control of the servo drive and for use by the reference clock generator in generating control signals used during playback operations. The color frame related sync signal is obtained from a color frame generator 301, which receives the 30 Hz field index pulse signal over line 327 and frequency divides it by 2 to obtain the 15 Hz color frame signal. The color frame signal is sent over line 329 to the disc drives 73 and the reference clock generator 98.

The specific circuitry that can be used to carry out the operation of the block diagram shown in FIG. 10A is illustrated in FIGS. 34A through 34D, which together comprise an electrical schematic diagram of the reference logic circuitry. Since the operation of the circuitry shown in the detailed schematic diagram is carried out generally in the manner as has been previously described with respect to FIG. 10A, it will not be described in detail herein. However, with respect to the digital burst phase decoder 307 shown in the upper portion of FIG. 34A, the digitized video subcarrier synchronizing signal or color burst in the form of 8 bits that is derived from the output of the analog-to-digital converter 95, appears on lines 308 which are connected to arithmetic logic units 335 which in turn connect to shift registers 336. The shift registers 336 are clocked by the logic circuitry, indicated generally at 337, which is activated upon receipt of the precision burst sampling gate over line 307a and together with the arithmetic logic units 335 perform the arithmetic steps that are necessary to determine the signal of the phase of the digitized color burst on line 309. The error of any sampling is determined by examining the quadrature component of the samples which would be zero if the samples are taken at the proper phase of the subcarrier color burst signal. More specifically, the quadrature component is proportional to the function X1 - 1/2(X2 + X3) where the samples X1, X2 and X3 are 120.degree. apart. The clocking logic 337 performs the sequence that enables the arithmetic units 335 and shift registers 336 to carry out the arithmetic computation which will produce either a plus or minus signal on line 309 indicating an error in the phase of the actual samples.

Turning now to FIG. 34A which contains circuitry 324 for generating the PALE flag signal on line 323, the H drive signal is inverted by inverter 342 and is applied via line 338 into the clock input of an FF 339 which is a divide by 2 having output line 340 applied to the input of a second FF 341 that is clocked by the burst gate or flag signal on line 328. Line 340 also extends to a NAND gate 343 as does the outline 344 from the FF 341. The operation of the PALE flag generator 324 will now be described in connection with the timing diagrams shown in FIG. 10B which has the H drive signal (line 326) shown in FIG. 10B(1), the signal on line 340 shown in FIG. 10B(2), the signal on line 344 shown in FIG. 10B(3), the burst gate clock on line 328 shown in FIG. 10B(4) and the output of the NAND gate on line 345 appearing in FIG. 10B(5). The PALE flag signal on line 323 is the inverse of the signal on line 345 by virtue of inverter 346. While the PALE flag signal occurs at a rate of H/2, FIG. 10B(5) shows it to be nonsymmetrical because the output of FF 341 appearing on line 344 and applied to the NAND gate 343 is delayed with respect to the output from the first FF 339 because the FF 341 is clocked by the burst gate rather than by H drive.

The reference clock generator 98 produces the basic timing signals for the apparatus during playback, data transfer, diagnostic and other like operations during which input video signals are not recorded and uses as its input timing reference the regenerated SC (3.58 MHz) that is produced by the input circuitry 93B and passed through the reference logic circuitry 125B. The reference clock generator has phase shifting capability for shifting the entire system phase and includes a phase locked loop and assorted counters and logic circuits to generate the timing signals with the desired system phase. It also generates control signals used by the data decoder and time base corrector 100 and the chroma separator and processing circuit 101. Also, in accordance with the present invention, the reference clock generator 98 identifies alternate reproductions of the recorded two field picture frame and issues a frame delay switch signal employed in the blanking insertion and bit meeting circuitry 127 to prevent jittering in the display of the output video signal that would otherwise exist because of the use of an H sync related timing control signal synchronized with the reference color subcarrier signal to control the processing of the reproduced video information.

The operation of the reference clock generator 98 will now be described in more detail in conjunction with the block diagram shown in FIG. 11A. As is shown therein, the top half of the circuitry produces various timing signals including several clock signals and the bottom half uses reference synchronizing information, such as color frame from the reference logic circuitry 125B and field index and horizontal drive signals from the reference input circuitry 93B and generates the control signals used by the time base corrector 565 (FIG. 13A) and chroma circuitry 101 and blanking insertion and bit meeting circuitry 127. More specifically, the SC signal is applied to the reference clock generator 98 at input line 340', causing the generator to produce 1/2SC, SC, 3SC and 6SC clock timing signals and various time base corrector pulse timing signals as indicated at the right of FIG. 11A. The reference clock generator 98 includes circuitry that is controllable by an operator through, for example, a thumb wheel switch 349 so that the phase of the output signals can be adjusted relative to the phase of the regenerated SC signal on the input by introducing various amounts of phase shift into the circuit and thereby set the playback system phase. Using the horizontal sync position control included in the reference input circuitry 93B and the SC phase control together enables an operator to determine and control the delay introduced to the playback signal channel over a large range in small increments. To control the phase of SC, the input regenerated SC signal on line 340' is divided by 2 by a divider 343', the output of which appears on line 344' that extends to two locations, one being the programmable counter 345' while the other is another divide by 2 divider 346' which in turn is connected by line 347 to a phase comparator 348. The thumb wheel switch 349 introduces a ten bit BCD number, ranging from 0 to 399, into the programmable counter 345' which has the effect of varying the phase of the subcarrier over a range of 0.degree. to 399.degree. in 1.degree. increments. The output of the programmable counter, which is a periodic signal whose duty cycle may be varied in increments of precisely 1/720 of its basic period by means of the thumb wheel switch 349, extends to a current switch 351a which modulates the current from a current source 351 of one of two matched current sources 351 and 353. This modulated current is coupled to to low pass filter 354a which develops a DC voltage proportional to the duty cycle of the signal on line 354.

A circuit of identical DC characteristics comprising the other matched current source 353, a current switch 353a and a low pass filter 355a, develops a DC voltage on line 355 which is proportional to the duty cycle of the output of the phase comparator 348. The voltages on lines 354 and 355 are applied to a differential amplifier 356, the output of which is extended by line 357 to the control input of a voltage controlled oscillator 358, which operates at a nominal frequency of 6SC. A number of dividers 360 (divide by 6), 363 (divide by 2) and 365 (divide by 2) sequentially operate on the output of the oscillator 358, producing a signal with a nominal frequency of 1/4SC on line 342' which extends to the second input of phase comparator 348, so that the duty cycle of the signal at the phase comparator output varies with the phase angle between its inputs. Under steady state conditions, the duty cycle of the signal on line 352 is forced to be equal to that of the signal on line 350 within a very small margin of error due to close matching of the current sources 351 and the DC impedance of filters 354a and 354b.

A change in the duty cycle of the signal at the phase comparator 348 output of 1/720 of its basic period requires a change of phase of 0.25.degree. between its inputs, which have a frequency of 1/4SC, and this in turn requires a change of 1.degree. between lines 340' and 361, where the frequency is SC. Thus, changing the value by one on the thumb wheel switch 349 causes a 1.degree. change in phase of the SC signal on line 361. The total range of phase comparator 348 (180.degree. at 1/4SC) corresponds to 720.degree. at 1SC. For convenience, the thumb wheel switch is limited to 399.degree., which still insures adequate overrange capability with respect to the necessary 360.degree..

The phase controlled oscillator 358 provides the phase continuous 6SC clock timing signal over its output line 341' and, through the cooperative operation of the chain of dividers 359, 360 and 363, causes phase continuous 3SC, SC and 1/2SC clock timing signals to be provided at the outputs as designated in FIG. 11A. The dividers also furnish 3SC and SC clock signals to the logic circuitry 362 that produces phase continuous SC rate read/write (R/WR) mode, write enable (WR EN), demultiplex (DMPLX) clock and multiplex (MPLX) clock signals used by the time base corrector circuitry 565 (FIG. 13A). The details of the logic circuitry are shown in FIGS. 35C and 35D and relationships between the signals provided by the logic circuitry can be found by reference to FIG. 11C. The schematic diagram illustrated by FIGS. 35A through 35D together with the timing diagram of FIG. 11B disclose the operation of one embodiment of logic circuitry 362 for providing phase continuous time base corrector clock signals with the desired timing relationships.

With respect to the lower portion of the circuitry shown in the block diagram of FIG. 11A, the circuitry redefines an H sync related, namely, H/2, signal so that it is synchronous with the phase continuous 3SC signal that is produced by the upper portion of the circuitry and occurs at the first reference horizontal line following alternate reference vertical sync's. As will become apparent upon consideration of the description of H/2 vs SC definition or reclocking circuit 367 hereinbelow, maintaining H/2 synchronized with respect to reference subcarrier and also so that it occurs at the first line of the first field of every two reference field sequence (which corresponds to the placement of the sync word in the video signal), requires frame rate phase inversion of the subcarrier rate clock controlling the reclocking circuit 367 to redefine H/2 with respect to the phase of SC. Subsequent reclocking of the redefined H/2 signal with the phase continuous 3SC clock signal within the circuit 367 introduces a 46 nsec (1/2 cycle of 3SC) picture frame to frame motion of redefined of H/2 relative to reference H sync use of the redefined H/2 in the time base corrector circuitry 565 to correct a repetitively reproduced video signal transfers the .46 nsec picture frame to picture frame motion to the video signal output by the time base corrector. This motion occurs because the reclocked and redefined H/2 is mispositioned relative to the proper reference H sync position on alternate picture frames and causes the time base corrector circuitry 565 to misposition the sync word a corresponding amount, or 1/2 cycle of 3SC, on alternate reproductions of a frame. As will be explained hereinbelow upon consideration of the sync word insertion circuitry portion of the encoder 96 (FIG. 12), the H/2 rate sync word is inserted in the video signal on alternate picture frames at a position that is displaced 1/2 cycle of SC from that corresponding to the reference H sync. This is because the sync word inserter is reset every frame and because the sync word is positioned on the first line of every picture frame, it being understood that the first line of successive picture frames have oppositely phased SC. The time base corrector circuitry 565 inherently removes all of this displacement except for the aforementioned 1/2 cycle of 3SC. A frame delay detector 368 of the reference clock generator 98 generates a frame delay switch signal used by the blanking insertion and bit muting circuitry 127 to correct for such motion. Also, it is not desirable to have the H/2 positive going transition of the unredefined H/2 signal coinciding exactly with a subcarrier transition in the reclocking circuit 367 because an ambiguously timed redefined H/2 pulse signal will be produced for use by the time base correctors 565 and errors in time base correction will result.

To produce an H/2 signal redefined with respect to the phase of the phase adjusted, phase continuous regenerated subcarrier signal, SC, provided by divider 360 is coupled to one input of a phase inverter 393 formed by an exclusive OR gate circuit. The other input of the phase inverter is coupled through a NAND gate circuit 397 to receive the 15 Hz color frame pulse signal generated by the reference logic circuitry 125B (FIG. 10A) and present on input line 396a. The level of the color frame pulse signal at the input of the phase inverter 393 determines the phase of SC at the output of the inverter, a high level resulting in the inversion and a low level not. Inversion of the phase of SC is necessary because an H/2 signal is preferred that is phase coherent with H sync. (In the recorded video signal, a sync word is inserted in the same lines for all picture frames of the video signal, which in this apparatus is the odd numbered lines of the 525 lines forming an NTSC picture frame.) Without inversion of the phase of SC, the phase of the redefined H/2 signal would change at a 15 Hz rate with respect to H sync by one half of an SC cycle. Such an H/2 signal would not be suitable as a reference for use in processing reproduced video signals during playback operations. The SC signal output by the phase inverter 393 is provided to the reclocking circuitry 367 and is used together with the reference H drive signal received over line 396 and the field index signal received over line 395, both signals provided by the reference input circuit 93B (FIG. 8A), to generate the H/2 signal defined with respect to the phase of SC. The reclocking circuitry 367 includes logic circuitry to assure that an unambiguously timed H/2 signal is produced and defined with respect to the phase of SC.

The output of the reclocking circuitry 367 is then applied to the frame delay detector 368 which produces the frame delay switch signal on line 369 that identifies the first or second playing of a reproduced still, composed of two television fields or a picture frame, so that the clocking circuitry for the blanking insertion and bit mating circuit 127 will know whether to insert an additional 1/2 period of 3SC offset for correcting the previously mentioned 45 nsec picture frame to frame motion of H/2.

The redefined H/2 pulse signal generated by the reclocking circuitry 367 appears on line 386 that is gated through gate circuitry 370 and 371 to appear on line 372 for use as the basic reference in the time base corrector circuitry 565 during playback operations, which is signified by an enabling signal provided on line 373 by the encoder switch 126 (FIG. 8A) from control signals issued by the computer control system 92. During playback operations, a high level signal appears on line 373 and the playback H/2 on line 386 will satisfy AND gate circuitry 370 and will appear on line 372.

In other operations, such as E to E and transfer, involving the processing of video signals in a playback channel, the H/2 signal as ordinarily generated by the H/2 vs SC definition circuitry 367 is not used. In E to E operations, continuous time base correction is not necessary since the video signal does not experience a record and reproduce process. Hence, the EE or PB command provided by the encoder switch 126 from control signals issued by the computer control system 92 is coupled over line 398 to the reference clock generator 98 associated with the playback channel selected for use to disable the phase alteration of SC. The phase alteration is disabled through the operation of the NAND gate circuit 397 placing a low level signal on the second input of the phase inverter 393. Furthermore, the EE or PB command is coupled to logic circuitry 399 that responsively generates an EE TBC disable signal used to allow the time base corrector circuitry 565 to operate for approximately ten lines at the beginning of each color frame and, thereby, generate the proper timing correction for each color frame or every 15 Hz. This timing correction is required because during the sync word insertion process for E to E operations the sync word generator is reset every two fields, i.e., picture frame. This results in a discontinuity of one half SC cycle in the position of the sync word every other picture frame or every 15 Hz.

When the apparatus is performing a transfer operation through a playback channel, a low level signal is placed on line 373 of the reference clock generator 98 associated with that playback channel. This enables the AND gate circuit 374 to pass a transfer H/2 signal present on line 375 to the OR gate circuit 371 which gates the transfer H/2 to the output on line 372. The transfer H/2 is derived from the sync word inserter portion of the encoder 96 circuitry. An output pulse from the encoder 96 that is coincident with the sync word or line identification is produced and that pulse is used as the time base corrector reference. The pulse appears on line 376 and passes through a shift register delay circuit 377 which correctly positions the pulse that is present on line 376. The transfer H/2 signal is positioned so that the digitized video signal provided to the encoder 96 during a transfer operation has a correctly identified location for insertion of a new sync word. Specific circuitry that can be used to carry out the operation of the block diagram shown in FIG. 11A is shown in FIGS. 35A through 35D. The operation of the specific schematic circuitry will not be described in detail since it carries out the operation as has been previously described with respect to FIG. 11A. However, with respect to the generation of the H/2 signal so that it is unambiguously redefined with respect to SC, the reclocking circuitry 367 includes an H/2 signal generator 378 comprising a divide by 2 counter and following pulse former respectively formed by an edge triggered flip-flop and following self resetting flip-flop. The counter receives at its clock input H drive signals present on input line 396 and provides an H/2 signal at its output. The H/2 signal is formed into a train of negative pulses, each at a positive going transition, by the H/2 generator's pulse former. The 30 Hz field index signal resets the counter portion of the generator 378 at beginning of the first field of every picture frame so that the phase of the H/ 2 signal is the same at the time of the first line of the first field of every picture frame.

The SC signal provided by the phase inverter 393 is also formed into a train of negative pulses by a pulse former 393a. A pulse coincidence detector circuit 378a formed by a low level AND gate and following D latch examines for a coincidence of the SC transition related pulses received from the pulse former 393a and the H/2 transition related pulses provided by a timing selection circuit 379 in response to each negative pulse provided by the pulse former portion of the generator 378. If the positive transition of the H/2 signal provided by the generator 378 becomes too close in time to the positive transition of the SC signal, the transition related pulses will overlap in time at the coincidence detector circuit 378a resulting in the toggling of the latch of the detector circuit. Toggling of the latch changes the level at an input of an exclusive OR gate 379a included in the timing selection circuit 379 to change it between its inverting and non-inverting mode. The timing selection circuit 379 includes a self resetting, edge triggered flip-flop 379b having its clock input coupled to the output of the exclusive OR gate 379a. By selectively inverting and not inverting the negative pulses provided by the H/2 signal generator 378, the positive edge of the pulse output of the exclusive OR gate is moved relative to SC. The timing selection circuit 379 cooperates with the coincidence detector circuit 378a to position the positive edge of the pulse output of the exclusive OR gate 379a so that unambiguous redefinition of H/2 will always result.

Redefinition of H/2 is performed by the reclocking edge triggered flip-flop 367a having its reset input coupled to an output of the timing selection circuit 379 and its clock input coupled to receive the SC signal provided by the phase inverter 393. Each H/2 transition related pulse resets the flip-flop 367a and the immediately following positive transition of the SC signal received at the clock input changes its state to thereby generate the redefined H/2 transition. A following latch 367b couples the redefined H/2 transition signal to a delay means 391 composed of a counter and following shift register operated to provide a properly timed H/2 signal on the line 380 extending to the frame delay detector circuit 368. The redefined H/2 transition signal output by the latch 367b is coupled to reset the delay means 391 and an SC signal, opposite in phase to that utilized in the reclocking circuitry 367, provided over line 392 clocks the delay means to effect issuance of the redefined H/2 transition signal to the detector 368.

With reference to the frame delay switch signal that appears on line 369 in FIG. 35D, it is a signal that changes level on alternate picture frames and is used in accordance with the present invention in the blanking and bit muting circuitry 127 for adjusting the half cycle of 3SC mispositioning of alternate picture frames as previously discussed. The operation of this portion of the circuitry will now be discussed in connection with FIG. 11C. The signal appearing on line 380 is an H/2 rate pulse signal which has been unambiguously redefined with respect to the phase of the regenerated SC that is inverted on alternate frames so as to insure that the SC redefined H/2 transition signal is stationary with respect to H sync reference. This transition signal is clocked into a shift register 381 by a phase continuous 3SC signal appearing on line 394 and appears on the first output line 385 delayed and synchronized to the 3SC signal. Because the continuous phase 3SC clock is an odd multiple of one half the picture frame frequency, its phase during a first picture frame is 180.degree. different with respect to H sync reference than its phase at the same time during the next picture frame and, hence, is also 180.degree. different frame to frame with respect to the redefined H/2 pulse. Because of this 180.degree. phase relationship difference, the positive transition of the 3SC clock shifts one half cycle picture frame to picture frame relative to the redefined H/2 pulse and consequently the clocking of the shift register 381 relative to the occurrence of the stationary H/2 pulse will change frame to frame by one half of the 3SC clock period. To detect the relationship between the redefined H/2 signal and the phase continuous 3SC clock signal, a stationary pulse is generated from the positive transition of the redefined H/2 signal and is used by the frame delay detector latch or D type flip-flop 368a to determine the phase of the 3SC clock at the beginning of alternate picture frames and provide the phase indicative frame delay switch on line 369 as shown in FIG. 11C. More specifically, the pulse forming circuitry comprised of inverter 382, resistor 388, capacitor 387 and NAND gate 389 generates a stationary pulse from the leading edge of the H/2 pulse signal present on line 380 at the input of the shift register 381. The stationary pulse has an interval of 3/4 of a cycle of 3SC and its leading edge (as well as that of the H/2 pulse signal) corresponds to the positive transition of the redefined H/2 signal. Because the shift register 381 is clocked by the phase continuous 3SC clock, the H/2 pulse signal will appear on the shift register's output line 385 at different times relative to its presence on the input line 380 depending upon the phase relationship of the redefined H/2 signal and 3SC signals. When the signals are in phase, the H/2 pulse signal appears on line 385 one cycle of 3SC after its presence on the input line 380. When the signals are out of phase, the H/2 pulse signal appears on line 385 1/2 cycle of 3SC earlier. The signal level on line 385 is strobed into the D flip-flop 368a by the positive going transition of the stationary pulse on line 384, which occurs 3/4 of a cycle of 3SC after the occurrence of the redefined H/2 pulse signal at the input of the shift register. The output of latch 368a on line 369 indicates whether the H/2 pulse was present on line 385 after a delay of 3/4 period thereby determining if the time delay between the positive going signals on lines 394 and 385 is 1/2 period of 1 period of 3SC. This signal on line 369 in turn is coupled to the blanking insertion and bit muting circuitry 127 to selectively insert a 1/2 3SC period offset in the clocking of the video data, compensating for the aforedescribed 46 nsec picture frame to frame motion of the redefined H/2.

With reference to the frame phase inverter switch signal that appears on line 356a in FIG. 35D, it is a signal that changes levels on alternate picture frames and is used in the chroma separator and processing circuitry 101 to effect inversion of the chrominance component included in the reproduced video signal on alternate reproductions of the two field color video signal. The playback burst is provided on input lines 361a by the data transfer circuitry 129 and is phase compared with the phase continuous SC by the exclusive OR gate 362a. SC and playback burst alternate between in phase and out of phase conditions with alternate reproductions of the two field color video signal, causing the level of the output of the exclusive OR gate 362a to change at a 15 Hz rate with the change occurring at the time of playback burst. The frame phase inverter switch signal is obtained by clocking the output of exclusive OR gate 362a through a latch 363a with one properly timed clock at every burst flag. The latch 364areceives at its D input the burst flag signal provided on line 360a by the reference input circuitry 93B and is clocked by the phase continuous SC provided at its clock input by the divider 360. Each time a burst flag signal is present on input line 360a, the latch 364a issues a pulse to the latch 363 a defined with respect to the phase of SC. This pulse is used to clock the level at the input of the latch 363a to its output. Because the level at the input of the latch 363a changes with alternate reproductions of the two field color video signal, the level at the output of the latch 363a also changes with alternate reproductions to produce the 15 Hz frame phase inverter switch signal on line 356a that defines when the chrominance should be inverted or not in the chroma separator and processing circuitry 101.

The encoder switch 126 described with respect to the block diagram of FIG. 8A is interconnected with the computer control system 92 and, upon receiving the appropriate commands, performs the principal function of selecting either the video data streams from the analog-to-digital converter 95 when in the record operating mode or the data streams that originate at the data transfer circuitry 129 when a transfer operating mode occurs. In the transfer mode, the recorded picture frame is transferred from one disc drive to another so that the video information does not go through the chroma separator and processing circuitry 101. Instead, it is directed to the encoder switch 126 to be thereafter encoded and recorded on another one of the disc drives. The encoder switch 126 also switches between the appropriate clock signals, i.e., 6SC and 1/2SC. It switches to clock signals generated by the reference logic circuitry 125A which are used when the video information from the analog-to-digital converter 95 is being recorded. During the transfer mode, it switches to the 6SC and 1/2SC signals provided by the reference clock generator 98 and are used as the basic reference clock signals during the recording of the transferred video signal, all of which is generally shown in the block diagram of FIG. 8A.

The encoder switch also performs functions in addition to switching the proper reference signals, depending upon whether the regular record or transfer modes are being performed. Circuitry is included for generating a blinking cross picture display signal, one diagonal line of which is supplied by one field and the other by the second field which provides an indication that the track has been deleted and is available to receive a still in that particular location. The encoder switch also includes circuitry that generates a PALE switch signal which terminates PALEing during the transfer process, the PALE switch (or flag) signal extending to the data transfer circuitry 129 which normally PALEs the data going to the chroma circuitry 101. The PALEing by the transfer circuitry is stopped because there is no need of aligning the samples line to line during a transfer mode of operation. The encoder switch also includes circuitry for performing diagnostic testing, which circuitry selectively generates a recurring sequence of digital information, as well as a random word for use in such testing. During the record mode, the encoder switch 126 couples the outputs of the analog-to-digital converter 95 and timing signals provided by the video input circuitry 93A and reference logic circuitry 125A to the encoder 96. The details of the encoder switch 126 capable of performing the various other operations generally described above besides a record operation are described in the above-identified related application, Ser. No. 763,371.

The encoder 96 shown in the block diagram of FIG. 8A of the video signal system contains circuitry which performs functions in addition to channel encoding the digitized data on each of the eight video data bit lines, the parity bit and the data track sequence as described hereinbelow. One of the additional functions involves the use of a parity generator to perform a parity check to verify that the data is correct on all of the eight data bit lines. The parity bit is optional and requires an extra data bit line such as is available in the apparatus described herein. The encoder 96 also generates and inserts the sync word (also referred to herein as the line identification or line ID). The sync word is in the form of a 7-digit binary number which is placed in alternate television lines, generally where the horizontal sync pulse had been previously located, it being understood that the horizontal sync had been stripped from the composite video signal by the video input circuitry 93. The sync word is inserted within one cycle of SC of the location previously occupied by the horizontal sync pulse, and the encoder 96 inserts the sync word into each of the eight video data lines, the parity bit line and the data track line before the channel encoding is performed so that the output of the encoder 96 which is connected to the electronics data interface 89 contains the sync word in each of the 10 data streams recorded on a disc pack 75 (or sent to the playback chanel 91 during E to E operations).

The encoder 96 operation will now be described in conjunction with a block diagram shown in FIG. 12 and schematic circuit diagrams of FIGS. 36A-D. NRZ-L data from the encoder switch 126 enters on input line 450 and exits on output line 451 of each data bit line after having been (i) checked for parity, (ii) had the sync word inserted in alternate (odd) lines and (iii) channel encoded in a format that is conductive to magnetic recording and reproducing the digitized information with respect to one of the disc packs 75. The input data on each data bit line is applied to one input of a data input AND gate 452 which is connected to a channel encoder 453, which may be switched between two channel encoding formats, both of which will be described hereinafter. In the schematic circuit diagram of FIGS. 36A-D, identical channel encoders for two video data bit lines are shown in their entirety. Identical channel encoders for the other video, parity and data track lines are contained in dotted line enclosures below the encoders shown in their entirety. A sync word input AND gate 454 in each of the 10 bit lines is used to gate the sync word into the encoder at the proper time. These AND gates are also arranged to insert a test signal in the 10 bit lines if desired, the test signal being provided on line 450a (FIGS. 36A and 36B) by a suitable test signal source, such as digital test pattern generator. A first clock generator 455 has input signals 6SC and 1/2SC applied thereto by the encoder switch 126 and provides various SC and 3SC outputs as shown. Two of the 3SC outputs are applied by lines 472 and 473 to a second clock generator 456 which provides two time displaced 3SC clock signals on the two lines 474 and 475 that are extended to the channel encoder 453 for clocking the same. The clock signal on line 475 is a .phi.1 clock and is displaced one-half cycle of 3SC from the clock signal on line 474, which is a .phi.2 clock. During recording operations, these time displaced clocks are derived from the continuous phase 6SC, 1/2SC signals generated by the reference logic circuitry 125A and provided to the encoder 96 by the encoder switch 126. During other operations, such as recording the blinking cross delete signal, the reference clock generator 98 provides the clock signals. The .phi.1 and .phi.2 3SC clock signals are used to drive the channel encoder 453 so that a continuous channel encoded digital signal without phase discontinuities is provided at the output on line 451.

The clock generator 455 has an SC clock output 471a driving a .div. 455 divider 457 which can also be reset by a reset pulse provided by the encoder switch 126 on line 463 at a 30 Hz rate. The divider 457 sets a flip-flop (FF) 458 through the start line 464 and subsequently resets the FF 458 when a pulse appears on the stop line 465 extending to the reset pin. The START and STOP pulses define a window during which a single 7-digit binary sync word provided at the output of a sync word generator 459 can be inserted in all data bit lines simultaneously.

During the vertical blanking period, a pulse is applied to a monostable multivibrator (MS) 460. The multivibrator is active for a period of about 10 lines of the vertical blanking period by switch vertical signal provided on line 466 by the encoder switch 126 and its output is applied to one side of gate 461 (shown in this block to be an NAND gate), the other side of which is supplied by the output of the window generating FF 458. The output of the NAND gate 461 extends to the other input of the AND gate 454 as well as through an inverter 462 to one side of the AND gate 452.

During the operation of the encoder circuitry 96, it is desired that the data stream for each bit be applied on an input such as input 450 which is representative of the eight separate data input lines, each of which is connected to a separate encoder 453 and the associated data and sync word input AND gates 452, 454 and inverter 462 so that a data output line 451 exists for each of the data bits and each of the data streams is properly channel encoded and has a sync word inserted therein. Since it is desired that the sync word occur close to the former location of the horizontal sync pulse and since it is also desired that it not be confused with data of the data stream, the data bit lines input to the channel encoders 453 are disabled by the data input gates 452 when the sync word is inserted during a sync word gate window that is generated by the divider 457 and FF 458. More specifically, the divider 457 provides a START pulse for setting the FF 458 and this enables one input of each AND gate 454 while simultaneously disabling each AND gate 452 thereby blocking the data entering on lines 450. The divider 457 issues a pulse to over line 467 the sync word generator 459 twelve data bit intervals after the generation of the START pulse and the sync word generator 459 then generates the 7-digit binary word which is applied to the upper input of all AND gates 454 which have previously been enabled. The AND gates 454 pass the sync word into each channel encoder 453 where it is encoded onto the data stream. After the sync word has been generated, the divider 457 issues a STOP pulse 29 data bits later which resets the FF 458, disabling all AND gates 454 and simultaneously enabling all AND gates 452 so that the data on lines 450 will be passed into the channel encoders. It should be understood that the data stream line 450 is continuous in its flow and that disabling the AND gates 452 merely blocks it from passing. Hence, the information is only discarded in a sense during the insertion of the sync word. However, since the sync word is inserted approximately at the previous location of the horizontal sync pulse, no active video informational data is lost.

During the vertical blanking interval, the multivibrator 460 provides an output to the NAND gate 461 which occurs for an interval of about 10 lines. This disables the data input AND gate 452 during the 10-line interval of the blanking period so that the received data is prevented from passing to the channel encoder during this interval. Thus, the only data or logical 1 bits that appear on the output data line 451 during the 10 line interval of the vertical blanking period are those in the sync words that appear every other line, as previously described, and pass through the sync word gate 454. This insures that the decoder and time base corrector circuitry 100 will be locked on the actual sync word during playback rather than some randomly occurring sync word bit pattern that might be contained in the active video information during the flow of the data stream.

Another aspect of the operation of the encoder 96 will be more clearly understood by referring to FIGS. 8A and 8B. The electronics data interface 89, disc drive data interface 151 and data select switch 128 couple the encoder 96, disc drive 73 and decoder and time base corrector circuitry 100. It should be appreciated that during a seek operation when the heads in the disc drive 73 are moving between tracks, it is desirable to prevent the introduction of perturbances in the signal system. Ordinarily, the record signal processing system 88 will provide at the output of its encoder 96 digitized data even in the absence of an input video signal. While this signal will represent noise information, the digital signal processing electronics of the apparatus cannot distinguish between digitized noise and digitized video information. This factor is taken advantage of when the apparatus is performing a seek operation. During seek operation, the transducing heads create noise signals that do not conform to the channel encoded format of the digital data ordinarily present in the signal system. Such noise signals, if permitted to enter the playback channel 91, undesirably perturb the phase lock loops of the decoder and time base corrector circuitry 100. To avoid such perturbances, the disc drive data interface 151 is switched (as in an E-to-E operation) to reroute the output provided by the encoder 96 to the decoder and time base corrector circuitry 100. In this manner, the decoder and time base corrector circuitry 100 is receiving channel encoded digital signals that maintain the respective phase lock loops in the circuitry 100 within their normal operation range. Hence, when the heads of the disc drive 73 are properly positioned and playback data provided to playback channel 91, the decoder and time base corrector circuitry 100 are prepared to immediately provide the output decoded and time base corrected signals.

In addition, the encoder 96 also serves to cause black level data to be generated for use in recording on the disc surfaces as previously described during the first two revolutions of the disc pack 75 prior to the recording of the video signal information on the subsequent two revolutions of the disc pack. Accordingly, the prerecord line 470 (FIG. 36A) extending from the electronics data interface 89 is activated as a result of signals provided by the disc drive data interface 151 and causes NAND gate 461 to block any logical "1's" as may be present on the input lines 450 thereby producing the black level at the input of the channel encoder circuitry 453. It should be noted, however, that the encoder 96 still functions to insert the sync word in the black level signal.

The NRZ-L data in each data bit line 450 is channel encoded selectively by the channel encoder 453 into the DC free self clocking channel code described in the aforementioned U.S. Pat. No. 4,027,335 or the self clocking channel code described in U.S. Pat. No. 3,108,261. As will be described further hereinbelow, the two position code selection switch 480 selects between the two channel codes. In both codes, the NRZ-L data bit stream on a data bit line is broken into discrete bit times commonly designated as data bit cell times. For the channel code described in the U.S. Pat. No. 3,108,261 Patent, the code rules followed result in logical first bits, e.g., logical 1's to be represented by signal transitions at a particular location in the respective bit cells, specifically at mid-cell, and logical second bits or logical 0's to be represented by signal transitions at a particular earlier location in the respective cells, specifically at the beginning or leading edge of each bit cell. Any transition occurring at the beginning of one bit interval following an interval containing a transition at its center is suppressed.

In the channel code described in the above-identified 4,027,335 patent, the input data stream in each data bit line may be viewed as the concatenation of variable length sequences of three types: (a) sequences of the form 1111---111, any number of logical 1's but no logical 0's; (b) sequences of the form 0111---1110, any odd number of consecutive 1's or no 1's, with 0's in the first and last positions; (c) sequences of the form 0111---111, any even number of consecutive 1's preceded by a 0. A sequence is of type (c) only if the first bit of the next following sequence is a zero. Sequences of types (a) and (b) are encoded according to the code rules described in the U.S. Pat. No. 3,108,261. A sequence of type (c) is encoded according to the U.S. Pat. No. 3,108,261 rules for all bits except the last logical 1, and for this 1 the transition is simply suppressed. By this means, the type (c) sequence, viewed in isolation, is made to appear the same as a type (b) sequence, that is, the final logical 1 looks like a logical 0.

By definition, the type (c) sequence is followed immediately by a logical 0 at the beginning of the next sequence. No transition is allowed to separate the type (c) sequence from the following 0. Therefore, the special coding is distinctive for decoding purposes. The decoder must merely recognize that when a normally encoded logical 1 is followed by two bit intervals with no transitions, then a logical 1 and logical 0 should be output successively during those intervals. Other transition sequences are decoded as for the Miller code.

The encoding procedure for this code requires that a modulo-2 count be maintained of the number of logical 1's output by the encoder since the last previous 0 which was not the final bit of a type (b) sequence. If the count is 1 (odd number of 1's) and the next two bits to be encoded are 1 and 0 in that order, then no transitions are output during the next two bit intervals. If the next subsequent bit is another 0, then it is separated from its predecessor by a transition in the usual aforementioned U.S. Pat. No. 3,108,261 code fashion. This channel code provides for the transmission of data in binary form over an information channel such as a magnetic record/playback system, incapable of transmitting DC, the information being transmitted in self-clocking fashion.

With respect to the channel code, it makes no difference which binary state is considered logical 1 and which binary state is considered logical 0. In the foregoing and following decriptions the state normally marked by mid-cell transitions is considered the 1 state, whereas the state normally indicated by cell edge transitions is considered the 0 state,

The channel encoders 453 illustrated by the FIG. 36A through 36D operate in accordance with the aforedescribed code rules. FIG. 36E is a timing diagram depicting the operation of the channel encoder 453 included in one of the data bit lines 450. With switch 480 shown in FIG. 36B in the indicated position, the channel encoders 453 provide encoded data in accordance with the code rules of the aforementioned U.S. Pat. No. 4,027,335. In its other position, the channel encoders 458 provide encoded data in accordance with the code rules aforementioned U.S. Pat. No. 3,108,261.

The channel encoder will now be described with the code selection switch 480 set as shown in FIG. 36B to effect channel encoding of one of data bit streams according to the code rules of the aforementioned U.S. Pat. No. 4,027,335. A description of the differences in the operation of the encoder when the switch 480 is set in its other position to effect channel encoding of the data bit stream according to the code rules of the aforementioned U.S. Pat. No. 3,108,261 will follow.

As described above, data encoded according to the U.S. Pat. No. 4,027,335 code rules requires examining two successive data bits to be encoded whenever the modulo-2 count of logical 1's previously encoded is odd. For this purpose, each channel encoder 453 includes a pair of serially connected input latches 481 and 482 clocked by the trailing positive edge of the .phi.2 3SC clock signal (FIG. 36E - (2)) on line 474a, which is coupled to line 474 by an inverter 483. The input latches provide a two bit cell delay from the input of latch 481 to the output of latch 482. At each trailing positive edge of the .phi.2 clock, latch 481 is operated to latch the present data level of the bit stream at its input so that it appears at its output (FIG. 36E - (3)) and latch 482 is operated to latch the preceding data level of the bit stream contained in latch 481 so that it appears at its output (FIG. 36E - (2), (3) and (4)). Therefore, the outputs of the latches 481 and 482 contain the data bits of two consecutive bit cells that are to be encoded.

The outputs of the latches extend to the inputs of three NAND gates 486, 487 and 488 for separately gating through pulses corresponding to logical 1's and 0's in the data bit stream. NAND gate 486 receives three inputs; one from the output of latch 481, one from the output of latch 482 and .phi.1 clock pulses (FIG. 36E - (1)) placed on line 475 by an inverter 484 connected to the output line 475a of the clock generator 456. This NAND gate is enabled to provide an output pulse 489 (FIG. 36E - (6)) upon receipt of a .phi.1 clock whenever its other two inputs are at a low level, which occurs only when successively received data bits are logical 0's. Consequently, NAND gate 486 issues logical 0 related pulses that are marked by transitions in the channel encoded format of the data stream output by the channel encoder 453. A logical 0 bit that immediately follows a logical 1 bit is blocked from passage by the NAND gate because the latch 482 will be high when, for example, the .phi.1 clock pulse 490 (FIG. 36E - (1)) occurs. Hence, the channel encoder 453 follows the code rules described in the aforementioned U.S. Pat. No. 3,108,261 for successively occurring logical 0 data bits.

On the other hand, the NAND gate 487 has two inputs and is enabled to provide an output pulse (FIG. 36E - (5)) upon receipt of a .phi.1 clock for all logical 0 data bits. Because the output of latch 482 enables the NAND gate 487, the logical 0 related pulses are provided one data cell time after the data has been latched into the channel encoder 453.

NAND gate 488 has three inputs and is enabled by the inverted output of the latch 482 to provide an output pulse (FIG. 36E - (7)) upon receipt of a 100 2 clock for all logical 1 data bits, unless a high level bit suppression command 491 (FIG. 36E - (10)) is placed on the input of the NAND gate by a line 492 extending from a bit suppression NAND gate 493 as will be described hereinbelow. NAND gate 488 generates the logical 1 related pulses during the interval of the .phi.2 clock, hence, before the latch 482 is clocked by the trailing positive edge of the .phi.2 clock. The logical 1 related pulses are provided by the NAND gate 487 one data cell time after the data has been latched into the channel encoder 453 at latch 481.

An OR gate 494 has two inputs connected to receive the logical 0 pulses 489 (FIG. 36E - (6)) provided by NAND gate 486 according to the 3,108,261 Patent code rules and the logical 1 pulses 515 (FIG. 36E - (7)) provided by the NAND gate 488. The output of the OR gate 494, which appears on the encoder output line 451, will, therefore, be a train of pulses (FIG. 36E - (14)) that occur according to the code rules for the channel encoder. Hence, the NAND gates 486 and 488 together with the OR gate 494 serve to encode the incoming NRZ-L data stored by the latches 481 and 482 into the selected channel code format. The NAND gate 487 operates with bit suppression logic circuitry 500 described below to control the selective suppression of logical 1 data bit related transition in the channel encoded data. By disabling the bit suppression logic circuitry 500, as would occur by changing the position of the switch 480 from that shown in FIG. 36C, the NAND gates 486 and 488 will encode the data according to the U.S. Pat. No. 3,108,261 rules.

To encode the data bit stream according to the aforementioned U.S. Pat. No. 4,027,335, the bit suppression logic circuitry 500 includes two modulo-2 counters 495 and 496 for counting encoded logical 1's and 0's and, together with cooperating gate circuitry, effecting the generation of the bit suppression command on line 492 that suppresses selective logical 1 bit related transitions in the channel encoded data appearing on line 451. The modulo-2 counter 495 counts the logical 0 related pulses coupled to its clock input by the NAND gate 487. Logical 1 related pulses provided by NAND gate 488 are coupled to the clock input for counting by the modulo-2 counter 496. Counter 495 recognizes the beginning of each sequence by toggling in response to logical 0 pulses each time a logical 0 is encoded and being cleared each time a logical 1 related transition is suppressed. As can be seen from the aforedescribed code rules, counter 495 toggles twice during a type (b) sequence and never changes state during a type (a) sequence, and therefore is in its cleared state before the start of any sequence. The bit suppression logic circuitry 500 must recognize the end of a type (c) sequence. Modulo-2 counter 496 is employed in the performance of this function by toggling in response to logical 1 pulses each time a logical 1 is encoded and being cleared in response to logical 0 pulses each time a logical 0 is encoded. Waveforms (8) and (9) of FIG. 36E illustrate the respective operations of the modulo-2 counters 495 and 496 if their outputs are not connected together at the wired-OR 501. Waveform (13) of FIG. 36E illustrates the actual state at the wire-ORed connection 501. As should be appreciated from the foregoing, if counter 496 is not in its cleared state, the counter 495 is in its cleared state, the present bit to be encoded is a logical 1 and the next following bit is a logical 0, the bit suppression command is provided by NAND gate 493 on line 492 to disable the NAND gate 488 and thereby suppress the encoding of the present logical 1 bit.

Considering the cooperating gate circuitry for controlling the clearing of the two modulo-2 counters 495 and 496, counter 496 has its set terminal coupled to the NAND gate 487 so that its output is set high each time a logical 0 related pulse is output by the NAND gate 487. The counter 495 has its set terminal coupled to the output of a NAND gate 497 so that its output is set high each time a logical 1 related transition is suppressed in the channel encoding of the data bit stream. For reasons that will become apparent from the following description, a pair of capacitors 498 and 499 are connected in the output circuits of the modulo-2 counter 495 and NAND gate 493, respectively, to delay the set logic level of counter 495 appearing at the wired-OR 501 and removal of the bit suppression command from NAND gate 488.

The bit suppression command is generated by the NAND gate 493 that examines the first of consecutive data bits to be encoded and which is present in inverted form at the output of the latch 482, the next following of the consecutive data bits to be encoded and which is present at the output of the latch 481 and the counter states of the modulo-2 counters 495 and 496. If either one of the counter outputs at the wire-OR 501 is high, the NAND gate is disabled. However, whenever the beginning of a type (c) sequence occurs, both counters 495 and 496 will be low, thereby placing an enabling signal at the input of the NAND gate 493. If the next two bits to be encoded are a logical 1 followed by a logical 0, the bit suppression command 491 will be generated and placed on line 492 upon the occurrence of the .phi.2 clock pulse 502 (FIG. 36E - (2)) immediately preceding the .phi.1 clock pulse 490 that would effect the generation of the logical 1 related pulse through NAND gate 493. Hence, when the .phi.1 clock pulse 490 (FIG. 36E (2)) occurs on line 474 that would cause the NAND gate 488 to generate a logical 1 bit pulse, the NAND gate 488 is disabled by the bit suppression command on line 492 and the logical 1 bit pulse is suppressed as represented by the pulses 512 shown in phantom at line (14) of FIG. 36E. The bit suppression command is terminated upon setting the counter 495. The set pulse 505 (FIG. 36E - (12)) is provided by the NAND gate 497 in response to the bit suppression command 491 (FIG. 36E - (10)) on line 510 and the aforementioned .phi.1 clock pulse 490, which occurs 1/2 cycle of 3 SC after the .phi.2 clock pulse of about 47 nanoseconds. To insure that the counter 495 is not set and the bit suppression command not removed until after the .phi.1 clock pulse 490 has ended, the delay capacitors 498 and 499 are provided to delay the return of the counter 495 to its high set state, hence, disabling of the NAND gate 493 and to delay the return of NAND gate 493 to its low disabled state, hence, extending the duration of the bit suppression command 491. The effect of the delay is seen at the rounded portions 508 and 509 of the waveforms (10) and (13) of FIG. 36E.

To disable the bit suppression logic circuitry 500, switch 480 is placed in the position that places a high level signal (ground in the channel encoder 453 of this apparatus) on the set line 510 for the counter 495. This places the counter permanently in its set state, thereby placing a disabling high level signal permanently at the wire-OR input of the NAND gate 493. Hence, bit suppression commands 491 can not be generated and bits will not be suppressed.

Commonly, self clocking channel encoded data code formats carry data and clock information as particularly placed transitions between two signal levels. When such encoded data is sent through a transmission channel, it usually experiences some timing distortion because of the non-linear characteristics of most transmission channels. If the timing distortion is significant, errors may result because of the inability of the channel decoder to determine the correct location of the transmitted transitions. Furthermore, at high data rates, such as found in the apparatus described herein, the timing distortion may result in unacceptable errors in the transmitted data. This is particularly the case where, as in the case of the channel codes selected for use in the apparatus herein, oppositely directed transitions carry the data and timing information. Non-linear transmission channels will alter the positively and negatively going transitions in a non-linear manner with respect to time. Hence, level sensitive data detectors commonly used at the terminal of a transmission channel to restore the transmitted data so that it has properly positioned transitions will position the positive and negative transitions differently. Different positioning occurs because positive transition with substantial timing distortion will reach the level selected for sensing the presence of transitions after a time after its

In the disc drives utilized in typical computer processing apparatus, such as the aforementioned Ampex model DM 331 disc drive, the disc spindle motor drive is free running. To provide desired servo control for the disc spindle motor drive, the motor drive circuits have been modified for the unique application in the present apparatus. The operation of the motor driving the disc will now be described in connection with FIG. 27 which is a block diagram illustrating the operation of such circuitry for controlling the driving of the motor in the computer disc drive so that it is locked to vertical sync and correctly positioned relative to the timing so that recording, playback and transfer operations are carried out with the proper timing.

Referring to FIG. 27, a block diagram of the circuitry which operates the drive motor and servo control system is illustrated. The detailed electrical circuitry of the modified Ampex model DM 331 disc drive that carries out the functions that will be generally described with respect to FIG. 27 are contained in FIGS. 32A and 32B which are schematic diagrams of the disc drive phase lock control and FIGS. 45A and 45B which are schematic diagrams of the disc drive motor logic and predriver circuitry which is used during start up of the disc drive motor. Referring to FIG. 27, when the three phase induction motor 2040 for the drive is to be started up, it is started using three phase AC power from the power lines 2041 which pass through relays 2042 and power the motor nominal position that is different from that required by a similarly distorted negative transition.

To enhance the reliability of transmission of channel encoded data in which oppositely directed transitions carry the data and clock information, each of the channel encoders 453 encodes the data bit stream at its input by providing pulses in accordance with the rules of the selected channel code at the transition locations of the channel encoded format. In the particular channel encoder used in the apparatus described herein, logical 1 data bit pulses 515 (FIG. 36E-(7) and )14)) are provided at the data cell boundaries to define logical 1 bit related transitions that appear in the channel encoded data and logical 0 data bit pulses 489 (FIG. 36E-(6) and (14)) are provided at center of a data cell to define logical 0 bit related transitions that appear in the channel encoded data. The transition-related pulses are generated by the clock generator 456 to have a precisely defined edge, the leading edge being selected. The second clock generator 456 includes two one-shot multivibrators that are clocked by the oppositely phased 3SC clock signals provided by the first clock generator 468 over lines 472 and 473. Since the leading edges of the positive pulses generated by each of the one-shot multivibrators are defined by rapidly switching the multivibrators from its stable state to its quasi-stable state (there being no significant time constant determining components involved), each leading edge will be identical to all others and occur at a precise time following the occurrence of the positive clocking transition of the clocking signal. The two multivibrators of the second clock generator 456 thusly provide .phi.1 and .phi.2 clock pulse trains, which in the embodiment described herein have a pulse width of about 17 nsec, with the leading edges of the pulses of each train precisely defined with respect to each other and those of the other train. As described hereinbefore, the .phi.1 clock pulses provided on line 475 are gated through the NAND gate 488 as logical 1 data bit transition related pulses that appear in the channel encoded data and the .phi.2 clock pulses provided on line 474 are gated through NAND gate 486 as logical 0 data bit transition related pulses that appear in the channel encoded data. Since the NAND gates 488 and 486 and in an enabled condition at the times the .phi.1 and .phi.2 are received for transmission as transition related pulses (FIG. 36E- (4), (7) and (14) for logical 1 bit pulses and FIG. 36E-(3), (4), (5), (6) and (14) for logical 0 bit pulses), their respective leading edges will not be noticeably affected by the transmission through the NAND gates. Because the transmission channel over which the pulses are sent will act on identical pulse edges the same, the precise locations of the transition-related positive pulse edges, hence, data signal transitions themselves, are not lost as a result of any distortion that may be introduced to the pulses by the action of the transmission channel.

The channel encoded transition related pulses output by the encoder 96 over lines 451 are coupled by the electronics data interface 89 to the transmission line 152 extending to the disc drive data interfaces 151 associated with the disc drives 73. The electronics data interface 89 includes conventional logic converters which convert the TTL logic on lines 451 to emitter coupled logic levels which provide complementary level pulses on two lines in a manner that is described hereinbelow with reference to FIGS. 43A and 43B. The interface 151 of the disc drive selected for recording the video data passes the data to the selected drive's record amplifier and head switch circuitry (FIGS. 44A and 44B). A divide by two JK flip flop 1070 included in each data bit line receives the transition related pulses and is responsive to the leading edges of the transmitted pulses to be rapidly switched between its two stable conduction states. This converts the transmitted pulse form of the channel encoded data to the level transition form for recording as transitions between two signal states. Prior to being converted by the JK flip-flop 1070, the transmitted pulses in each data bit line are passed through a differential amplifier line receiver 2020 included in the disc drive data interface (FIG. 46A) of the kind described hereinafter with respect to the decoder portion 525 (FIG. 37A) included in the data decoder and time base corrector circuitry 100 to regenerate the transmitted pulses with precisely defined leading edges after passage through the associated transmission line of the transmission line (FIG. 8B) bus 152.

The data bit streams of channel encoded data, comprising 8 video data bit streams, 1 parity bit stream (if a parity bit is added) and 1 data track bit stream, transmitted by a disc drive 73 (FIG. 8B) over a transmission line bus 154 are received by one or more of the playback channels 91 (FIG. 4) selected by the data select switch 128. At the input of each playback channel, each of the 10 transmitted data bit stream is received by a separate data decoder and time base corrector included in the circuitry 100 for decoding the channel encoded data back to the NRZ-L form of digital code and then time base correcting the NRZ-L data to remove any intra channel and inter channel bit time displacement errors that may be present in the received data streams. Bit time displacement errors result from the data transmission channel acting on the transmitted data to introduce intersymbol interference and reflections caused by impedance discontinuities in the transmission channel. This disturbs the timing of the data transmitted in the channel. In a video recorder data transmission channel, bit time displacement errors commonly are a result of changes in record medium dimensions, usually caused by environmental changes, of differences in the relative head to medium record and reproduce velocities of the relatively transported head and record medium and of machine to machine mechanical variations resulting in geometric differences between the heads and record medium. Video disc recorders utilizing rigid record media, such as the disc packs 75 used in the apparatus described herein, ordinarily do not cause large time displacement errors in the transmitted apparatus, particularly, at the data rates common for analog type video disc recorders that are in wide use today. The rigid record media used in such recorders are dimensionally stable and the servo mechanisms used are able to maintain the relative transport of the heads and rigid record media within sufficient tolerances so that time displacement errors are kept small. In some applications of video disc recorders, the time displacement errors are so small as to be insignificant and time base correction is not necessary.

However, as described herein, the present apparatus in which the time base corrector circuitry is used employs (with little modification) highly reliable disc drives that have been specifically designed and manufactured for computer data processing. Unfortunately, the computer disc drives do not maintain the relative head to disc velocity stable enough to avoid the introduction of intolerable bit time displacement errors into the data bit streams when such disc drives are used in the present apparatus to process video data. This is because the disc pack spindle in the drive is not servoed but instead is driven by a common three phase AC motor referenced to a relatively unstable line voltage and the rotational position of its disc pack is not controllable with respect to an external reference. The resulting position errors and bit time displacement errors are particularly detrimental at the high data bit rates, i.e., 10.7 MHz, required to faithfully process broadcast quality video data without reduction in the quality of the video information. Therefore, to take advantage of the mechanical reliability of the existing computer disc drive design, the apparatus described herein is provided with a positional servo for the AC motor and time base corrector circuitry to remove any unacceptable time displacement errors introduced into the data bit streams rather than altering the reliable design of the computer disc drives.

As described above, before the received data bit streams are time base corrected, each channel encoded data bit stream is decoded back to its original NRZ-L digital form. For this purpose, and with reference FIGS. 37A and 37B, the data decoder and time base corrector circuitry 100 includes for each data bit line a channel decoder circuitry portion 525 having a pair of input terminals 526 coupled to the data select switch 128 (FIGS. 8A and 8B) for receiving channel encoded data, which as described hereinbefore, is in the form of channel encoded transition related pulses, such as pulses 515 and 489 shown in FIG. 36E-(14). The pair of input terminals 526 are coupled to a differential amplifier line receiver circuit 527 connected to reject common mode noise in the pair of complementary transition related pulses received from the transmission line pair included in the transmission line bus 154 after passage through the data select switch 128 (FIG. 8B). In addition, the differential amplifier line receiver circuit 527 regenerates a single transition related pulse from each transmitted pair of complementary transition related pulses so that the regenerated pulse has a well defined leading edge properly positioned according to the code rules of the channel code selected for originally encoding the video NRZ-L data. More specifically, the differential amplifier line receiver circuitry 527 provides a single regenerated transition pulse with leading and trailing edges provided when the levels of the edges of the received complementary pulses are the same. By examining the edges of the transmitted complementary pulses in this manner, the leading edges of all regenerated pulses will be properly positioned according to the channel encoding rules because the same sense, i.e., leading positive going and leading negative going, edges of each pair of the complementary pulses are employed to define the occurrence of the leading edge of each regenerated transition related pulse. Because the transmission channel through which the transition related pulses are sent to the decoder circuitry 525 affect identical pulse edges the same, any time distortion introduced to the pulse edges will not effect the regneration of the transition related pulses.

Following the regeneration of the transition related pulses, they are coupled over line 528 to clock a one shot multivibrator 529 at each occurrence of a regenerated pulse, using the defined leading edge to effect clocking. The one shot 529 is rapidly switched from its stable conduction state to its quasi-stable conduction state to provide the precisely defined leading edge of the transition related pulses. The one shot 529 has one of its outputs connected to line 530a that extends to the clock input of a divide by two flip flop 531. Upon the occurrence of each regenerated transition related pulse, the flip flop 531 is rapidly switched between its two stable conduction states by the leading edges of the regenerated pulses and thereby converts the pulse form of the channel encoded data to the level form for subsequent decoding of the data back to its original NRZ-L digital form as will be described hereinbelow.

The one shot 529 provides complementary outputs of the channel encoded data on line 530a and 530b. The complementary outputs are coupled to a 6SC clock generator 532 which provides complementary 6SC clock signals on its output lines 533 and 534 for use by the data decoder circuitry 525 for decoding the received data. The clock generator includes a 6SC voltage controlled oscillator 537 which is locked by an operatively associated phase detector 535 to the phase of the data clock carried by the channel encoded data. The complementary transition related data pulses output by the one shot 529 on lines 530a and 530b are coupled to the input of the phase detector 535, which has its output on line 536 coupled to the control input of the 6SC voltage controlled oscillator 537. The phase detector 535 examines the phase of the 6SC clock provided by the oscillator 537 with respect to the received and regenerated transition related data pulses and provides an error correction signal to the oscillator via the phase error smoothing capacitor 538. A change in the phase of the received data causes the phase detector 535 to change the average voltage level on the capacitor 538 by a corresponding amount and thereby cause the phase of the 6SC clock provided by the voltage controlled oscillator 537 to be adjusted to clock carried in the channel encoded data.

The phase detection operation is performed by a pair of matched current sources 540 and 541, each having an output line 542 and 543 respectively connected to the line 536 coupled to the error averaging capacitor 538. In the absence of a transition related data pulse, the line 530b extending from the one shot 529 is high, which enables the current source 541. Because the base electrodes of each transistor of the differential pair forming a current switch 545 at the output of the current source 541 are grounded, the current provided by the current source 541 divides equally in the two current paths defined by the current switch 545. Current in the path defined by the current switch 545 connected to thw output line 543 flows onto line 536 to change the error smoothing capacitor 538 to a level which, when a data stream is not input to the decoder circuitry 525, will cause the voltage controlled oscillator 537 to provide a 6 SC clock at a nominal frequency and phase. Thus, even in the absence of a data bit stream at the input of the decoder circuitry 525, a 6 SC clock is provided at its nominal frequency. This facilitates rapid synchronization of the oscillator 537 to data clock when a data bit stream is initially received and proper decoding of the channel encoded data.

When a transition related data pulse is received on the input line 526, the one shot responsively provides a high level signal on line 530a and a low level signal on line 530b for an interval determined by its time constant circuit 529a, which in the decoder circuit described herein is about 17 nsec. The low level signal on line 530b disables the current source 541, thereby terminating the provision of charging current through the current switch 545 to the error smoothing capacitor 538. However, the high level signal on line 530a enables the other current source 540, which provides charging current to the error detection capacitor 538 in accordance with the relative conduction periods of the halves 544a and 544b of a current switch 544 formed by the transistors arranged in circuit as a differential pair. The transistors forming the two halves 544a and 544b of the current switch have their respective base electrodes coupled to receive the 6 SC clock provided over line 533. When the clock is at a low level, transistor 544a is disabled. However, the other transistor 544b is allowed to conduct because the long time constant RC circuit 547 holds the voltage at its base electrode at an average voltage level which is more positive than the low level of the 6SC clock. Consequently all of the current furnished by the current source 540 will flow through the one enabled transistor 544b to the output line 542 of the current source 540.

When the 6 SC clock goes high, the base of the transistor 544a goes more positive than the base of the transistor 544b. Therefore, transistor 544a is enabled and transistor 544b disabled. This removes the current flow to the error smoothing capacitor 538. If the transition related data pulse received by the current source 540 is positioned in time relative to the 6 SC clock provided to the current switch 544 so that low to high level transitions in the 6 SC clock occur at the center of the transition related data pulses, each transistor 544a and 544b of the current switch will be enabled for equal intervals and the voltage on the error detection capacitor 538 will be maintained at an average level corresponding to a correctly phased 6 SC clock. Any change in the data bit rate of the received channel encoded data bit stream changes the position of the transition related pulses at the input to the current source 540 relative to the low to high level transitions of the 6 SC clock at the input to the current switch 544. If this occurs, one of the transistors of the current switch 544 will be enabled during the period that the current source 540 is enabled (by the transition related pulse) for a longer interval than the other transistor, with one of the transistors enabled for a longer interval depending upon whether the data bit rate increased or decreased. This causes a corresponding change in the current provided to the error smoothing capacitor 538 and a corresponding corrective change in the average voltage level on the capacitor. A change in the voltage level on the capacitor causes the voltage controlled oscillator 537 to change its phase and frequency until the transition related pulses are centered with respect to the low-to-high level change in the 6 SC clock provided to the current source 540. With the low to high level change in the 6 SC clock adjusted to be centered with respect to the duration of the transmission related pulses, the two halves, 544a and 544 b, of the current switch will individually pass current from the current source 540 for equal intervals. Hence, the average voltage on the capacitor 538 will be maintained at the level required to lock the frequency and phase of the 6 SC oscillator 537 to the data clock rate of the received channel encoded data.

If the 6 SC voltage controlled oscillator 537 fails to lock to the received data or data is not received by one of the decoder and time base correctors 100 included in one of the 10 bit lines of a playback channel, a frequency unlock signal is provided on an output line 550 that extends to the reference clock generator circuitry 98. All of the lines 550 from the 10 decoder and time base correctors of the playback channel are ORed in the reference clock generator circuitry 98 for coupling a frequency unlock command to the computer control system 92 in the event that one or more frequency unlock signals are generated in a playback channel. The computer control system 92 responds to the frequency unlock command by providing a video mute command to the blanking insertion and bit muting circuitry (FIGS. 41A and 41B) that blocks the sending of data to the requesting station. In the channel decoder 525, the frequency unlock signal is generated by detecting the failure of the channel decoder to provide a data bit for 16 cycles of 6SC. The frequency unlock signal is provided by a divide by two circuit 546 that has its clock input coupled to receive a clock pulse provided on line 548 each time the channel decoder 525 fails to detect a data bit for an interval of four cycles of 3SC, hence, 8 cycles of 6SC. If a second clock pulse appears on line 548 before the divide by the two circuit 546 is reset by the NAND gate 549, the divide by two circuit 546 provides the frequency unlock signal on line 550. The NAND gate 549 resets the divide by two circuit 546 each time a coincidence occurs between a low level of the 6SC clock provided by the oscillator 537 and a low level on line 530b, which occurs when a transition related data pulse is received at the input 526 of the channel decoder.

After the divide by two flip flop 531 converts the channel encoded data from the transition related pulse form to the channel encoded NRZ-L form, the data is coupled by line 531a to a pair of latches 551 and 552 (FIG. 37B) at the input of the decoding circuitry 525a. The decoding circuitry is able to decode data that is channel encoded according to the code rules of U.S. Pat. No. 3,108,261 (FIG. 37E (1)) and the U.S. Pat. No. 4,027,335 (FIG. 37E- (2)). The latches are clocked by .phi.1 and .phi.2 3SC clocks, respectively, derived from the 6SC clock generated by the oscillator 537.

The 6clock on line 534 is coupled to one input of each of the NAND gates 553a and 553b. The other input of each of the NAND gates receives complementary 3SC square waves generated by the divide by two flip flop 534a from the 6SC clock on line 534. The NAND gates are enabled when their inputs are low to issue the positive .phi.1 (FIG. 37E - (4)) clock pulses to clock the latch 552 and positive .phi.2 (FIG. 37E - (3)) clock pulses to clock the latch 551. The .phi.1 and .phi.2 clock pulses are displaced in time by one half cycle of 3SC. Hence, the time that the level of the channel encoded NRZ-L data on line 531a is latched by latch 551 is displaced one half cycle of 3SC from the time the level is latched by latch 552 (FIG. 37E - (5) and (6)). Both latches are coupled to the two inputs of an exclusive OR gate 554a. The exclusive OR gate serves to detect the occurrence of a change in state in the level of the channel encoded NRZ-L data at the input of latches 551 and 552 between the times they are clocked by the displaced .phi.1 and .phi.2 clocks (FIG. 37E - (7)). To determine if the change in state at the input of latches represented a logical 1 bit, the output of the exclusive OR gate 554a is coupled to one input of a NAND gate 555. The other input of the NAND gate receives inverted .phi.1 3SC clock pulses coupled from the NAND gate 553a by the inverter 555a. If the change in state at the input of the latches represents a logical 1 bit, the output of the exclusive OR gate 554a will be low at the occurrence of an inverted .phi.1 3SC clock pulse. The NAND gate 555 will be enabled, placing a high level on its output. To assure safe latching of the detected logical 1 bit pulse at the output of the NAND gate 555, a delay circuit 556 is connected to the input of the NAND gate 555 receiving the inverted .phi.1 clock so that the output of the NAND will be maintained high for an interval longer than the .phi.1 3SC clock pulse (FIG. 37E - (8)). This permits the following latch 557 to be clocked with the positive trailing edge of the .phi.1 3SC clock to latch the delayed high level provided by the NAND gate 555 (FIG. 37E - (9)). If the input data is channel encoded according to the U.S. Pat. No. 3,108,261 code rules, the output of latch 557 will be the channel decoded NRZ-L data. This is represented by the doted lines in the timing diagram shown by FIG. 37E. In the decoder shown by FIGS. 37A and 37B, however, an additional latch 558 is needed to permit decoding of data channel encoded according to the code rules of the aforementioned U.S. Pat. No. 4,027,335. However, for the U.S. Pat. No. 3,108,261 channel code, the additional latch 558 only delays the output of the decoded data by one cycle of 3SC.

When data is encoded according to the code rules of the U.S. Pat. No. 4,022,335, specified logical 1 bit related transitions are suppressed. If a logical 1 bit related transition has been suppressed, there will be an absence of data transitions for an interval greater than 11/2 cycles of 3SC. This is detected by a modulo-4 counter 559 having its clock input coupled to receive .phi.0 clock pulses provided by the NAND gate 553b and its reset input to the output of the edge detecting exclusive OR gate 554a. The exclusive OR gate 554a provides a reset pulse to clear the counter 559 each time a transition occurs in the channel encoded data (FIG. 37E - (10)). The output of the modulo-4 counter 559 is coupled to one input of an AND gate 560 which also receives .phi.0 clock pulses at its other input. Both inputs are low 1/2 cycle of 3SC after the modulo-4 counter has counter four .phi.1 3SC clock pulses without being reset, which corresponds to an absence of data transitions for an interval of 21/2 cycles of 3SC (FIG. 37E - (11), (12) and (13)). Ordinarily, this signifies that a logical 1 bit has bit suppressed in the channel encoded data. To make certain that no errors have been introduced to the data stream, a following NAND gate 561 examines an output of the latch 558 at the time when AND gate 560 provides the low state signal representing suppressed logical 1 bit. If the examined output of the latch 558 is also low, it verifies that a logical 1 bit has been suppressed and outputs pulses on line 562 (FIG. 37E - (14)) by the NAND gate 561 that is wire ORed with the output of latch 557. Line (14) of FIG. 37E repesents the state of NAND gate 561 as if it was not wire ORed with the output of latch 557. The second pulse 563 (FIG. 37E-(14) provided by the NAND gate 561 occurs at the time of and is latched into the latch 558 by the .phi.1 3SC clock. This prevents the output of the latch 558 from being returned low, thereby, inserting the suppressed logical 1 bit into the decoded NRZ-L data (FIG. 37E - (15)) appearing on line 566. In the data track bit line, the decoded data is coupled by line 566 to the computer control system 92. The decoded data clock provided by the flip flop 534a on line 574 and the line 10 or sync word from the first shift register and sync word detector circuitry 572 are also coupled to the computer control system 92.

If the phase of the 3SC decode clock provided by the flip flop 534a is incorrect, a one-shot multivibrator 534b is enabled by the coincidence of the 6SC clock on line 534 and a pulse provided on line 564. This pulse will be generated 3 cycles of 3SC before the line ID is first detected by sync word detector portion of the circuitry 572 if the level of the decoded data at that time is low, therefore, incorrect. A counter 590 (FIGS. 13A and 37C) receives 3SC decoded data clock and, as will be described hereinbelow, provides an advanced end of count pulse at H/2 rate, designated advanced EOC pulse on line 591. Because of the known data bit pattern of the sync word interval, which interval ordinarily occurs when the advanced end of count pulse is generated, the decoded data level can be examined at the shift register portion of the circuitry 572 to determine if decoding is performed correctly. The gating circuitry 592 issues a pulse on line 564 when the examined decoded data level is low that enables the one-shot 534b to provide a disabling signal at the clock input of the flip flop 534a for one cycle of 6SC. This results in a shift in the phases of the .phi.1 and .phi.2 clocks by 1/2 cycle of 3SC, thereby establishing the right phase for correct decoding of the channel encoded NRZ-L data.

During playback operations, each bit stream of channel decoded NRZ-L data provided at the output line 566 of the decoder circuitry 525 will contain time base errors in the form of bit time displacement errors as previously described. Furthermore, bit line to bit line or skew time displacement errors will be present in the 9 data bit streams that carry the 8 parallel bits of digitized video and 1 parity bit, if included. To remove these bit time displacement errors from the NRZ-L data, a time base corrector 565 is provided in each data bit stream and corrects such errors by electronically adjusting a variable delay through which the NRZ-L data is passed. Each time base corrector contains circuitry which processes the received data so that the data bit rates in all video data and parity bit lines are frequency and phase coherent with respect to the reference 3SC provided by the reference clock generator 98 for the playback channel 91. Furthermore, each of the time base correctors 565 also aligns the data bits in the data bit lines with respect to a common redefined H/2 reference provided by the playback channel's reference clock generator 98. As a result of these combined functions, any relative time displacement errors between the data bits in the 9 bit lines are removed, i.e., line to line or skew errors removed, and any bit time displacement errors within a bit line corrected. However, as described hereinbefore, the redefined H/2 signal, while being synchronized to a particular phase of SC and thereby facilitating processing of the reproduced video data, it is not stationary with respect to reference H sync. For this reason, use of the H/2 signal by the time base corrector 565 results in a mispositioning of the sync word in the video data that is output by the time base correct for alternate reproductions of the video data.

The operation of the time base corrector 565 included in each data bit line will be described in connection with the block diagram shown in FIG. 13A and the timing diagrams of FIGS. 13B and C Specific circuitry which can be used to carry out the operation of the time base corrector is shown in FIGS. 37B, 37C and 37D. The decoded data in each data bit line received from the decoder 525 over line 566 is time base corrected independently of the other 8 data bit lines by using a periodically occurring time reference common to all of the data bit lines and defined in terms of the frequency and phase of a higher rate clock used to encode the data. In video recording and reproducing apparatus such as described herein, horizontal line related H/2 signals derived from the periodically occurring sync words synchronously inserted in each data bit stream in the horizontal blanking interval as hereinbefore described is defined in terms of the frequency and phase of the higher rate (455 times H/2) signal color subcarrier component and the 3SC data clock (1365 times H/2) and is available for use as the periodically occurring timing reference.

To effect time base correction of the reproduced and channel decoded data, the data in all data bit lines are retimed to a common reference 3SC clock by directing the decoded data in each bit line through a phase 567. In the illustrated embodiment, a multiple port shift register 568 performs the retiming by having data written into addresses determined by the write address generator 569 clocked by the decoded 3SC data clock provided by the channel decoder 525 on line 574. The data is read out of the register 568 under the control of the read address generator 570 clocked by the reference 3SC clock provided on line 571 by the reference clock generator 98 (FIG. 8A). Because all of the phaser read address generators 570 in the 9 data bit lines are clocked by the same reference 3SC clock, the data in all of the data bit lines are retimed to the desired stable 3SC reference clock, which for an NTSC television signal standard is 10.7 MHz.

The write and read address generators 569 and 570 are preset and reset respectively to their starting addresses by the sync word included in the data being corrected, with the starting write address in advance of the starting read address by four addresses. Each time a sync word is detected in the received decoded data by the first shift register and sync word detector circuitry 572, a reset signal is provided and coupled to reset the read address generator 570. The decoded data on line 566 enters a seven bit shift register included in the circuitry 572 and is examined by logic circuits forming the sync word detector portion of the circuitry 572 for the occurrence of the 7-bit sync word pattern. After passage through the shift register, the data is clocked into the multiple port shift register 568. The register 568 has an 8 bit capacity and is initially operated to read an address four 3SC cycles following writing of data at the address. Because the write address generator 569 is clocked by the 3SC data clock and the read address generator 570 by the reference 3SC clock, data bit displacement errors in the received data will change the time an address has data written into it relative to the time the address is read. This change in the time between writing data at an address and reading data from the address results in the received data being retimed to the stable 3SC reference. Furthermore, the phaser 567 will properly retime the received data to the stable 3SC reference even if the sync word is not detected by the first sync word detector 572 as long as unanticipated large time displacement errors do not occuur that exceed the storage capacity of the register 568. Even if large time displacement errors occur, the video data emerging from the phaser 567 will be at the proper reference 3SC rate although incorrectly positioned in phase.

The sync word detector 572 provides a first input to the gating circuitry 592 (FIG. 37C) each time a sync word is detected in the decoded data. The seven bit shift register is clocked by the decoded data clock on line 574 to enter the decoded data received over line 566 for examination by the logic circuitry. The sync word detector 572 is enabled for sync word detection by the sync word enables pulse generator 600. This generator is enabled by a divide by 1364 counter 590 clocked by the 3 SC data clock on line 574. The generator 600 provides a sync word detection enable pulse on line 601 (FIG. 13B-(3)) which is initiated by the advanced EOC pulse (FIG. 13B-(2)) issued by the counter 590 issued over line 591 three counts in advance of the expected occurrence of a sync word at the first sync word detector circuitry 572 (FIG. 13B-(6)). This advanced EOC pulse also is coupled by line 591 to the gating circuitry 592 that responsively examines the output of the shift register to determine the data logic level and, hence, the phase of the decoded data clock. Upon the detection of a sync word by the second sync word detector 575 (FIG. 13B-(6)), a reset signal is issued over line 608 to the generator 600. The reset signal terminates the enable pulse on line 601 before the counter 590 reaches a count of fifteen. The counter position 15 in the counter 590 terminates the enable pulse if a sync word is not detected by the second sync word detector 575 (FIG. 13B-(6)). The shift register 604 provides the automatic EOC reset pulse to the counter 590 over line 610 upon the occurrence of the third 6 SC clock pulse following the advanced EOC reset pulse (FIG. 13C (2) and (5)). The shift register 604 and the pulse generator 605 cooperate to allow the sync word enable pulse to follow changes in the time of the occurrence of consecutive sync words in the amount of .+-. 1 cycle of 3SC. The pulse generator 605 simultaneously examines three outputs of the shift register 604 and generates a gating waveform (FIG. 13B-(4)) that prevents the sync word enable pulse from resetting the counter if it occurs within 1 clock time of the occurrence of the automatic EOC reset pulse generated by the shift register 604. If the reset enable pulse derived from a sync word arrives one count before the automatic EOC reset pulse, the counter 590 will not be reset (FIG. 13B-(4) and (8)). If the reset enable pulse is provided one count after the occurrence of the EOC reset pulse, the counter 590 will not be reset again (coincidence with the second positive pulse of the gating waveform provided by the pulse generator 605). If a sync word is not detected during the interval of the sync word enable pulse, the counter 590 will continuously reset itself through shift register 604 and line 610 (FIG. 13B-(5)) and, thereby, with generator 600 retain, as a memory, knowledge of when to provide sync word enable pulses until a sync word is detected as long as the detected sync word is not in coincidence with the positive gating waveform (FIG. 13B-(4)) provided by generator 605, NAND gate 612 is enabled to permit the sync word to be placed on line 613 for resetting the counter 590.

The vertical blanking signal on line 606 (FIG. 13B-(1)) is coupled to place the sync word enable pulse generator 600 in the enabled state for an interval of ten horizontal lines by disabling gate 611 and prevent the coupling of the count 15 position of the counter 590 to generator 600. This enables the time base corrector circuitry to lock onto the sync word detector 572 and 575 to be enabled at sync word time and set the phaser 568 and error gate 582 for proper operation.

The data is read from the multiple port shift register 568 with the 3SC reference clock into the shift register portion of the second sync word detector circuitry 575 (FIG. 37B). The shift register portion has three of its output lines 576 coupled to the data input of a serial to parallel converter 577. The multiplex clock provided on line 578 by the reference clock generator 98 is at the SC rate and latches the data in blocks of three data bit cells from the shift register portion of the circuitry 575 into the converter 577. Each cycle of SC the contents of the serial to parallel converter are transferred to a following RAM 579. The three output lines 580 of the converter 577 extend to the input of a RAM 579. The final time base correction is performed in RAM 579 whose write address generator 614 is clocked at reference SC, since the data rate at the input of the RAM is SC. The read address generator and latch and subtraction circuitry 623 and 615 is also clocked at reference SC to cause the reading of the RAM addresses. Read/write mode signals and write enable signals from the reference clock generator 98 of of FIGS. 35A-D control the reading and writing of the RAM addresses so that a read cycle occurs during one part of a subcarrier cycle and a write cycle at a different part of the cycle (refer to FIG. 11B) sync word than the record. The amount of the time displacement error required to be corrected is determined by the error gate 582. Upon the detection of the sync word by the second sync word detector 575, a signal placed on line 608 opens the error gate and allows reference 3SC clock pulses placed on line 571 by the reference clock generator 98 to pass to a divide by three counter 583. One output of the counter 583 extends to the read error address generator 623 to provide SC rate clock pulses to the generator. When the reference H/2 is received on line 581 from the reference clock generator 98, the error gate 582 is closed, delayed a terminating the coupling of reference 3SC clock pulses to the counter 583. Consequently the SC rate clock pulses are no longer provided to the read error address generator 623 and the number being provided at such time represents the time displacement between the video signal's sync word and reference H/2 in a whole number cycles of SC. Also, a delayed pulse is generated by the delay and pulse former 621 in response to the closure of the error gate 582. The delayed pulse is coupled to the read error address, generator 623 and latches the error count in the read error address generator 623. Subsequently, a reset pulse is generated from the latch pulse to reset the .div. 3 binary counter 583 and read error address generator 623. The counter sets the read address in accordance with the timing difference between reference H/2 and the sync word detected by the second sync word detector 575 measured in cycles of 3 SC divided by three. The measured value of the timing difference is coupled to a latch and subtractor 624 and is subtracted from the write address to generate the correct read address. Because the clocks representing error are divide-by-three, the RAM 579 will adjust for errors of integral numbers of subcarrier cycles. A 3-bit shift register 617, error latch 618 and gates 619 provide correction in fractions of one cycle of 3SC of any residual error remaining after the data has passed through the RAM 579. The parallel to serial converter 620 at the output of the RAM 579 receives the demultiplex clock from the reference clock generator 98 and converts the data rate back to 3 SC at the input of the shift register 617. FIG. 13C shows the typical correction performed by the phaser 567 and following time base correction by the RAM 579 and shift register 617. The corrected output of the time base corrector 565 appears at terminal 622. However, the use of the reference H/2 signal, which is redefined with respect to a particular phase of subcarrier, in measuring the time displacement error through the operation of the error through the operation of the error gate 582 results in the 42 nsec 15Hz jitter in the video signal provided by the time base corrector 565.

The 9-bit parallel output of the time base corrector 565 is coupled to the data transfer circuitry 129. The data transfer circuitry also clocks the data received at the input to the output is clocked using a 3SC PALE clock to reposition the samples into the desired vertically aligned positions that was achieved by the original PALEing during sampling in the analog-to-digital converter 95. When the signal was channel encoded, the alignment was changed due to the fact that a line to line continuous phase 3SC clock was used to channel encode the NRZ data. The data emerging from the time base corrector circuitry 565 is thus aligned the same way as the encoded data at the output of the encoder 96. Accordingly, the data transfer circuitry 129 again PALEs the data to realign the samples in the manner as shown in FIGS. 8C(10) and 8C(11).

Referring to the block diagram of the data transfer circuitry 129 shown in FIG. 14, time base corrected data provided by the decoder and time base corrector circuitry 100 over nine bit lines i.e., eight bit lines containing video information and one parity line, are applied at nine input lines of the data transfer circuitry. The line 625 in FIG. 14 represents the most significant bit line and is representative of each of the nine input lines provided for each bit stream. The data is clocked into FF 626 and FF 627 using a 3SC PALE clock signal which appears on lines 628 629. The PALE clock is generated by a PALE clock generator shown at the lower portion of the block diagram from 6SC and 1/2SC signals received from the reference clock generator 98 on lines 630 and 631, respectively, and a PALE flag signal received from the reference logic circuitry 125B via the encoder switch 126 on line 632. The PALE flag signal is applied through an inverter 633 and line 634 to one input of an AND gate 635. The line 634 also connects to a second inverter 636 which extends to one input of another AND gate 637 via line 638. The 1/2SC signal on line 631 passes through a pulse former 639 and clocks a divide by 2 FF 640 which produces 3SC output signals of opposite phases on output lines 641 and 642 which respectively extend to the other inputs of the AND gates 635 and 637. The outputs of the AND gates are connected to line 643 and extend to complementary dual output buffer 645 which clocks FF 626 and FF 627. The PALE flag signal on line 632 is a two state or level signal that changes state at an H/2 rate and, upon changing levels, alternately disables AND gate 635 and enables the ANd gate 637 to gate one of the 3SC signals from lines 641 and 642 onto the output line 643. Thus, in effect the PALE flag alternately changes the phase of the 3SC signal that is used to clock the data on line 625 through the FF 626 and FF 627 so that consecutive horizontal lines of video data are clocked with opposite phased 3SC signals. This retimes the video data bits from the continuous phase clock back to the PALE clock so that the vertical alignment of samples of consecutive lines is re-established for subsequent chroma separation and processing. As previously described, the video data bits are not to be retimed in the transfer mode of operation. To prevent the retiming, the encoder switch 126 blocks the coupling of the PALE flag from the reference logic circuitry 125B to the data transfer circuitry 129 and instead places a low level signal on line 632. This places an enabling signal on the input of the AND gate 635 and a disabling signal on the input of the AND gate 637, where a line to line continuous phase 3SC clock signal is provided on line 643 through the AND gate 635.

The data on the output of FF 627 extends to an AND gate 647 via line 648 and AND gate 647 has output line 649 connected to the first of three FFs 651, 652 and 653 which serve to shift the serial bits to the output of the last FF which appears on line 654. Line 654 also extends to one input of another AND gate 655. A parity tree error detecting circuit 656 is coupled to receive the data bits of the nine bit streams as described below and has two output lines 657 and 658 which extend to AND gates 655 and 647, respectively. When an error is detected, it disables AND gate 647 to block the bit containing the error and enables AND gate 655 so that the output data on line 654 can be clocked through AND gate 655 onto line 649. This has the effect of replacing the incorrect bit with the third previously occurring bit in the data stream and effectively masks the error with the bit that is approximately correct for the reasons that had been previously discussed.

Five bits, i.e., bits 2 through 6, or the next most significant bit through the sixth most significant bit are also sampled through a resistor ladder network 659 having weighted values to produce an analog conversion of the digital information which approximates digitally encoded analog information and is used to detect if chroma phase needs to be inverted. The output on line 660 extends to the reference clock generator 98 and is compared with the phase of the burst of the station reference video signal to determine if the chroma phase needs to be inverted. The digital-to-analog conversion occurring in the data transfer circuitry 129 is gated to reject all but the burst and produces an imprecise, but sufficiently accurate determination of the burst phase for use by the reference clock generator 98.

A television picture with a region of saturated color bounded along the bottom by a region of no color defines a vertical color transition along the horizontal boundary or color edge. Given three successive television lines A, B and C of a field, wherein the lines are within the saturated color region immediately above the color edge, a conventional comb filter generates the vectors representing chrominance in accordance with the relationship, - 1/4A + 1/2B - 1/4C.

However, the color subcarrier of an NTSC television signal has a 180.degree. phase shift between alternate lines A, B and C. Thus 180.degree. inversion of, for example, lines A and C and subsequent summation of the vectors +1/4A +1/2 +1/4C generates a full chrominance vector, herein termed 1B or simply, +B, the chrominance on line B. When this chrominance vector +B is subtracted from the wideband signal (which also contains the chrominance vector +B), the chrominance vectors cancel. The comb filter has effected complete chrominance and luminance separation, i.e., all chrominance is in the chrominance channel.

However, in a second case, if lines A and B are in the saturated color region, with line C in the region of no color, line A provides a chrominance vector equal to B in the negative direction and line B a vector equal to B in the positive direction. But line C provides a zero chrominance vector since it lies in a region of no color. When combining the vectors in accordance with the previous relationship, -1/4 of vector A is inverted and added to +1/2 of vector B, thereby providing a sum of +3/4 of a full vector B. It follows that when the chrominance +3/4B is subtracted from the wideband signal, i.e., line B, there is a residual of +1/4 of the chrominance vector left in the luminance channel, while only +3/4 of the chrominance vector is extracted into the chrominance channel.

A third case exists wherein only line A is within the saturated color region, and lines B and C are in the region of no color. The third case is similar to the second case above, wherein however, the signs are opposite.

The consequence of the second (and third) case given above, wherein line C (or B and C) lies in a region of no color, prove disadvantageous when attempting to reconstitute a composite NTSC color television signal from a signal stored color field, or picture frame. As is well known, when reproducing the composite video signal from a single stored picture frame, in one picture frame the chrominance is added directly back to the luminance previously separated therefrom, whereas in the second picture frame the chrominance components is first inverted and then is added to the luminance. Therefore, in the second case mentioned above wherein line C is in a region of no color, in the non-inverting frame, the +1/4 chrominance vector which remained in the luminance channel due to incomplete separation, is added to the +3/4 chrominance vector separated into the chrominance channel. Thus the full vector B, i.e., the full chrominance signal, is recovered to define a correctly reconstituted color television signal for the non-inverted picture frame. However, when reconstituting the second frame of color video from the single stored picture frame, the chrominance (+3/4B) is first inverted, providing a +3/4 chrominance vector, when when subseqently added to the +1/4 vector in the luminance channel provides only a -1/2 chrominance vector for the inverted frame. Thus, in the non-inverted frame, the chrominance is reproduced with full saturation, whereas in the alternate, inverted frame the chrominance is reproduced at 1/2 saturation. Thus the color saturation defining the color edge between the region of full color and that of no color will flicker at a 15 Hz rate between 1/2 saturation and full saturation. This visible flicker is objectionable when reproducing the composite NTSC four-field color coded television signal.

The chrominance separating and processing system provides various embodiments of digital circuits which perform the inversion process digitally in combination with a digital comb filter and digital bandpass filter, while further providing a conditioned chrominance signal which, when digitally re-combined to form the composite NTSC color television signal, minimizes or cancels completely the objectionable 15 Hz flicker at the vertical transitions.

Although the combination is hereinafter particularly described utilizing a three times subcarrier (10.7) megaHertz phase alternating line encoding (PALE) sampling technique with a PCM encoded NTSC video signal, it is to be understood that other encoding techniques, sampling techniques, frequencies, etc., may be employed. Furthermore, the single lines depicting the inputs and outputs of the block diagram components are representative of digital words of selected numbers of bits, as exemplified in the detailed schematics of FIGS. 38, 39 and 40.

Referring to FIG. 15, there is shown a digital chrominance separating and processing system wherein a 10.7 megaHertz (MHz) PALE PCM color video signal is introduced via input terminal 700 to digital comb filter means 701. The filter means 701 is per se generally typical of digital comb filters presently utilized in various television signal processing systems, but herein is adapted via a specific clocking technique further described hereinbelow, to separate the chrominance from the digital wideband color signal. The outputs from filter means 701 and the associated clocking techniques, include a 1H delayed wideband signal (delayed by one-horizontal-line delay period) on line (terminal) 702, and an extracted chrominance signal (with low frequency components still included) on line (terminal) 703a. The term "extracted" is herein used to define the chrominance signal which is separated into a chrominance channel, whether the separation is complete, or incomplete, as previously described with respect to case two (and three) hereinabove.

The extracted chrominance signal is fed to bandpass filter means 704 which removes vertical resolution losses due to the comb filter means, by passing only that frequency band occupied by the chrominance information. The bandpass filter means 704 is centered at 3.58 MHz (the NTSC subcarrier frequency) and has a bandwidth of, for example, 1.5 MHz.

The resulting combed chrominance signal is fed via line (terminal) 703b to a digital circuit for inverting the chrominance signal on alternate frames at the frame rate. In FIG. 1 the inverting circuit comprises a digital transversal filter with odd symmetry 705, which herein may be further identified as a modified, digital, "Hilbert" transformer. It is understood that the transversal filter 705 provides one form of inversion; i.e., employs what is basically known as the Hilbert transform, but which is herein modified to provide a specific form of transversal filter with odd symmetry, while further providing a digital rather than analog inversion implementation. The transversal filter of interest has the property of rotating th phase 90.degree. of all frequencies of a selected range which herein, for example, may be two to four MHz.

Thus the term inversion, or inverting is employed to define the circuitry and process of digitally conditioning the chrominance at the frame rate (or field rate if one field is used to reconstitute the four field color coded NTSC color television signal) as by phase shifting, rotating, inverting or otherwise handling the phase. Further, the successive playbacks of either a single stored field of picture frame is referred to generically as "alternate repretitive reproductions".

The chrominance signal is also fed to a negative input of digital adder (subtractor) means 706. The 1H delayed wideband video signal of terminal 702 is fed to the positive input of the adder means 706. The transversal filter 705 includes a control input at 707 which determines the conditioning of the chrominance signal phase. In one embodiment, for example, the transversal filter may provide a plus and then a minus 90.degree. phase rotation of the chrominance with respect to the luminance signal in alternate repetitive reproductions. The chrominance and luminance signals are then summed in digital adder means 708 to provide the composite color television signal on output terminal 728.

Control means 709 includes various timing and clock inputs thereto which, for example, relate to the overall apparatus timing and thus originate upstream in the apparatus. In turn, the control means 709 generates specific control signals for the comb filter means 701, for the transversal filter control input 707, for the bandpass filter means 704, etc., which control signals include a PALE clock and 1H delay line, four-phase clocks, inter alia. The control means 709 and the various inputs and outputs are further shown and described in detail in FIGS. 38A, 38B, 39A, B and C and thus are not further described here.

Briefly, in FIG. 15, the comb filter means 701 combines the three adjacent television lines A, B, C of previous mention, and includes a pair of digital, one-horizontal-line (1H) delay lines 710, 711 and a pair of adder means 712, 713. The 10.7 MHz PALE video signal is fed to delay line 710 as well as to adder means 712. The 1H delayed signal is fed to 1H delay means 711, and to the adder means 713. The 2H delayed signal is fed to the other input of adder means 712, whose output in turn is fed to the negative input of the (subtractor) adder means 713.

The digital comb filter means 701 and digital bandpass filter means 704 exemplified in block diagram herein, generate (eight bit) digital words corresponding to the separated chrominance and 1H delayed wideband signals, and are further depicted in the schematic diagrams of FIGS. 38A-B and 40A-B.

The combed chrominance signal is subtracted from the 1H delayed wideband video signal via the digital adder means 706, wherein the resulting combed luminance signal is fed to the digital adder means 708.

FIG. 16 shows the digital transversal filter 705 wherein the digital combed chrominance signal is fed to a series of one-sample-period delays 714a-714c, and also to the positive input of an adder means 715b. The negative input of adder means 715b is coupled to the output of the last delay 714c. The positive and negative inputs of an adder means 715a are coupled to the input and output respectively of the delay 714b. The outputs of adder means 715a, 715b are coupled to respective multiplier programmed read-only memories (PROMs) 716a, 716b, and thence to adder means 717. The latter is coupled via an inverter stage 718 to the adder means 708 of previous mention, along with the combed luminance signal from adder means 706, whereby means 708 generates the composite color television signal. The control input 707 is coupled to the inverter stage 718.

In operation, the transversal filter 705 provides digital circuits for conditioning the phase of the chrominance signal with respect to the luminance signal; i.e., for providing the digital implementation of phase inversion of the chrominance on alternate picture frames. To this end, the 1H delayed wideband signal, and the chrominance signal are introduced to the adder means 706 via terminals 702, 703b respectively, whereupon the resulting luminance signal is introduced to adder means 708. The chrominance signal is delayed one-sample period (e.g., 93 nanoseconds) in each of the delays 714a-714c, whereby the undelayed chrominance and the three-sample delayed chrominance are introduced to the adder means 715b, and the one-sample and two-sample delayed chrominance signals are introduced to adder means 715a. The delays 714a-714c may comprise a single stage of a shift register. The adder means 715a, 715b provide signals to multiplier PROMs 716a, 716b respectively, which perform a multiplication of the respective signals by 0.575 and 0.096 in a digital approximation of a conventional convolution operation. The resulting signals are then summed via adder means 717, and the summed signal has all of its frequency components advanced 90.degree. with respect to the luminance signal, to define the conditioned chrominance signal of previous mention. The output of adder means 717 is delivered to the adder means 708 via the inverter stage 718. During one color frame the inverter means 718 has a high, or "1", introduced thereto via the control input 707 from control means 709, whereby the (eight) bits of the output word are delivered unchanged to the adder means 708. On alternate video picture frames the control input 707 is a low (or "0") invert enable signal (see FIGS. 39A-F). Data is represented in this device in the signed two's complement negative system where negative numbers have a "1" in the sign bit position and the magnitude is the 2's complement of its absolute value. Therefore, inversion amounts to changing the sign and forming the 2's complement, via the "0" invert enable input 707. Thus the conditioned chrominance signal (which is rotated +90.degree.) is directed added to the luminance in one frame, and is inverted and then added to luminance in the alternate frame, to provide the composite color television signal on output terminal 728. Alternately, the chroma first may be rotated -90.degree. on each frame by reversing the inputs to the adder means 715a, b and then adding directly in one frame and inverting 180.degree. and adding in the next.

In another alternative, the transversal filter 705 may be implemented whereby during one picture frame the phase of the chrominance signal is advanced by 90.degree., and during the alternate picture frame is retarded by 90.degree., to provide in essence the 180.degree. inversion of the frequency components between frames.

FIGS. 38A-C, 40A-B and 39A-B illustrate one schematic implementation of the embodiment of FIGS. 15 and 16 utilizing the digital transversal filter with odd symmetry 705. FIGS. 38A-C illustrate one implementation of the digital comb filter means 701, and part of the control means 709 of FIG. 15; FIGS. 40A-B illustrate one implementation of the digital bandpass filter 704; and FIGS. 38A-B illustrate one implementation of the digital transversal filter 705, signal re-combining adder means 706, 708, and the remaining circuits of the control means 709. In all figures, like components of FIGS. 15 and 16 are indicated by similar numerals.

Thus, in FIG. 38A the 10.7 MHz PALE video signal is introduced via the input terminal 700 to the digital comb filter means 701. The output thereof (FIG. 38C) comprises the separated chrominance and 1H delayed wideband signals on terminals 703a and 702 respectively. The inputs at terminal 719, 725 comprise a group A and B control signals and a symmetrical PALE clock, generated in the respective portion of control means 709 of FIG. 39B further described below. The terminals 719, 725 are coupled to a four-phase clock generator 720 of control means 709 depicted in FIG. 38A. The clock generator 720 forms part of the timing circuits for clocking the shift registers which comprise the 1H digital delay lines 710, 711. The delay lines 710, 711, adder means 712, 713 and terminals 702, 703a, are interconnected via integral latching circuits 712a, 713a and 721 which conventionally temporarily store the respective digital products of the preceding shift registers, adders, etc. Terminal 703a provides the input to the succeeding digital bandpass filter means 704 of FIGS. 40A-B, and the terminal 702 provides the input to the adder means 706 of succeeding FIG. 39B.

The delay lines 710, 711 further each include a series of two-phase shift registers 750, 751 respectively, employing two-phase clocks, wherein the register stages are further arranged into groups 750A, 750B of delay line 710, and 715A, 751B of delay line 711. Shift register stage selectors 752A, 752B select portions of the digital word corresponding to specific clock phases of groups A, B of delay line 710, and shift register stage selectors 753A, 753B do the same for delay line 711. Wideband signal selectors 754, 755 of delay lines 710, 711 respectively, then provide selection of the digital words corresponding to the 1H and 2H delayed wideband signals, respectively.

The wideband video signal word is split, and is clocked into four bit stages of the shift registers 750A, 750B by the four-phase clocks, which are in effect, four phases of the symmetrical PALE clock. The stage selector 752A receives and loads the pairs of four bits in response to PALE clock, alternately from different pairs of stages of shift register 750A. Stage selector 752B does the same with shift register 750B stages. The group A stage selectors 752A unload into one (four bit) wideband signal selector 754, while the group B stage selectors 752B unload into the other (four bit) selector 754, in response to timed PALE clocks respectively. At selected times, the group B selectors are clocked whereby the combined group A and B registers provide a total of 680 bits per television line. One NTSC horizontal television line sampled at three times subcarrier rate will contain 6821/2 samples. However, as will be described in more detail hereinbelow, the clocks for the shift registers are generated and applied to the registers so that the total bits per television line output by the register for each bit line is equal to an integral number of samples. In the embodiments described herein, 680 samples per television line are clocked through the registers. The clocking of the registers is arranged so that the discarded interval of 21/2 sample intervals occurs outside the active video information portion of the television line during the horizontal blanking interval.

The control circuits 720 of FIG. 38A, which provide the four-phase clocks for the shift registers 750A, 750B and 751A, 751B, and which receive a symmetrical PALE clock, are further described in operation in the block diagram and clock waveforms of the combined control means 709 in FIGS. 39C-D infra, with one implementation thereof illustrated in the schematic diagrams of FIGS. 38A, 39A-B.

FIGS. 40A-B depict the bandpass filter means 704 with terminal 703a providing the incoming extracted chrominance signal from the comb signal 701 output, FIG. 38B. The combed chrominance signal from the bandpass filter means 704 is provided at terminal 703b of FIG. 40B, which forms the input to the transversal filter with odd symmetry 705 of succeeding FIGS. 39A-B. Immediately preceding the terminal 703b is an adder/latch stage 756, wherein the latches are clocked by a chroma invert enable signal via a terminal 757. In the embodiment employing the transversal filter 705 (FIGS. 15, 16, 39), the chroma invert enable signal does not enable the clear input of the latches, and the signal into the adder/latch stage 756 appears at the terminal 703b. The PALE clock of terminal 725 couples to various inverters (FIG. 40B) to provide a plurality of clocks for the adders and latches that comprise the bandpass filter means 704. The latches are thus clocked by the PALE clock to deliver the digital output from the preceding logic processor component (viz, the adders) to the succeeding logical processor components (also adders).

The final adder/latch stage 756 of the bandpass filter means 704 delivers the combed chrominance signal.

One-horizontal-line delay lines are required to provide the comb filtering process of chrominance signal separation from a wideband signal. The delay lines, and thus the comb filter 701, must be in synchronism with the overall system timing, which inter alia is represented by the input termed PALE flag. As discussed herein with reference to the video signal system of FIG. 8A and the reference logic circuit 125B of FIG. 10A in particular, the PALE flag signal is asymmetrical, i.e., has one phase for a longer time period while the alternate phase has a shorter time period, and the phase of the PALE clock changes coherently with the asymmetrical PALE flag. However, the PALE clock utilized by the instant chrominance separating and processing circuit utilizes a symmetrical PALE clock, i.e., one in which clock has alternate phases for the same time duration.

A paramount problem when attempting to reconstitute the composite color television signal from a single stored color field or frame, stems from the fact that each line of a field is of a duration equal to 2271/2 cycles of subcarrier f.sub.SC. That is, it is equal to an integral number of cycles plus one-half cycle of subcarrier time. It follows that a required condition of 1H delay lines, when they are formed of digital shift registers such as, for example, those in the comb filter means 701, is that there is an integral number of samples per line of television and thus one horizontal line of delay.

Accordingly, the present invention provides the control means 709 which inter alia generates the symmetrical PALE clock from the overall apparatus asymmetrical PALE flag, and which, during the horizontal blanking period deletes an intergral number plus one-half of subcarrier cycles, to shift by 180.degree. with respect to previous samples at the line rate. The PALE clock thus is in the proper phase relationship with the subcarrier frequency as required to reconstitute the four fields required to color encode the television signal, while also being in proper timing relationship with the overall apparatus.

Accordingly, FIG. 39C depicts in block diagram form the digital control means 709 shown in one schematic implementation in FIGS. 38A-B and 39A-B. FIG. 39D is a timing diagram of the waveforms generated at various points along the circuit of FIG. 39C, as well as FIGS. 38A-B and 39A-B. Inputs from the overall system include the asymmetrical PALE provided by the reference logic circuitry 125B, a six times phase continuous subcarrier frequency (6f.sub.SC) and a one-half times phase continuous subcarrier frequency (1/2f.sub.SC) provided by the reference clock generator circuitry 98 and a field index pulse provided by the reference input circuit 93B, on respective terminals 758, 759, 760 and 761. The signals are introduced to a PALE clock generator generally indicated at 762, which in turn is coupled to the four-phase clock generator 720 of that portion of control means 709 in FIG. 38A. The latter provide the four-phase clocking of the shift registers 750A-B and 751A-B, as further described below.

The PALE clock generator 762 receives the PALE flag via terminal 758, and feeds it to an exclusive OR 763. The latter is coupled to a D-type flip-flop 764, together with the 1/2 f.sub.SC clock from terminal 760. The exclusive OR 763 and flip-flop 764 define a gated phase detector. A D-type flip-flop 765 is coupled to flip-flop 764 and is clocked by a correction pulse on line 766 corresponding to the group A control signal provided by a count decoder 772, further described infra. A JK-type flip-flop 767 is coupled at pin K thereof to flip-flop 765, and is clocked by the 6 f.sub.SC clock from terminal 759. The flip-flop 767 is coupled to an AND gate 768 and back to the clear pin of the flip-flop 765. The flip-flops 765, 767 and the AND gate 768 together define a gated phase corrector. AND gate 768 also received the 6 f.sub.SC clock, and is coupled in turn to a divide-by-two (.div. 2) JK-type flip-flop 769 and to a divide-by-1365 (.div. 1365) counter 770. The .div. 1365 counter 770 receives the field index pulse from terminal 761, and is coupled to the .div. 2 flip-flop 769 via a reset pulse generator means 771. As shown in FIG. 39B, the field index pulse first is reclocked to inverted 2 f.sub.SC via a flip-flop stage. The counter 770 is also coupled to a count decoder 772 which generates the group A and B control signals on terminal 719. The group A control signal defines the correction pulse 766 which clocks the flip-flop 765. The output of the .div. 2 flip-flop 769 comprises the symmetrical PALE clock which is fed back to the second input of the exclusive OR 763 to define a closed loop in the PALE clock generator 762. The PALE clock is also fed via terminal 725 to the four-phase clock generator 720 of FIGS. 38A-B and 49C, which as shown, only generates group A four-phase clocks.

In operation, referring to FIGS. 39C and 39D, when the chrominance separating and processing system is turned on, the counter 770 is not properly set and accordingly is reset via the reclocked field index pulse. The latter is a 30 Hz pulse which occurs on a selected field wherein sync pulses coincide with vertical interval. After reset, the PALE clock generator starts generating an initial PALE clock which resembles true PALE clock. However, the PALE clock must be in phase with the apparatus PALE flag, during the active part of a television line. That is, when PALE flag is up, the rising edge of 1/2 f.sub.SC is supposed to coincide with a rising edge of PALE clock, and vice versa. To this end, the (initial) PALE clock, which may resemble the waveform of either FIG. 39D - 17 or 18 when the circuit is turned on, is fed back to the exclusive OR 763 together with the PALE flag. When PALE flag is high, the exclusive OR output is low when PALE clock is low. When PALE flag is low, the exclusive OR output is low when PALE clock is high. Thus the PALE clock is de-PALEd to provide 3 f.sub.SC which is fed to the flip-flop 764 together with 1/2 f.sub.SC. The flip-flop 764 compares the de-PALEd signal and the 1/2 f.sub.SC signal (waveforms 39D - 16, 17, and 18). If flip-flop 764 takes the data the PALE clock is not in phase with the PALE flag, and vice versa. Thus the exclusive OR and the flip-flop 764 provide the gated phase detection.

If the PALE clock is not in proper phase, the gated phase corrector formed of flip-flops 765, 767 and AND gate 768, deletes one cycle of the 6 f.sub.SC clock to shift the phase by 180.degree. and bring PALE clock into the proper phase relative to PALE flag. The correction pulse 766 delays the time that the detection and correction is made, i.e., during the active part of the television line where it is known that the phase is the same. Because PALEing of the sampling clock used in the video signal system does not occur during the horizontal blanking interval as described hereinbefore with reference to FIGS. 8A-C and 10A,B, detection of the proper phase of the symmetrical PALE clock cannot occur during the horizontal interval. However, once the proper symmetrical PALE clock phase is detected, the PALE clock phase thereafter changes during the horizontal blanking interval in the chrominance separator and processing circuitry 101.

The counter 770 counts down 1365 counts of 6 f.sub.SC (FIG. 39D - 1) corresponding to one television line, and delivers a carryout (FIG. 39D - 3) to the reset pulse generator 771 on a rising edge of 2 f.sub.SC (FIG. 39D - 2). The latter includes a series of D-type flip-flops which provide six counts after carryout goes low, and thus the succession of highs depicted in FIG. 39D - 4 through 8. The inverse output signals corresponding to the waveforms of FIG. 39D - 6, 8 provide the start and end of a low state to the .div. 2 JK-type flip-flop 769 (FIG. 39D - 9), which in turn generates the symmetrical PALE clock at 3 f.sub.SC (FIG. 39D - 10) which appears at terminal 725.

As may be seen by comparing FIG. 39D - 10, 11, the phase of the PALE clock is shifted by 180.degree. by deleting 21/2 cycles of the phase continuous 3 f.sub.SC signal. To this end, after the input to the .div. 2 flip-flop 769 goes low, the rising edge of the PALE clock corresponding to the next rising edge of 6 f.sub.SC stays low, as do the two following rising edges of PALE clock. On the following rising edge of 6 f.sub.SC after the input to flip-flop 769 goes high, the PALE clock goes high, but with 180.degree. phase shift relative to its phase during the prior line (FIG. 39D - 11), thus the requirement of deleting the 1/2 cycle of subcarrier each television line is accomplished.

The count decoder 772 is coupled to the counter 770 and generates the group A and B control signals after a selected count, the signals being introduced via terminal 719 to the four-phase clock generator 720. The group A control signal is also fed to the gated phase on line corrector as a pulse 766 as previously mentioned.

The four-phase clock generator 720 provides for selected timing control of the comb filter shift registers 750A-B and 751-B, whereby the outputs thereof fulfull the requirement of generating an integral number of samples per television line, e.g., 680, utilizing the symmetrical PALE 3f.sub.SC sample clock. This circumevents a further problem caused by the integral number of subcarrier cycles plus one-half cycle per line, wherein the one-half cycle prevents proper sampling from line-to-line and must be deleted, or otherwise compensated for. To this end, the four-phase clock generator 720 includes a divide-by-four (.div. 4) binary counter 773 coupled to the PALE clock via terminal 725, and thence to a one-out-of-four binary decoder 774, and to the shift register stage selector 752A (and selector 753A) of previous mention in FIGS. 38A-B. The binary decoder 774 data input is connected to a high, wherein the selected output equals a low, and the unselected outputs equal highs. The shift register stage selectors 752A and 752B are coupled to the wideband selector 754 (FIG. 38) which selects digital words from shift register group A or B in response to the group A and B control signals from the count decoder 772. Binary decoder 774 is coupled to a latch 775 and thence to four D-type flip-flops 776a-d. The latch 775, whose output follows its input, is coupled to PALE clock and flips-flops 776a-d are also coupled thereto via an inverter 777. The four phase clocks are generated on outputs .phi.1, .phi.2, .phi.3 and .phi.4 of the flip-flops 776a-d via inverter staes, and are shown in FIG. 39D- 12 through 15. The clocks .phi.1, .phi.2, .phi.3 and .phi.4 are introduced to the shift registers 750A (and 750B) of 1H delay line 710, as well as to shift registers 751A and (751B) of comb filter 711 (FIGS. 38A-B). The video input signal is introduced to the shift registers at terminal 700.

In operation, the overlapping four-phase clocks .phi.1- .phi.4 (of the order of 150 nanoseconds) are applied to the multi-stage, two-phase shift registers 750A (750B) to clock successive four bit pairs into alternate stages to provide the clocking rate required, which rate the shift registers could not handle without employing the four-phase clocking into alternate stages. Note that the four-phase clocks are disabled, FIG. 39D- 12-15, during the 21/2 cycles of PALE clock FIG. 39D- 10, to provide the exact 1H delay. In addition, since shift registers having a capacity of 512 bits are readily available, they are employed to provide the 680 bits corresponding to one-horizontal-line delay.

Although only the group A shift registers 750A and timing controls therefor, of only the 1H delay line 710, are shown in FIG. 39C, it is understood that the PALE clock line 725 and group B control signal 719 also are introduced to the group B shift registers of the 1H delay line 710 (FIG. 38). Furthermore, the 1H delay line 711 (FIG. 38C) is identical to the 1H delay line 710 and similarly employs the PALE clock and group A and B control signals.

FIGS. 39A-B depict one digital implementation of the control means 709 of FIG. 39C, and also of the transversal filter with odd symmetry 705 of FIG. 16, the latter including terminals 703b and 702 for receiving the combed chrominance and 1H delayed wideband signals, respectively.

The various components 714-718 of the filter 705 are shown in schematic, and define means for rotating the phase of the chrominance signal +90.degree., whereupon inverter means 718 inverts the signal 180.degree. in response to the control input 707. A-90.degree. rotation may be generated by corresponding sign changes, i.e., by clocking the latches of adders 715a, 715b, with inputs that are opposite in sign to those shown in FIG. 16. Inverter mans 718 is defined herein as a plurality of exclusive ORs which essentially perform the 180.degree. inversion.

The bandpass filter inherently has a gain of 27/32, hence the gain of the wideband signal must match this gain. Therefore, in FIGS. 39A-B, the 1H delayed wideband signal is coupled to a 27/32 multiplier PROM 722 which multiplies the wideband signal by 27/32, to provide an overall gain of unity. The wideband signal is then fed through an (eight stage) delay 723, which equalizes the delays in the wideband channel with the delays in the chrominance channel caused by the bandpass filter means 704, and thence to the adder means 706. Various latches 724 are provided between the adder means 706 and 708, which provide a temporary store of the intermediate signal while clocking the luminance signal from adder means 706. The composite color television signal is provided an output terminal 728 via the adder means 708 of FIG. 39B, by combining alternate repetitive reproductions of the stored video signal.

The block diagram of the PALE clock generator 762 of FIG. 39C is shown in schematic in FIGS. 39A-B while the four-phase clock generator 720 of FIG. 39C is shown in schematic in FIGS. 38A-B. Since the operation of the generators 762 and 720 were described in FIG. 39C, no further explanation is required in the schematic diagrams of FIGS. 39A-B, wherein like components are identified by similar numerals.

However, in addition, FIG. 39A includes a terminal 778 for receiving a chroma switch and a frame switch input, which are provided by the computer control system 92 via the blanking insertion and bit muting circuitry 127 and the reference clock generator 94, respectively. The frame switch input is a chrominance inversion enable signal which determines the color frame which is to be inverted and that which is not. Thus the frame switch input generates the control input 707 to the transversal filter 705 in the form of the chroma invert enable signal, as further described below, which is the same chroma invert enable signal which is fed to the adder/latch stage 756 (FIG. 40B) on the terminal 757 of FIGS. 39B and 40A. As previously described in FIG. 16, the chroma invert enable is high during one picture frame to pass the input unchanged through the inverting exclusive ORs 718. In the alternate frame, the invert enable is low to change the sign and form the 2's complement to thus invert the chrominance. The chroma switch input of terminal 778 couples to the frame switch input via AND gate 779 and prevents the frame switch signal from enabling inversions when the apparatus is not receiving signals from the (disc/tape) storage, e.g., when the apparatus is in electronic-s-to-electronics mode and chrominance inversion is not desired.

Referring still to FIG. 39A-B, the PALE clock generator also provides the PALE clock on lines 781, 782 via the inverse pin of the .div. 2 JK flip-flop 769 and inverters 780. The PALE clock is used conventionally to clock the various latches associated with the adder means 715a, b, the multipler PROMs 716a, b, the one-sample delay lines 714a, b, c and the delay 723.

Referring now to FIG. 17, there is shown an alternative embodiment of the chrominance separating and processing system, wherein like components are similarly numbered as in FIG. 15. The transversal filter 705 of FIGS. 15, 16, 39 is replaced by digital inverting means 705a, which is selectively enableed via a control input 707a thereto. In one frame the inverting means passes the incoming signal from the bandpass filter mans 704 without changing it, whereas in the alternate frame the control input 707a provides an invert enable signal to the inverting means to shift the bits of the incoming digital word by 180.degree. prior to introducing them to the adder means 708. The luminance signal derived from adder means 706 is delivered to the adder means 708, which latter means generates the composite color television signal on terminal 728, as previously described.

FIG. 18 depicts a modification of the alternative embodiment of FIG. 17, wherein adder means 706 is deleted and the inverting means 705a is replaced by inverting means 705b. Like components in the block of FIG. 18 are also similarly numbered. The inverting means 705b constitutes a digital multiply-by-two (.times. 2) stage 756a coupled to the bandpass filter 704, and thence to a negative input of an adder means 708a adapted to perform a subtraction process. As shown in FIGS. 39E-F, the .times. 2 stage 756a is actually disposed at the output of the bandpass filter means 704, and corresponds to the adder/latch stage 756 of FIG. 40B. The 1H delayed wideband signal on terminal 702 is introduced to the positive input of the adder means 708a.

In operation, the .times. 2 stage 756a is controlled via the control input 707b, i.e, the chroma invert enable signal, whereby in one frame the stage provides a zero output such that the adder means 708a reconstitutes the composite color television signal from only the 1H delayed wideband signal. On alternate frames the chroma invert enable (707b) disables the x 2 stage 756a to allow passage of the digital signal to the negative input of the adder means 708a, together with the wideband signal from the comb filter means 701. The multiply-by-two process is actually performed by shifting the lines one bit, whereby subtraction of the doubled chrominance signal from the wideband signal via the adder means 708a sums the alternate repetitive reproductions to define the composite color television signal on terminal 728.

It may be seen that the system of FIG. 18 is simplified in that the adder means 706 is deleted. In any event, the systems of FIGS. 17 and 18 provide a lesser degree of conditioning of the chrominance signal on repetitive playbacks than does the system of FIGS. 15, 16, 39. Thus the systems of FIGS. 17,18 provide full saturation of the chrominance in the non-inverted frame, with of the order of 1/2 saturation in the inverted frame. However, the stability improvement provided by the all-digital processing, including the inversion process, correspondingly visually improves the color edges.

FIGS. 39E-F depict in schematic the inversion means and control means therefor, for the digital chrominance separating and processing system shown in FIG. 18. To this end, the 1H delayed wideband signal is introduced from the comb filter means 701 (FIG. 38B) via terminal 702, and the bandpass filter means 704 output of the combed chrominance signal is introduced via the digital X2 stage 756a (which herein forms part of the inverting means) from the terminal 703b of FIG. 40B. To simplify the specification, the portion of the inverting means 705b corresponding to the adder/latch stage 756 of FIG. 40B, is depicted in FIG. 39E hereof, by the dashed block 756a inserted after the terminal 703b. The control input 707b, corresponds to the chroma invert enable signal of terminal 757 as previously described. Thus the latter enable signal enables the clear input of the latches of stages 756a on the non-inverting frame, to prevent passage therethrough of the signal and provide in effect the zero input from the bandpass filter to the adder means 708a. On the inverting frame, the chroma invert enable signal disables the clear input of the latches of stage 756a to pass the chrominance signal. The multiply-by-two process is conducted by shifting the wire connections to provide a bit shift of the digital word to double the chrominance signal.

The 1H delayed wideband signal is introduced to a delay 723a (FIG. 39E) similar to delay 723 of FIG. 39A, which equalizes the delays in the wideband signal with those of the chrominance signal introduced via the bandpass filter means 704. The wideband signal is then introduced to a 27/32 multiplier, 722a (FIGS. 39E-F), which performs a gain adjusting function. The wideband signal from the 27/32 multiplier 722a is introduced to the adder means 708a, along with the output from the digital X2 stage 756a. The composite video signal is provided on terminal 728 via the subtraction process conducted on alternate frames, i.e., on alternate repetitive reproductions, by adder means 708a.

As in the circuit of FIGS. 39A-B, FIGS. 39E-F include the control means 709 having the inputs 758, 759, 760 and 761, the PALE clock generator 762, and the count decoder 772, as well as the group A, B control signals on terminal 719, and the PALE clock on terminal 725. As previously mentioned, the chroma invert enable on terminal 757 is introduced to the digital X2 stage 756a. The PALE clock provided by the JK flip-flop 769 via inverters 780, is introduced via lines 781, 782 to the various latches associated with the delay 732a, the 27/32 multiplier 722a, and adder means 708a, to clock the digital signals from the preceeding logical processor component to the succeeding logical processor component, as well known in the art. The various logical elements of FIGS. 39E-F are thus essentially similar to those of FIGS. 39A-B.

FIG. 19 illustrates in block diagram a digital chrominance separating and processing system which generally functions as those previously described, but which reconstitutes the composite color television signal by repetitive reproductions of a single stored color field. As in the previous figures, like components are similarly numbered. Thus the chrominance signal is separated from the color field wideband signal via comb filter means 701, and is introduced to bandpass filter means 704 via terminal 703a. The 1H delayed wideband signal is introduced to the adder means 706 via terminal 702. The combed chrominance isgnal is introduced via terminal 703b to an inverting means 705c, and more particularly to: a transversal filter with odd symmetry 705 similar to that of FIGS. 15,16,39,a third input to an electronic switch means 737; and a first input to a second electronic switch means 738. The number of the inputs of the switches corresponds to the playback number of the single field used to reconstitute the four fields of the composite color television signal. Accordingly, the output from the transversal filter 705 is coupled to a second input to the switch means 737, and to a fourth input to the switch means 738. The output from switch means 737 is coupled to an inverting means similar to 705b of FIGS. 18, 39E-F (or inverting means 705a of FIG. 17), which in turn is coupled to second and third inputs of switch means 738. The output of the latter is coupled to one input of the adder means 708, and the output of adder means 706 is coupled to the other input of adder means 708. Control means 709 provides switching signals via control input 707c to step the switch means 737 and 738 through the inputs thereof at the field rate, to enable the transversal filter 705 and inverting means 705b, and to control the filter means 701, 704, adder means 706, 708, etc., as described above.

As is well known, a 90.degree. phase rotation is required between fields since there is an integer number plus three-fourths cycles of subcarrier in a field. Thus the inverting means 705c provides shifting of the single stored field by 90.degree. on each of four successive plays thereof, to reconstitute the four fields of the composite color television signal. To this end, on first playback of the stored field, the switch means 738 is stepped to the first input thereof, to deliver the combed chrominance signal from the bandpass filter means 704 directly to adder means 708 through switch means 738, together with the incoming luminance signal from adder means 706. The first field at 0.degree. phase shift is thus delivered to terminal 728.

On the second playback of the stored field, switch means 737, 738 are stepped to the second inputs thereof, and the chrominance signal is delivered to the adder means 708 via the transversal filter 705, switch 737, the inverting means 705b and the second input of switch means 738. The transversal filter 705 provides a phase shift, for example, of +90.degree. and the inverting means 705b a phase shift of 180.degree., to rotate the frequency components of the chrominance signal through +270.degree..

On the third playback of the field, switch means 737, 738 are stepped to the third inputs thereof, whereby the chrominance signal is delivered to the adder means 708 via switch means 737, the inverting means 705b and the third input of switch means 738. The chrominance signal is thus rotated +180.degree..

On the fourth playback, the switch means 738 is stepped to the fourth input, whereby the chrominance signal is delivered to adder means 708 via the transversal filter 705 only, to provide a +90.degree. rotation of the chrominance signal. The four fields are combined on successive playbacks via adder means 708 to generate the composite color television signal on terminal 728.

The sign of the phase shifting may be changed, and the circuit connections and clocks thereto adapted correspondingly, whereby on the second playback of the field the transversal filter 705 rotates the chrominance -90.degree. and is then coupled to the adder means 708. On the third playback the inverting means 705b rotates the chrominance -180.degree., and on the fourth playback the transversal filter 705 provides -90.degree. rotation, and is coupled to the inverting means 705b which provides -180.degree. rotation, wherein the combination shifts the chrominance -270.degree., thus providing the 90.degree. phase shift between playbacks.

The control means 709 provides the PALE clock, the four-phase clocks, the chroma invert enable signal, etc., to the various components of the inverting means 705c, the filter means 701, 704 and to the adder means 706, 708, as described and shown in the embodiments of the previous figures.

As well known, when a composite color television signal is reconstituted from a single field, the horizontal sync pulses are not aligned on successive playbacks without the addition of one-half horizontal line delay on alternate fields. Although the chrominance processor of FIG. 19 is not directly concerned with this problem and will deliver the desired succession of fields, the use thereof would require adjunct means for detecting the vertical interval and for inserting the one-half line delay in response thereto as required, and as conventionally known in the art.

Although a 3 f.sub.SC sampling rate is employed in the description above, other sampling rates may be used. For example, 4 f.sub.SC 16/5 f.sub.SC, etc., may be employed. A sampling rate which provides an integral number of samples per television line is advantageous since PALE clock is not required; i.e., the PALE clock generator 762 may be omitted. Thus, the PALE clock per se is not necessary to provide the chrominance separating and processing functions herein. In addition, components such as the 27/32 multiplier and multiplier PROMs may be deleted from the systems, in the event a bandpass filter of unity gain is employed.

The functions carried out by the blanking insertion and bit muting circuitry are primarily those of inserting a black level during the blanking period as well as inserting a gray level during the time in which one picture or still image has been played and another has been addressed for playback. The disc drive head movement may take from one to four fields of time duration in which to change from one still to another, the time increasing the greater the radial movement. Thus, if a track on the outside of a disc pack was being played and the next picture frame that was addressed happened to be on an inside track of the same disc pack, then almost four full fields of time would be required for the heads to move to the new position. Since it is aesthetically pleasing not to have a black picture during this tine period, a grey level is inserted. The circuitry also is adapted to perform bit muting operations which essentially enables one or more of the bits defining samples of a field to be set to the logical zero state for the purpose of performing special effects during playback. The circuitry shown in block 127 of FIG. 8A also generates a PALEd 3SC clock signal from a PALE flag signal for use by the digital-to-analog converter circuitry 102 and it also generates a continuous subcarrier sine wave signal that can be phase adjusted from the continuous phase 6SC and 1/2SC square wave signals that are applied to the circuitry by the reference clock generator circuitry 98. Moreover, in accordance with the present invention, the circuitry is also adapted to adjust the 1/2 cycle of 3SC that is present during the second playback of a picture frame that was detected in the reference clock generator circuitry 98 as previously discussed. The chroma inversion enable signal that enables the chroma separator and processing circuitry 101 to invert the phase of chrominance of alternate frames of the receive television signal during playback operations is also generated by the circuity 127 and is output over line 874 (FIG. 20).

The operation of the blanking insertion and bit muting circuitry 127 will now be described in connection with a block diagram shown in FIG. 20. The frame delay signal from the reference clock generator 98 is input on line 857 to one input of an exclusive OR gate 872, the other input of which is supplied by line 878 carrying the PALE flag signal received from the reference logic circuitry 125B. The output of the gate 872 appears on line 878' extending to steering logic 876. The frame delay signal serves to invert the PALE flag signal at a picture frame rate, thereby superimposing a fram-to-frame 178 6SC clock period offset onto the PALE clock, which is used at the output of the blanking insertion and bit muting circuitry 127 and following digital-to-analog converter circuitry 102 to effect the repositioning of the final output video.

In order to insure reliable repositioning of the video data and strobing of the data within the digital-to-analog converter 102 by the PALE digital-to-analog converter clock that is modified by the frame delay switch signal through the EXCL OR gate 872, the video data itself is selectively delayed by a 1/2 clock period, so that strobing of the data does not occur during a transition between bits. This is accomplished by the upper portion of the circuitry shown in FIG. 20 as follows. The video data from the chroma processing circuuity 101 is applied on lines 850 which extends to an 8 bit latch 851, the output of which appears on lines 852 which extend to another 8-bit latch 853 as well as to 4 to 1 by 8 bit data multiplexer 854. The latches 851 and 853 are clocked by the continuous phase 6SC clock on line 855 and the output of the 8 bit latch 853 is also applied to the multiplexer 854 on lines 856. Each of the latches effectively clocks the data from lines 850 through with a 1/2 cycle of 3SC delay, so that the data appearing on line 852 is delayed 1/2 cycle of 3SC, whereas the data on line 856 has a full cycle of 3SC delay, by virtue of having been clocked through the two latches. While the same data is applied to the multiplexer 854 by lines 852 and 856, the data on line 856 is 1/2 cycle of 3SC delayed relative to the data on line 852.

The frame delay signal from the reference clock generator circuitry 98 on line 857 also extends to address control logic, indicated generally at 858, which controls the multiplexer 854 through lines 859. In accordance with the present invention, the frame delay signal commands the address control logic during alternate frames to alternately pass the data from lines 852 and 856 to correct for the 1/2 cycle of 3SC offset that is present on the second playing of the picture frame as previously described.

When the black mute, or grey mute commands provided by the computer control system 92 are applied on lines 860 and 861, respectively, they are strobed into a latch 862 by the V drive (strobe 1) generated by the reference input circuit 93A and provided on line 862'. The latch 862 controls address control logic 858 in accordance with its stored command to cause the logic to provide the appropriate levels on lines 859 to insert the black or grey level digital information on lines 863 and 864, respectively, so that the black level or grey level data inserted in the video data stream appears on output lines 865. The black and grey levels are produced by setting switches 866 and 867 with the appropriate 8 bit word digitally defining the black or grey levels. When selective bits are to be muted, bit mute control lines 868 are applied to the multiplexer via line 869, assuming that gates 870 are enabled by a bit mute enable signal on line 871 that originates at the address control logic 858. Bit muting is inhibited during the blanking interval so as not to change the setup level of the video. The inhibiting is accomplished by the H and V gated blanking signal provided to the address control logic 858 by the D/A converter and sync insertion circuitry and 103 over line 858'.

With respect to the generation of the PALE SC signal, the continuous phase 1/2 SC and 6 SC inputs appear on lines 873 and 855, respectively, with the 1/2 SC signal being applied to a pulse former 875 that forms 1/2 SC pulses that extend to steering logic 876 via line 877. A PALE flag signal appearing on line 878 steers the 1/2 SC pulses to either the set (879) or reset (880) inputs of a divide by two divider 881 that is clocked by the 6SC signal on line 855. The output is a 3 SC signal on line 882 that is changed in its phase by the 1/2 SC pulses appropriately steered by steering logic 876 in accordance with the level of the PALE flag signal on line 878.

The 6 SC and 1/2 SC signals are also applied to a coarse burst phase circuit 884, the output of which appears on line 885 into a 6 bit shift register 886 that is clocked by 6 SC and has 6 lines to permit the picking up of every 60.degree. of burst phase and apply the selected phase burst signal over line 887 into a voltage variable capacitor network 888 which permits fine burst phase adjusting with control 889. The outlet is a SC square wave signal on line 890 that is applied to a limiter and filter 891 to produce a continuous sine wave SC signal on output line 892 for use in generating the burst for the composite analog television signal.

Specific circuitry that can be used to carry out the operation of the block diagram of FIG. 20 is illustrated by the detailed electrical schematic diagrams of FIGS. 41A and 41B. Since the operation of the circuitry shown in FIGS. 41A and 41B operates substantially as does the circuitry exemplified in the block diagram of FIG. 20, it will not be explained in detail.

However, with respect to the address control logic 858, it provides appropriate commands on lines 859, 871 and 874 to operate the blanking insertion and bit muting circuitry 127 to pass data to the following D/A converter and sync insertion circuitry 102 in accordance with the control inputs at lines 860, 861, 862' and 874'. The EE/PB signal provided by the encoder switch 126 over line 874' from control signals provided by the computer control system 92 is strobed into the latch 862 by the V drive signal on line 862'. When playback operations are performed, the latch 862 places the chroma invert enable command on line 874 which extends to enable two circuits. One of the circuits is the chroma separator and processing circuitry 101 as previously mentioned. The other is a NAND gate 857a in the frame delay switch line 857. The NAND gate 857a is enabled by the command to pass the frame delay switch to the address control logic 858 for use as previously described. During E to E operations, the chrominance of the video signal is not inverted and the previously mentioned frame to frame 46 nanosecond jitter does not occur in the video signal processed by the playback system 91 because a continuous four field color encoded television signal is provided to the electronics of the playback system 91. The EE/PB signal latched into latch 862 disables the NAND gate 857a and removes the chroma inversion enable signal status from line 874.

The address control logic 858 includes NAND gates 883a, 883b and 883c and a multiplexer 858a for directing the commands provided by NAND gates 883a and 883b onto the appropriate multiplexer control lines 859. NAND gate 883c inhibits bit muting during blanking for reasons discussed above and is provided with three inputs connected to receive the gated blanking signal over line 858' and the black and grey mute commands from the latch 862. Should any of these three functions become active, associated inputs of 883c will go low forcing line 871 high, disabling the bit muting circuitry. Consequently, NAND gate 883c provides a bit mute enable signal on line 871 except during blanking intervals and grey and black mute operations.

NAND gates 883a and 883b have their inputs connected so that during normal playback operations NAND gate 883b provides a low level signal output and NAND gate 883a provides a high level signal output. The multiplexer 858a switches these output signals at the two lines 859 each frame in response to the frame delay switch signal 857 to cause the 4.times.1 multiplexer 854 to alternately pass the data received from the two latches 851 and 853 as previously described

When a grey mute command is placed on line 861, latch 862 provides a low disabling signal to one of the inputs of the NAND gate 883c, thereby removing the bit mute enable signal from line 871. However, the inverter 861a inverts the low level provided by the latch 862, causing the output of the NAND ggate 883a to be low. The multiplexer 858a activates lines 859 to cause the 4.times.1 multiplexer 854 to couple grey level digital information from lines 864 to the data output lines 865.

Black level mute operations are selected by the switch 860a being placed in a condition to couple the black mute command output of the latch 862 to one input of each of the NAND gates 883a, b and c. The black mute command causes all of these gates to issue high level signals. Hence, the bit mute enable signal is removed from line 871. Also, the multiplexer 858a activates lines 859 to cause the 4.times.1 multiplexer 854 to place black level digital information from lines 863 on data output lines 865.

The final playback processes that are performed in the signal system shown in the block diagram of FIG. 8A and 8B involve the converting of the digitized video signals to an analog signal in a proper manner as well as generating and inserting the color burst and the composite sync signals. However, before these processes are performed, the video data, delayed on alternate picture frames by 1/23 SC and present at the output of data multiplexer 854 (FIG. 20) is clocked into a latch 901 with the PALE 3 SC clock provided by the blanking insertion and bit muting circuitry 127 on line 902, which effects the reclocking of the correctly repositioned video data. The functions that are carried out will be described in connection with the block diagram of FIG. 21 which has the digitized video information on the eight bit lines 900 that extend from the blanking insertion and bit muting circuitry 127 to latches 901. The latches serve to fix the repositioning of the video data to remove the aforementioned 47 nsec picture frame-to-picture frame jitter and, also, latch each of the bits on the bit lines to align them so that the digital-to-analog conversion can be made. The 3 SC PALE clock generated by the blanking insertion and bi muting circuitry 127 is applied on line 902 which clocks the latches 901 as well as the following timing circuits, including a second latch 903 and a resampling gate 904. The output of the latches 901 containing the digitized video information is clocked through output lines 905 into current switches 906 that have reference current generators connected thereto. The current switches 906 are connected via lines 907 to a resistor ladder network 908 that provides a weighted analog value of each eight bit digital word, thus providing an analog value having 256 possible levels.

The analog output signal from the ladder network appears on line 909 that splits into two paths, an upper path 910 and a lower path 911, the upper path 910 of which represents the normal path during which the video information is passed into a switch 912. The lower path 911 extends to a blanking filter 913 which is switched during the blanking time for the purpose of shaping the blanking pulse so it has the proper transition rate.

If the reshaping filter is not utilized, then the rapid video to blanking transition time can cause ringing in many television receivers. Accordingly, the output of the filter 913 appears on line 914 into the switch 912 which is controlled by line 915 that comes from the latch 903 which is clocked by the 3 SC PALE clock on line 902. During operation, the analog signal on line 909 extends through both paths 910 and 911 and the switch 912 is in the upper position passing the video information except during the blanking period. During the blanking periods the switch 912 is switched to the lower position which connects to the resample gate 904 the signal that has been filtered by the blanking filter 913.

The signal from the switch 912 appears on line 916 that is connected to the resampling gate 904 which operates to sample the level of the signal immediately before a level transition at a point where all transients from the previous transition have disappeared. For example, in the eight bit digital word, a change in value may result in up to seven or eight changes between logical states, i.e., from 1 to 0, each of which will produce a transient condition in the switch. The resampling gate 904 provides a sample and hold operation while blocking the transients so that they do not affect the analog information that is present on line 917 that extends to the buffer and low pass filter 918.

The output of the low pass filter is connected to an amplifier and equalizer 919 via line 920 which performs a sine x/x roll off compensation. The compensated signal is then applied to a black clipper circuit 921 which clips any luminance components of the video signal which appear below black level. The output 922 of the equalizer 919 is also part of a DC restoration loop comprising a switch 923, and a loop amplifier 924 which produces a feedback signal to the low pass filter. The switch 923 is controlled by a clamp pulse on line 925, effecting DC restoration of the video signal on line 922. The clamp pulse is contained in the blanking and composite sync signals provided on a pair of lines 933 by the reference input circuitry 93B.

The output of the black clipping circuit 921 appears on line 927 that extends to the sync and burst adder 928. Burst is added to the signal by line 929 and sync is added by line 930 so that a complete composite analog television signal appears on line 931 to output amplifiers 932. The sync signal is generated by a sync shaping circuit 934 that utilizes a sync pulse contained in the blanking and composite sync signals appearing on line 933, with the sync shaper providing the proper 140 nanosecond rise time and proper shaping. The burst is produced by a burst flag signal provided by the burst envelope generator 936 in response to a burst flag signal provided by the reference input circuitry 93B on line 935. The burst flag signal triggers the burst envelope generator 936 to modulate the SC sine wave generated in the bit muting and blanking insertion circuitry 127 previously described. The output on line 929 contains the burst envelope with the 9 to 11 cycles of burst therein which are added in the sync/burst adder 928 to the analog video signal supplied on line 927.

One embodiment of specific circuitry that can be used to carry out the operation of the block diagram of FIG. 21 is shown in FIGS. 42A and 42D which operates in the manner as described with respect to the block diagram of FIG. 21 and therefore will not be described in detail. However, referring to FIGS. 42A and 42B, a blanking signal is applied to line 950 which extends to the latch 903 and produces an output which extends via lines 915 to a number of switching transistors 953. These transistors 953 together with two transistors 954 and 955, respectively comprise the switch 912 that selects either the signal on the upper path 910 or on the lower path 914 from the filter 913. When blanking occurs, the transistors 953 effectively cut off transistor 954 while placing transistor 955 into conduction and during all other times, the reverse switching occurs.

With respect to the resampling gate 904, a clock appearing on line 902 extends to a number of inverters 957 and 958 which have the effect of providing a small amount of propagation delay to the signal so that the clock signal on line 902 that extends to transistors 961 and 959 are out of step with one another which has the effect of providing a positive transition in the primary side of a transformer 960 secondary of the transformer 960 is connected to a diode bridge 904 that blocks signal flow during the period of the pulse to prohibit passing of the transients or spikes during switching of the digital-to-analog converter switches 906.

FIG. 22 shows a portion of the data detector and equalizer 99 of the record/playback channel, including a reproduce head 1008 coupled to a preamplifier 1009, the combination of elements 1008 and 1009 being designated as block 1001. The magnetic flux patterns recorded on a disc drive surface are picked-up by the reproduce head 1008 and amplified by the preamplifier 1009. Due to the differentiating action of the reproduce head, which is well known in the magnetic recording art, the output signal of block 1001 at terminal 1006 is a voltage proportional to the time-derivative of the recorded flux. Hence, the transfer function of block 1001 in the conventional symbolic notation of the Laplace transformation is

where

G.sub.1 is a complex transfer function

k.sub.1 is a gain constant, and

s is the complex Laplace variable.

Note: With respect to the above-indicated symbolic notations G; k; s; these will be maintained throughout the specification while only the indexes thereof will be changed, indicating specific circuits to which the notations pertain. In the following equations symbolic notations R,C with indexes attached thereto indicate respective resistance and capacitance values pertaining to corresponding circuit elements indicated by identical notations and indexes in the specification and drawings.

To the output of block 1001 of FIG. 22 an equalization circuit 1000 is coupled, the later circuit being shown in an idealized form suitable for theoretical explanation of the equalization operation which follows. The equalization circuit 1000 has an input terminal 1006, to which the output signal of block 1001 is fed. To the input terminal 1006, inputs of an integrating circuit 1002 and a differentiating circuit 1003 are coupled, respectively. The transfer function of the integrating circuit is

and the transfer function of the differentiating circuit is

In the differentiating signal path, a variable gain control circuit 1004 is shown which enables to change linearly the high frequency boost effected by the differentiating circuit 1003, as it will be explained later in more detail. The difference of the respective output signals of the integrating and differentiating circuit is taken, as it is schematically shown by a subtraction circuit 1005. The resulting difference signal at output terminal 1007 of the equalization circuit 1000 is the required amplitude and phase-equalized signal with respect to the input signal at terminal 1006. The resulting record/playback channel has an overall flat amplitude response and linear phase response for all transmitted signal frequencies, as will be seen from the more detailed description below.

The overall transfer function of the portion of the record/playback channel shown in FIG. 22 comprising block 1001 and the equalization circuit 1000 coupled thereto is

and after substituting for G.sub.1, G.sub.2 and G.sub.3 from (1), (2) and (3) ##EQU1## When substituting s = jw, the following is obtained ##EQU2##

The overall phase shift introduced by the portion of the record/playback channel shown in FIG. 22 is determined by ##EQU3##

Since the expression on the right side of equation (6) is a real number (the imaginary part being zero), the overall phase shift determined by equation (7) is zero. At zero phase shift, the requirement of a linear phase response for all frequencies transmitted through the channel is satisfied.

It is essential for the equalization circuit to provide a difference signal at the output terminal 1007, rather than a sum of the respective output signals of the integrating and differentiating circuit. Each of the latter circuits introduces an equal phase shift of 90.degree. but opposite in sense, lagging in the integrator and leading in the differentiator. Thus, the respective output signals of circuits 1002 and 1003 in FIG. 22 are out of phase by exactly 180.degree. with respect to each-other and a difference signal yields a resulting signal combination, for which the respective signal amplitudes are added together, rather than subtracted from each other. Besides that, a -90.degree. phase shift of the integrator output signal combined with the +90.degree. phase shift of the differentiating action of the reproduce head yields an 0.degree. overall phase shift. On the other hand, the +90.degree. phase shift of the differentiator output signal combined with the +90.degree. phase shift of the differentiating head yields a 180.degree. overall phase shift which is simply an inversion. Whether the resulting overall phase shift of the record/reproduce channel is 0.degree. or 180.degree., that is, whether the output signal at the terminal 1007 is in phase or inverted with respect to the polarity of the recorded flux, depends on the sense of the 90.degree. phase shift introduced by the equalizer 1000 as it will be described later in more detail.

Besides providing a linear phase response for all the frequencies transmitted through the channel, the equalization circuit also compensates for the non-constant amplitude-frequency response of the reproduce head, as it will be disclosed below. As it is well known in the art, the output voltage of the reproduce head 1008 and preamplifier 1009 combination of FIG. 22 rises at low frequencies at a rate of 6dB/octave, levels off at mid-band frequencies and falls at high frequencies. Such an amplitude response is shown as an example at G.sub.R in FIG. 25. Consequently, if an overall flat amplitude response of the record/playback channel is to be obtained, it is necessary for the equalizer to boost the amplitude at both low and high frequencies. This required equalizer characteristic is obtained by the circuit of FIG. 22 in a following manner. As an example, FIG. 26 shows a graph representing the gain G.sub.2 of the integrating circuit 1002 and the gain G.sub.3 of the differentiating circuit 1003 in dB, respectively, as dependent on frequency, the frequency values being plotted on a logarithmical scale. The characteristic G.sub.2 falls and the characteristic G.sub.3 rises with frequency at a rate of 6dB/octave. There are also shown diagrams of two other transfer functions G.sub.3 ' G.sub.3 " of the differentiating circuit, representing linear variation of these functions with variation of the gain control circuit 1004 output signal, as it will be described in more detail later. At G.sub.E a resulting transfer function of the equalization circuit 1000 is shown, obtained by adding the linear magnitudes G.sub.2 and G.sub.3. It can be seen that the transfer characteristic G.sub.E of the equalization circuit 1000 is complementary to the transfer characteristic G.sub.R of the reproduce head. Consequently, when combining the two characteristics G.sub.R and G.sub.E, as it is provided by the circuit shown in FIG. 22, the equalizer characteristic G.sub.E compensates for the departures from flatness of the reproduce head characteristic G.sub.R both at low and high frequencies and an overall flat amplitude characteristic results.

There is an additional advantage provided by the presently described equalization circuit which allows linearly varying the amount of high frequency boost provided by the differentiating circuit. For this purpose a variable gain control circuit is utilized in the differentiating signal path, shown for example at 1004 in FIG. 22. By adjusting the gain of the differentiating signal path by means of circuit 1004, the frequency at which the high frequency boost of the equalizer amplitude response begins may be changed. For this purpose a variable resistor or potentiometer may be utilized or in case an amplifier is employed in the differentiating signal path, the gain of that amplifier may be changed in a well known manner, as it will be described in connection with the embodiment of FIG. 24. The group of curves G.sub.3, G.sub.3 ', G".sub.3 shown in FIG. 26 is obtainable for three different values of gain provided by the differentiator 1003 in FIG. 22 and adjusted by the variable gain control circuit 1004. The gain adjustment affects only the gain constant k.sub.3 in the transfer function (3) presented above and therefore, it changes only the corner frequency at which the high frequency boost begins, in accordance with the formula for the corner frequency ##EQU4## As the corner frequency increases, the amount of signal amplitude boost decreases linearly as the curves obtained move from G.sub.3 to G'.sub.3 to G".sub.3 , etc. Increasing the amplitude boost linearly at the high frequency end of the equalizer response is an important feature because it enables to compensate, for example, for changes in the relative head-to-recording medium speed, such as due to the variations in track length of a magnetic disc. When recording digital signals on magnetic disc, this feature allows to compensate for higher density of recorded bits, also called pulse crowding which occurs on the inner tracks of the disc.

Examples of practical implementation of the above-described idealized form of the equation circuit shown in FIG. 22 are shown in the form of block diagrams in FIGS. 23 and 24. Elements similar to those previously described and shown in FIG. 22 are designated in FIGS. 23 and 24 by the same reference characters as in FIG. 22. With respect to the relatively low signal level at the output of playback amplifier 1009, it is necessary for practical purposes to amplify the signal in both the integrating signal path as well as in the differentiating signal path. Thus, in the diagram of FIG. 23 the integrating circuit of FIG. 22 is implemented by inverting integrating amplifier circuit 1002, comprising an inverting operational amplifier 1010, a negative feedback capacitor C.sub.1 and a series input resistor R.sub.1. On the other hand, the differentiating circuit of FIG. 23 is implemented by an inverting differentiating amplifier circuit 1003, comprising an inverting operational amplifier 1011, a negative feedback variable resistor R.sub.2 and a series input capacitor C.sub.2. The variable resistor R.sub.2 represents a variable gain control for the differentiating signal path. The transfer function of the integrating amplifier circuit 1002 of FIG. 23 is: ##EQU5## when comparing equation (9) with (2) we obtain ##EQU6## The transfer function of the differentiating amplifier circuit 1003 of FIG. 24 is

When comparing equation (11) with (3) we obtain

The subtraction circuit of FIG. 22 is implemented in the circuit of FIG. 23 by a differential amplifier 1005. The output of the inverting integrating circuit 1002 is coupled to an inverting input of the differential amplifier 1005 while the output of the inverting differentiating circuit 1003 is coupled to a noninverting input of amplifier 1005. The output signal at terminal 1007 is the difference signal which also represents the equalized signal of the recording/reproducing channel. The resulting equalized signal has 0.degree. phase difference with respect to the signal recorded on the magnetic medium, that is, it is in phase therewith. Thus, the phase response of the overall channel becomes linear when the equalization circuit 1000 is utilized.

However, the circuit of FIG. 23 is still considered idealized to the extent that exact implementation of the above transfer functions (9) and (11) would require unlimited gain in the integrating amplifier circuit 1002 at low frequencies and in the differentiating amplifier circuit 1003 at high frequencies. In practical applications both these extremities are avoided, for example, by adding a shunt resistor R" to C.sub.1 and a series resistor R' to C.sub.2 as shown in FIG. 23, to truncate the respective integrating and differentiating approximations at selected frequencies below and above the frequency range of interest. Considering the presence of the respective resistors R', R" in the circuit of FIG. 23, the respective transfer functions G.sub.2, G.sub.3 will be ##EQU7## where R.sub.1, R.sub.2, R', R", C.sub.1 and C.sub.2 are component values pertaining to corresponding circuit elements. When considering in equation (13 ##EQU8## We obtain ##EQU9## which is identical to the transfer function of (2). When considering in equation (14) ##EQU10## we obtain

which is identical to the transfer function of (3).

It follows from the above discussion that when substituting for s = jw, the respective transfer functions of the integrating and differentiating circuit of the equalization circuit 1000 shown in FIG. 23 will approach that of an ideal integrator and differentiator in the frequency range ##EQU11##

In FIG. 24 still another example of practical implementation of the equalization circuit is shown. The integrating circuit. of FIG. 22 is here implemented by a passive integrating network 1002 comprising series resistor R.sub.A and parallel capacitor C.sub.A followed by a non-inverting amplifier 1012 providing the necessary amplification in the integrating signal path. Analogously, the differentiating circuit of FIG. 22 is implemented in FIG. 24 by a passive differentiating network 1003 comprising a series capacitor C.sub.B and a parallel resistor R.sub.B followed by a non-inverting amplifier 1013 providing the necessary amplification in the differentiating signal path. Similarly as in the circuit of FIG. 23 the subtraction circuit is implemented by a differential amplifier 1005. In the circuit of FIG. 24 the integrated and subsequently amplified signal at the output of amplifier 1012 is fed to a non-inverting input of the differential amplifier 1005, while the differentiated and subsequently amplified signal at the output of amplifier 1013 is fed to an inverting input of amplifier 1005. The output signal at terminal 1007 of the circuit in FIG. 24 is the resulting difference signal which represents the equalized signal of the record/playback channel. The resulting equalized signal has a 0.degree. phase difference with respect to the signal recorded on the magnetic disc. Thus, the phase difference caused by the presently described equalization circuit does not introduce non-linearities in the phase response of the overall channel, but to the contrary, it yields an overall linear phase response.

The respective transfer functions of the integrating and differentiating circuit of FIG. 24 are ##EQU12## where A.sub.2 is the gain of amplifier 1012 and A.sub.3 is the gain of amplifier 1013. When comparing equation (20) with (2) we obtain for w>>(1/R.sub.A C.sub.A) ##EQU13## When comparing equation (21) with (3) we obtain for w<<(1/R.sub.B C.sub.B)

a potentiometer 1014 in FIG. 24 connected to the amplifier 1013 in the differentiated signal path represents a variable gain control circuit. By adjusting the gain A.sub.3 of amplifier 1013, the gain constant k.sub.3 expressed by (23) and the corner frequency of the boost changes as it has been described above in connection with the description of FIG. 26 and equation (8).

A detailed electrical circuit diagram of the data detector and equalizer 99 is shown as an example in consecutive FIGS. 43A and 43B and will be now described. In the video frame storage recording and reproducing system, a color television signal is encoded in digital form and recorded on a magnetic disc. The digital code utilized is the DC free self clocking channel code described in the above identified U.S. Pat. No. 4,027,335. Upon playback, the digital data is reproduced by a reproduce head 1008 and amplified by a reproduce 1009 (reproduce head and preamplifier are shown in FIG. 44B). FIGS. 43A and 43B show two identical playback equalizer and data detector circuits utilized for the ten separate data streams received from the disc drive data interface 151. However, only one of these circuits will be described. In the circuit of FIGS. 43A and 43B the preamplified playback data in the channel encoded format, for example, of the type described in the aforementioned U.S. Pat. No. 4,027,335 is equalized by an equalization circuit 1000 corresponding to the abovedescribed equalization circuit with reference to FIGS. 22 to 24. The equalized signal is filtered in a low pass filter circuit 1018, and thereafter amplified and amplitude limited to produce a rectangular pulse sequence in an amplifier-limiter circuit 1019. The pulse sequence from the limiter is fed through a pulse former circuit 1020 which forms output pulses for each detected signal transition. The pulses from circuit 1020 are fed to the data decoder and time base corrector circuitry 100 which decodes and removes timing errors from the playback data from which the original color television signal is recovered.

As shown in FIGS. 43A and 43B, the playback data from the preamplifier is applied to differential input terminals 1021 and 1022 of a differential amplifier 1033. The amplifier contains open-collector differential output transistors connected to output terminals 1034 and 1035. Resistor 1036 is the load resistor for the non-inverting output terminal 1034. The gain of the amplifier 1033 to output terminal 1034 is constant throughout the frequency range of interest. The non-inverted signal is buffered by emitter follower 1037 and then applied to a differentiating network 1003 comprising capacitor 1038 and resistor 1039. This network 1003 performs differentiation for signal frequencies below 60 MHz. Its transfer function is ##EQU14## Equation (23) corresponds to previously discussed equation (3) related to the block diagram of FIG. 22 where k.sub.3 = (R1039)(C1038). Since signals of interest in this particular embodiment extend only to about 10 MHz, this network 1003 may be viewed as a true differentiator. The output of the differentiator 1003 is applied to input terminal 1040 of differential amplifier-multiplier circuit 1041. Input terminals 1040 and 1042 of the circuit 1041 are differential input terminals biased by connection to +7.5V. The amplifier-multiplier 1041 receives a second input signal at differential input terminals 1043 and 1044 and at output terminal 1045 an output current is provided proportional to the negative of the product of the input signals at terminals 1040, 1042 and 1043, 1044. In the present circuit a direct current gain control voltage is applied to input terminal 1043, while terminal 1044 is grounded. The control voltage at 1043 corresponds to an output voltage from a remote variable gain control circuit (not shown on FIG. 43), such as previously described in connection with circuit 1014 of FIG. 24. In the presently described embodiment of the frequency equalizer, the gain of the circuit 1041 in the differentiated signal path is remotely and automatically controlled by a digital-to-analog converter to obtain desired gain variations dependent on the variations of the recording track length of the magnetic disc. A particular track number (corresponding to a specific track length) from which a particular data is being reproduced is decoded in a digital decoder and converted in the digital-analog converter to a direct current voltage level which is then applied as a gain control signal to input terminal 1043 of circuit 1041. As it has been mentioned before, the variable gain adjustment in the differentiated signal path is designed to compensate for higher pulse density on inner tracks of the disc while linearity of the high frequency boost of the equalized signal is maintained for the entire frequency band transmitted.

The magnitude of the current at output terminal 1045 of the amplifier-multiplier circuit 1041 is proportional to the input signal at input terminal 1040 and to the gain value determined by the control voltage at terminal 1043. The output current from terminal 1045 of the circuit 1041 is applied as an input current to the emitter of a common-base transistor amplifier serving as the subtraction circuit 1005 which has been previously described and shown in FIGS. 22, 23 and 24. This input current produces an output voltage at the collector of the amplifier which is proportional to both the input current and resistance of a collector load resistor 1047. Thus, the above-indicated part of the transistor 1005 output voltage is proportional to the negative of the signal derivative amplified by the amplifier-multiplier circuit 1041.

The inverting output terminal 1035 of the differential amplifier 1033 has a load resistor 1048 and a parallel load capacitor 1049. The direct current gain of the amplifier 1033 to output terminal 1035 is higher than the gain to the non-inverting output terminal 1034 by the ratio of the respective load resistances R1048/R1036, that is, by the factor of about 3. For signal frequencies above 80 kHz the gain to output terminal 1035 is determined by C1049 and is inversely proportional to the frequency. Thus, the output circuit R1048, C1049, connected to terminal 1035 functions as an integrating network for frequencies above 80 kHz and throughout the frequency range of interest which is approximately from 0.3 MHz to 10 MHz. The transfer function of the amplifier 1033 to the output terminal 1035 is ##EQU15## where A.sub.1033 is the gain of the differential amplifier 1033 to output terminal 1034. ##EQU16## Equation (25) corresponds to previously discussed equation (2) related to the block diagram of FIG. 22, where ##EQU17##

The inverted and subsequently integrated signal from the output terminal 1035 of amplifier 1033 is applied to the common emitter transistor amplifier 1005. Transistor 1005 inverts this input signal and multiplies it by the ratio of its respective collector and emitter load resistances R1047/R1050. The transistor 1005 operates as a common emitter amplifier in the integrating signal path and as a common base amplifier in the differentiating signal path. The resulting output signal at the collector of transistor 1005 is the sum of two input signal contributions, one proportional to the integral of the playback signal from the reproduce head and preamplifier combination, the other one proportional to the negative of the derivative of the playback signal. Thus, the resulting output signal at the collector of transistor 1005 corresponds to a difference signal, such as previously described with reference to the output signal at the output terminal 1007 of the previously described embodiments of the equalization circuit shown in FIGS. 22, 23 and 24. Thus, the output signal of the equalization circuit 1000 of FIGS. 43A and 43B corresponds to the equalized signal of the record/playback channel as previously disclosed with respect to the embodiments of FIGS. 22, 23 and 24.

Now the remaining part of the detailed circuit diagram shown in FIGS. 43A and 43B will be described. The equalizer 1000 converts the voltage peaks of the playback signal provided by the playback preamplifier 1009 (FIG. 44B), which represent zero crossings of the recorded flux, back into properly positioned zero crossings at the output of the equalizer. This equalized output signal is present at the collector of transistor 1005 of the equalizer and is filtered by a low pass filter circuit 1018 and thereafter fed through a first buffer amplifier 1051 arranged to provide complementary outputs of an amplifier-limiter circuit 1019. The output signal from the buffer amplifier is fed through a series of five amplitude-limiting amplifiers, preferably of the same type as the buffer amplifier. The equalized playback signal provided at the input of the amplitude-limiting circuit 1019 is in the channel encoded form with the transitions properly positioned. Amplitude limiting the playback signal serves to restore the rectangular shape of the playback data signal which has been considerably distorted by the record and reproduce processes. Furthermore, the buffer amplifiers of the amplitude-limiting circuit 1019 also serve to provide opposite phased waveforms of the restored data signal which are subsequently used to generate a pulse for each transition of the rectangularly shaped channel encoded playback data signal. As previously described herein with reference to the channel encoding of the data signals by the encoder 96 and subsequent recording of such signals, the transition-related pulses are generated so that a precisely defined edge, the leading edge being selected in this embodiment, can be sent through a transmission channel without introducing errors to the data although the data signal may be distorted by the channel. As described hereinbefore, the high bit rate data streams, such as processed by the apparatus described herein, are particularly susceptible to having errors introduced into them because of the differential response characteristics of transmission lines to different sensed signal level transitions such as twisted pair transmission lines used to couple channel encoded data between disc drives and the signal system.

To generate a pulse for each transition of the playback data signal so that only leading, positive edges of the pulses identify the data signal transitions, the amplifier-limiter circuit 1019 provides two opposite phased waveforms of the data signal. First, a sequence of transitions between signal levels of non-inverted polarity is provided at the output terminal 1052 of the last amplifier 1053 of the series of amplitude-limiting amplifiers and second, an identical sequence of inverted polarity is provided at the output terminal 1054 of the same amplifier 1053. Both these transition sequences have their transitions positioned according to the code rules of the channel code selected for originally encoding the video data and are applied respectively to clock two identical one-shot multivibrators 1055 and 1056 of the pulse former circuit 1020. Each multivibrator forms a positive pulse, respectively, for each positive going transition of the playback data signal received at its clock input. Consequently, the one-shot multivibrator 1055 receiving the non-inverted form of the playback data signal provides a positive pulse at each positive going transition in the data signal. On the other hand, the other one-shot multivibrator 1056 receiving the inverted form of the playback data signal provides a positive pulse at the location of each negative going transition in the data signal. Since the leading edges of the positive pulses generated by the multivibrators 1055 and 1056 are defined by rapidly switching the multivibrators from its stable state to its quasi-stable state (there being no significant time constant determining components involved), each leading edge will be identical to all others and occur at a precise time following the occurrence of the positive clocking transition of the playback data signal. Because the transmission channel over which the pulses are sent will act on identical pulse edges the same, the locations of the transition-related positive pulse edges, hence, data signal transitions themselves, are not lost as a result of any distortion that may be introduced to the pulses by the action of the transmission channel. If necessary, an amplitude level sensitive detector means can be coupled to the output of the transmission channel, such as is used at the input of the decoder circuitry portion of the previously described decoder and the time base corrector 100, to accurately redefine the relative locations of the playback data signal transitions.

For transmission of the transition related pulses to the signal system, the output pulses of both one shot multivibrators 1055 and 1056 are applied to separate inputs of a positive OR-gate 1057 which forms an output pulse for each input pulse. The output pulses of the OR-gate 1057 are applied to the disc drive data interface 151 (FIG. 8B) for transmission over lines 154 to the data select switch 128, which couples the transmitted pulses to the input of the data decoder portion of the decoder and time base corrector circuitry 100 of the selected playback channel 91 for decoding of the playback data and subsequent processing to recover the original color television signal. The disc drive interface 151 includes a conventional complementary output buffer amplifier arranged to receive a single input signal and generate coincident complementary output signal forms of the single input signal. The complementary buffer amplifier converts each transition related pulse provided by the OR-gate 1057 to a pair of coincident complementary level pulses, which are coupled to the data select switch 128 for transmission to the selected playback channel 91.

FIGS. 44A and 44B show consecutive parts of a detailed electrical circuit diagram including the record driver and playback preamplifier circuits of four identical data record and playback channels, designated 1058, 1059, 1060 and 1061 utilized in the video frame storage record and playback system. A fifth channel designated 1062 includes a servo track head permanently connected to a servo playback preamplifier and it also includes a data track record and playback channel. In the video frame storage record and playback system, five more data record and playback channels (not shown) identical with the above-indicated data record and playback channels shown in FIGS. 44A and 44B are utilized. A relay 1063 in channel 1058 is shown having its contacts in a position connecting one of the heads 1008a and 1008b for recording as occurs when a record command is received from the disc drive control circuitry on line 1066, as described hereinbefore. In absence of a record command, the relay 1063 is in the playback position. In this position, the contacts of relay 1063 are in their alternative positions. Heads 1008a and 1008b are utilized for both recording and playback and are switched alternatively for odd and even television fields. Switching of these heads 1008a or 1008b is controlled by the 30 Hz head switch signal continuously provided on line 1067 provided by the record timing circuit of FIG. 29A located in the disc drive electronics. The playback data received alternately from the heads 1008a and 1008b the respective channels 1058, 1059, 1060 and 1061 is fed into the playback equalizer and data detector circuits associated with the respective channels such as shown in previously described FIGS. 43A and 43B. The record/playback heads utilized in the video frame storage recording and reproducing system are conventional heads such as manufactured by Applied Magnetic Corporation or Information Magnetics Corporation, for digital recording on disc packs of the kind employed in the apparatus.

As previously mentioned, the disc drives 73 that are used in the present apparatus are preferably substantially unmodified so that advantage can be taken of the reliable operation that has been achieved through years of refinements in the design and manufacture of disc drives. Accordingly, the disc drives that are used in the present apparatus are relatively unchanged, except as previously mentioned, i.e., the 8 bits of video data together with one parity bit are simultaneously recorded on 9 parallel surfaces and the data track surface is also recorded with its information. The disc pack drive maintenance manual for the Ampex Model DM 331 disc drive, the manual having Ampex Part No. M300211 which has been incorporated by reference herein, includes Table 2-1 which illustrates the command decodes for the bus within the disc drive as well as the tag lines that control the operation that is occurring. In the Ampex Model DM 331 disc drive, tag line 11 relates to operation and status functions that are not particularly applicable to the operation of the disc drive when used with the present apparatus and, accordingly, several of the circuits that are used therein have been modified as well as replaced with circuits which are uniquely applicable to the present apparatus.

More particularly, the normal computer data processing application of the disc drive utilizes rapid switching between read and write operations within one revolution and also utilizes small sectors of the total disc circumference, of the standard tag 11 operation and status functions deal with this type of operation. However, with respect to the present apparatus, each revolution of the disc pack is used to either record or playback a single field of television information and a single picture frame will require two revolutions of the disc pack, with one field of video information being written on one set of 8 surfaces and the other field of video information being written on 8 different disc surfaces.

Since switching between read and write operations only occurs at the completion of whole revolutions of the disc, with respect to a defined point (specifically referred to as sector 000 or index) and it was chosen to be done during the vertical interval of the television signal, very rapid switching is not particularly critical with the present apparatus.

It should also be appreciated that normal data processing for disc drive recording and playback is at a data rate of about 6.5 megabits per second whereas the video information that is recorded on the disc pack surfaces in the present apparatus is at a rate of about 10.7 megabits per second. Since electronic switching of the heads between the record and play circuitry of standard disc drives causes some deterioration in the signal-to-noise ratio, the electronic switches have been replaced with relays which result in an increase of about 2 dB in the signal-to-noise ratio of the resulting signal that comes off the disc pack.

Since the majority of the circuitry that is associated with the disc drive remains unchanged, only those circuits which have been added or modified will be described herein in a general manner, since they must interrelate with existing circuitry that is not shown, but which has been incorporated by a reference herein.

Referring firstly to FIGS. 28A and 28B, which illustrate electrical schematic diagrams of record and play control circuitry, bus out lines 1820 through 1826 are shown to the left of drawing 28A (one bus line 1827 being shown on FIG. 28B) which are gated through NAND gates 1831 when an operate command valid appears on line 1832. This results when tag line 11 in the disc drive is raised and is checked and determined to be valid. The purpose of the circuitry of FIG. 28A is to latch in commands from the computer control system 92 relating to whether the relays controlling head currents should be placed in a record position or a play position for the purpose of recording on or playing from a disc pack 75 and to command through additional circuitry the spindle servo to provide correct rotational phase of the disc pack with respect to the reference vertical sync. This phasing is as follows: (a) during record, the servo reference signal coincides with the vertical sync pulse of television signal; (b) during play-transfer, the servo reference is advanced one horizontal line duration with respect to the vertical sync pulse of television signal; and (c) during play, the servo reference is advanced two horizontal lines duration with respect to vertical sync pulse of television signal. The signals on the top three bus lines 1820, 1821 and 1822, when gated through NAND gates 1831, are inverted and applied to a 1 of 8 decoder 1834. The decoder 1834 has three of its output lines 1835, 1836 and 1837 that determine in accordance with the input commands the spindle servo phasing defined to be legitimate. All other decoded outputs are ORed into NOR gate 1838 which, after being inverted, is sent via line 1839 to an NOR gate 1840 which generates an operate command reject. This indicates that an improper command has been sent on the first three lines 1820-1823.

Referring to the decoder 1834, output line 1835 is inverted and applied to a NAND gate 1842 which, when enabled, sets a latch, indicated generally at 1843, having output line 1844. This line 1844 provides a signal directing the spindle servo to rotationally phase the spinning disc pack to the record position. Output line 1836 is applied to NAND gate 1845 after having been inverted, which is ORed with a power up reset signal on line 1846 by NOR gate 1847. The output of the NOR gate 1847 resets the latch 1843 via line 1848 which also sets a latch indicated generally at 1850 and directs the spindle servo to provide the play rotational phase command which appears on line 1851. When line 1837 from the decoder is active, it is inverted and gated through a NAND gate 1852 which resets latches 1843 and 1850 and sets a latch 1854 which specifies a play-transfer rotational phase command on line 1855. Thus, any one of the three legitimate outputs of the decoder specify a transfer, record or play rotational phase when the NAND gates 1842, 1845 and 1852 receive an enable store command on line 1856.

Bus lines 1825 and 1826 carry mutually exclusive command signals to set the relays to the record or play position, respectively. When bus line 1825 is high, and the operate command valid is present, NAND gate 1831 will set a latch 1857 which will provide a high on line 1858 which places the relays in the record position and permits a recording to be carried out when the timing is correct. Bus line 1823, when gated through NAND gate 1831, sets a latch 1860 which provides a head select signal on line 1861 which is used for maintenance purposes.

Referring to FIG. 28B, a signal on bus line 1827 together with an operate command valid enabling NAND gate 1831 sets a latch 1862 provided that a store command is present on line 1863 which enables a NAND gate 1864. The output of the latch 1862 provides a record next frame signal that is used in the record timing circuitry shown in FIGS. 29A and 29B. The other commands that are generated by the circuit shown in FIGS. 28A and 28B are a signal on line 1865 indicating that the record sequence has been completed which is sent to the CPU 106 and also resets the record next frame latch 1862.

The circuitry shown in FIGS. 29A and 29B provides the 60 Hz reference signal for the spindle servo control sytem for the pack drive motor. Using the pack drive motor, the spindle servo controls the rotational phase of the disc pack utilizing as the servo reference the color frame shifted signal that is produced by the timing generator circuitry that will be hereinafter discussed. However, as previously mentioned, the television signal must be advanced either one or two television lines relative to its position during recording to compensate for delays that are experienced by the reproduced video data playback as a result of the operation of the playback channel 91 circuitry. The color frame shifted signal that is asserted in the record timing circuitry shown in FIGS. 29A and 29B is positioned correctly with reference to the required timing for each of the operating modes of record, playback and transfer. The circuitry shown on FIG. 29A provides the 60 Hz servo reference signal that is derived from the multiplex sync signal of 2H frequency which is provided by the signal system. In this regard, the 2H signal is divided by 525 to derive the basic 60 Hz reference signal which is phase position controlled by the color frame shifted signal from the timing generator.

The record timing circuitry also provides drive signals for placing the relays in the record or play positions and also provides signals back to the CPU 106 through the drive control lines informing the CPU of the relay position. Moreover, in the described apparatus, a head disable signal is also generated which inhibits head current for at least one revolution of the disc pack after the record/play relay has been switched between its two positions. The recording timing circuitry also generates the signal for switching from one set of recording heads to another set for recording one field on one set of disc surfaces while the other video field is recorded on a second set as described hereinbefore. A basic 30 Hz signal controls the head switching.

Referring specifically to FIG. 29A, a relay set line 1870, which is high when the relays are in the play position and low when they are in the record position, provides an input to a NAND gate 1871, the other inputs of which are essentially supplied by a pulse on line 1872 that indicates the sector 000 (index) on the disc passing the servo head which during normal operation occurs during the vertical interval. When the relays are in the record position and the pulse appears on line 1872, NAND gate 1871 sets a latch 1873 which is coupled to transistors 1874 that provide a relay drive signal that extends to the preamplifier circuitry (FIGS. 44A and 44B) via line 1875. The state of the latch 1873 also provides a signal on line 1876 extending to FIG. 29B indicating that the relays are in the play position or, alternatively, a signal on line 1877 extending to FIG. 29B indicates the relays are in record position.

To produce the reference signal for the servo, a 2H rate signal called multiplex sync and whose timing originates from the signal system circuitry is applied on line 1880 which is inverted and appears on line 1881 that extends to a divide by 256 counter 1882. The counter has output line 1883 that extends to the clock input of a divide by 2 FF 1884, thereby producing a divide by 512 resulting division of the 2H signal on line 1885, the divided signal being used to set a latch 1886 via NAND gate 1887. The latch 1886 is connected to a shift register 1888 that is clocked by the 2H signal on line 1881. The shift register 1888 has output line 1890 that is connected to a shift register 1892. The pulse clocked out on line 1891 from the shift register 1892 represents count 525 and clocks a FF 1893. The FF 1893 provides a pulse on line 1894 that is gated through NOR gates 1895 onto a line 1896 and clears shift registers 1892, 1888 as well as the counters 1882 and 1884. Thus, the terminal count of 525 resets the counters and shift registers. It should be appreciated that the rate of 2H divided by 525 is 60 Hz which appears on line 1897 that passes through an inverter 1898 onto line 1899 and to a NOR gate 1900 producing the 60 Hz signal servo reference on line 1901. The output of the shift register 1888 on line 1897 is also divided by 2 by the FF 1902 producing a 30 Hz rate on line 1903 that is gated to produce the properly phased head switch control signal on line 1904.

If a color frame detected signal appears on line 1906, a FF 1907 is set which inhibits the first NOR gate 1895 and thereby inhibits the clearing of the dividers and shift registers so that the later appearing color frame shifted signal on line 1908 will provide the clear pulse through the second NOR gate 1895 so that the color frame shifted signal will reset the shift registers and FFs to 0 rather than the terminal count. This permits the 60 Hz servo reference signal to be properly positioned relative to the line advancements that are required to have the video information at the proper location during the playback and transfer modes as has been previously described.

The head disable signal provided to the preamplifier circuitry (FIGS. 44A and 44B) for one revolution of the disc pack during a switching of the heads from playback to record is provided on line 1889' by the transistor 1889 in response to the latch circuit 1878 being clocked by the appearance of the index pulse on line 1872 when the latch circuit 1873 is in the record state.

Turning now to the remainder of the timing generator circuitry shown in FIG. 29B, circuitry is illustrated which generates the timing commands that are used to perform the record sequence. The 60 Hz servo signal present on line 1901 from the circuitry shown in FIGS. 29A and B together with a sync present signal on line 1953 enables a NAND gate 1909 whose output is ORed with the color frame shifted pulse on line 1936 by NOR gate 1910. A latch 1911 is set upon the occurrence of the 60 Hz servo signal to provide one input of a NAND gate 1912 associated with a shift register 1913. The NAND gate 1912 is satisfied with the latch 1911 being set together with the shift register 1913 having a low state in all pertinent outputs. Each time this occurs, the 60 Hz servo reference signal on line 1899 clocks the shift register, causing certain ones of a sequence of high signal states to be placed on output lines 1914, which lines are extended to various logic gates to perform the sequence of signals that are needed for as the shift register 1911 is clocked by a sequence of 60 Hz servo reference signals.

A record ready signal on line 1915 results when NAND gate 1916 is satisfied which happens when certain qualifiers are present, i.e., the relays are in the record position, a ready signal is present, a control or access disable reset is not activated, the disc pack has correct rotational phase and the sync is alright. When these qualifiers occur, the record/ready signal is exerted. Similarly, a record next frame signal is produced by NAND gate 1917 and sets a latch 1918 when certain qualifiers are present, including the sync alright signal, record next frame command, the relays are in record position signal, the timing from the shift register 1913, together with a disc being correctly positioned signal. If these conditions are met, the latch 1918 is set and a record sequence signal appears on line 1919. The latch 1918 is reset after four fields as timed by the shift register 1913 and the resetting thereof produces a record sequence complete signal on line 1920. A prerecord signal on line 1921 is generated by a latch 1922 which lasts for a time period of two fields and is reset two fields sooner than the record sequence latch 1918. During the prerecord interval, the black level signal is recorded on the first two revolutions of the four revolution sequence used by the apparatus described herein to record two fields of video data as previously described. It should be appreciated that the latches 1918 and 1922 are both set at the same time. Similarly, a data timing pulse appears on line 1923 for use by the data track circuitry if the record/playback relay is to be toggled at the end of a four field record sequence and it lasts for one field occurring during the last field of the four field record sequence. The data track circuitry provides a head disable switch to the preamplifier circuitry (FIGS. 44A and 44B) to prevent head current from flowing after the sequence when the record/playback relay is toggled.

The timing generator shown in the electrical schematic diagram of FIG. 30 generates the signals that are used to provide the timing functions of the drive, including the operation of the servo system such that the disc pack rotation is phased to the television signal during record and playback. The circuitry utilizes the multiplex sync signal received from the reference logic circuitry 125A and 125B that consists of narrow horizontal rate pulses in addition to a color frame signal that occurs in the form of 3 consecutive wide horizontal rate pulses every fourth television field. This multiplex sync signal is used to generate a horizontal rate signals as well as to provide a color frame output signal that, which is the basic drive operation timing pulse, for timing the functions of the drive. The color frame shifted signal, in addition to other functions, provides the basic phasing of the servo reference so that when a recording operation is occurring, the servo reference coincides with the vertical sync signal of the video signal being recorded. However, when a playback operation is occurring, the servo reference is shifted so that the television signal is advanced by a time period equal to two television lines to compensate for two television lines of delay that occur in the playback channels 91 of the apparatus.

More specifically, the time base corrector portion 565 of the data decoder and timebase corrector circuitry 100 of each playback channel 91 introduces one television line of delay during playback and the chroma separating and processing circuitry 101 of each playback channel 91 also introduces one television line of delay. Thus, when the video information is played back, it would be present at the output two lines later than it should and, accordingly, the servo reference position is adjusted so that the video information is advanced by the two lines during normal playback. However, when a transfer mode is being performed, i.e., a still frame of information is being transferred from one disc pack 75 to another, the playback channel of the apparatus produces only one television line of delay because the information goes through the decoder and time base corrector circuitry 100, but not through the chroma separating and processing circuitry 101. Since the delay introduced by the chroma circuitry is not present in the transfer mode, the position of the servo reference is advanced 1 television line so that it is recorded with a vertical sync pulse coincident with sector 000 (index) on the other disc pack 75. The circuitry associated with the timing generator provides the shifting of the color frame so that the servo reference is in the proper position and also produces a stable H rate signal that is not appreciably affected by reasonable noise levels or the occasional absence of pulses in the multiplex sync signal.

Referring to FIG. 30, the multiplex sync signal is applied at input line 1920' which occurs at H rate and which has the color frame information in the form of 3 consecutive wide pulses occurring every fourth television field. The multiplex sync is then converted from emitter coupled logic level to transistor-transistor logic level by a converter 1921' and passes through an inverter 1922' having output line 1923' that extends to a NOR gate 1924'. Line 1923' is also connected to two AND gates, through an inverter 1925 to one AND gate 1926 and directly to another AND gate 1927. The lower path of the signal to the AND gates 1926 and 1927 operate to detect the presence or absence of information indicating a color frame.

The color frame is detected by strobing the gates with a one-shot 1928, providing a short duration pulse to enable the AND gates 1926 and 1927 so that the pulses that are gated through will either increment or clear a counter 1929. When a color frame information is present, three successive counts will be passed by AND gate 1927 to the counter 1929, which responsively places a high output on both lines 1930. This loads a high into a shift register 1931. In the event that a color frame information is not present, then 3 successive pulses will not occur and the absence of either the second or the third pulse will satisfy AND gate 1926 which, when gated, clears the counter 1929. The shift register 1931 is clocked by a 2H signal on line 1932 to shift the signal placed on its input by the counter 1929 through the register 1931 to place in 1H intervals successively appearing high levels on lines 1933, 1934 and 1935.

The timing of the signals on lines 1933, 1934 and 1935 provide the 1 line, 2 line or 3 line delays (the 3 line delay being defined as a 0 advance, a 1 line delay being defined as a 2 line advance and a 2 line delay being defined as a 2 line advance) for the shifted color frame signal placed on output line 1936 by a decoder 1937. Two position select control lines 1938 provide a binary input command to the decoder 1937 that determines which one of the input lines 1933, 1934 or 1935 will be decoded to place a signal on the output line 1936, and thereby provide the basic shifted color frame reference timing information for the record timing circuitry.

The circuitry also generates a stable horizontal rate signal using a phase lock loop with a voltage controlled oscillator in an integrated circuit 1940 which receives the sync signal from the NOR gate 1924' through inverter 1941, AND gate 1942 and line 1943. The output of the oscillator 1940 appears on line 1944 which is divided by a divide by 10 counter 1945 having a 2H output on line 1946 which is in turn divided by a divide by 2 counter 1946 yeilding a 1H signal on line 1948 that, ultimately, appears as the H rate output signal. The line 1948 is also carried back to the phase comparator input of the circuit 1940. The filtered error signal input to the voltage controlled oscillator is carried by line 1949 which extends through a transmission gate 1950 which is conducting whenever a multiplex sync is present on the input line 1920'. This is detected by line 1951 which triggers one-shot 1952 that goes high for about 3H pulses before it times out. The output line 1953 of one-shot 1952 is always high whenever the multiplex sync is present.

If the multiplex sync is not present and does not resume after a 3H period, output line 1953 will go low disabling the AND gate 1942 as well as the gate 1950 and will, through inverter 1954, enable another transmission gate 1955 which produces an "artificial" error signal for use by the VCO in maintaining the H rate approximately at the correct frequency until multiplex sync is resumed. A NOR gate 1956 having its inputs connected to the phase comparator's outputs in circuit 1940 provides a lock indicating signal which drives a light emitting diode 1957 when the phase lock loop is not locked up. A signal indicating that sync is correct appears on line 1959, which is one of the qualifiers that is needed before a recording operation sync is permitted to be performed. The signal is generated when the servo is locked and the phase lock loop is locked, the status of these conditions being indicated by a signal provided at the input of the AND gate 1960.

The circuitry shown in FIGS. 31A and 31B illustrate error check logic that is similar in many respects to the error check logic of the existing disc drive circuitry that is used in computer data processing. However, with the present apparatus, additional fault conditions can occur and the error checking logic has been modified and expanded to provide this capability. Referring initially to FIG. 31A, the playing of a picture frame of video information requires two revolutions of the disc pack 75 as previously mentioned and the position of the heads are changed when a seek command is exerted on line 1975. However, since changing the position of the heads from one track to another would provide a discontinuity in the television picture, it is desired that the changing of the head position starts only during the vertical blanking interval. Accordingly, the seek command is timed to start at a specific time with respect to the vertical rate signal that is applied on line 1976 so that a timed start seek command appears on line 1977 that is properly timed with respect to vertical blanking interval. The vertical rate signal is provided by the timing generator circuitry shown in FIG. 30 and the record timing circuitry (FIG. 29A).

Referring to FIG. 31B which illustrates the other section of the error check logic circuitry, this section of the circuitry performs a check to determine if the recording current is behaving as it should, i.e., when it is turned on, it is checked to determine if it is in fact on and, conversely, after it has been turned off, the circuitry checks to see that it is off. It should be appreciated that if the instructed condition was not occurring, then data existing on the disc could be endangered.

More specifically, record current sense line 1978 is applied to the NAND gate 1979 as well as to an inverter 1980 which provides an input to a second NAND gate 1981. A record sequence line 1982 is also connected to the NAND gate 1979 and through an inverter 1983 to the NAND gate 1981. While the line 1978 actually indicates if current is flowing and originates from the record power supplies, the record sequence line 1982 should have a logical low level when current is flowing and a logical high level when it is off. When a strobe occurs on line 1984, one of the NAND gates 1979 and 1981 will provide an active signal on its respective output lines 1986, 1987 that set corresponding FFs 1988 and 1989 which are connected to NOR gate 1990. The NOR gate 1990 provides a signal that signifies conditions are unsafe and that the data on the track may be endangered whenever one of the NOR gate inputs is satisfied. In this regard, FF 1988 will indicate that current is flowing in the recording heads when it should not be and FF 1989 will provide an active signal to NOR gate 1990 when the recording head current has been turned on and no current is flowing. A horizontal rate signal appears on line 1992 and clocks FFs 1993 which produces an output on line 1994 that strobes the NAND gates 1979 and 1981 via connecting line 1984 to determine if the sensed record current is what it should be. In other words, after the record current is shut off, the operation of the FFs 1993 places a high level on line 1994 one horizontal line later to strobe the NAND gates and determine if the current is behaving properly. The strobe signal lasts for one television line and begins one horizontal line after the command has been given. The H rate is used because it provides adequate time for the current to reach its new level after a command has been given.

If an offset condition occurs, which indicates that the heads are mispositioned so that they are not following the center of a track of the disc pack 75, a signal on line 2000 will set a FF 2001, which responds to provide a true signal to a NOR gate 2002. The NOR gate 2002 is responsive to the true signal to provide a select lock on line 2003 which disables the drive because of conditions that could endanger the data and indicates to the disc drive that something is wrong.

The disc drive data interface 151 shown on the block diagram of FIG. 8B is adapted to receive the video data from the encoder 96 and send it to the associated disc pack 75 as well as receive the detected video data from the associated disc pack and send it to the data select switch 128. There are two disc drive data interface circuits that are used to interface the 10 bits of data that are sent to and taken from each disc pack 75 with only one representative interface being shown in FIGS. 46A and 46B. The data received from the encoder 96 appears on lines 2020 and are gated through AND gates 2021 onto the output lines 2022 for recording on the disc pack surfaces. The AND gates 2021 are enabled by a record sequence command on line 2023 originating in the record timing circuitry of FIGS. 29A and 29B. When data is reproduced from the disc pack 75, it appears on lines 2025. This reproduced data is gated through AND gate 2026 onto lines 2027 when the AND gate 2026 is enabled by a high level signal present on line 2028 when a low level signal is present on line 2029 that also comes from the record timing circuitry. When line 2029 is high, the complementary output buffer 2030 produces a low level on line 2028 and a high level on line 2031 which enables AND gates 2032. The enabled AND gates 2032 permit the data being received from the encoder 96 to be transmitted back to the data select switch 128 and following selected playback channel 91 via lines 2027. This condition occurs during the E-to-E and the seek operations during which the signal is processed by both the record and playback electronics, but the recording step is not carried out. The data on lines 2020 is converted by differential amplifier line receivers 2020' from emitter coupled logic having complementary levels to TTL logic before it arrives at AND gates 2021 and, conversely, the data on lines 2027 has been converted by differential amplifier line transmitters 2019 to emitter coupled logic from TTL logic for transmission.

In the disc drives utilized in typical computer processing apparatus, such as the aforementioned Ampex model DM 331 disc drive, the disc spindle motor drive is free running. To provide desired servo control for the disc spindle motor drive, the motor drive circuits have been modified for the unique application in the present apparatus. The operation of the motor driving the disc will now be described in connection with FIG. 27 which is a block diagram illustrating the operation of such circuitry for controlling the driving of the motor in the computer disc drive so that it is locked to vertical sync and correctly positioned relative to the timing so that recording, playback and transfer operations are carried out with the proper timing.

Referring to FIG. 27, a block diagram of the circuitry which operates the drive motor and servo control system is illustrated. The detailed electrical circuitry of the modified Ampex model DM 331 disc drive that carries out the functions that will be generally described with respect to FIG. 27 are contained in FIGS. 32A and 32B which are schematic diagrams of the disc drive phase lock control and FIGS. 45A and 45B which are schematic diagrams of the disc drive motor logic and predriver circuitry which is used during start up of the disc drive motor. Referring to FIG. 27, when the three phase induction motor 2040 for the drive is to be started up it is started using three phase AC power from the power lines 2041 which pass through relays 2042 and power the motor until it has come up to speed. After it has come up to speed, the relay 2042, which is controlled by coil 2043 from disc drive motor run logic circuitry 2044, is switched from the power lines 2041 to the three phase output lines of a switching inverter 2045. The inverter is powered by a DC power supply 2046 through line 2047 with the power supply being connected to the power lines 2041. The positional phase of the motor 2040 is derived from a servo read head 2049 that provides a signal to a preamp 2050 for every revolution of the disc drive, with the output of the preamp 2050 being amplified by amplifier 2051. Decoding circuitry 2052 provides a pulse for the sector 000 (index) mark of the disc when it occurs once during each rotation of the disc pack 75. The pulse appears on line 2053 at the input of a phase detector 2054. The phase of the index pulse is compared with the vertical sync appearing on line 2055 at the input of the detector 2054 and provides an error signal on line 2057 that is phase compensated by a phase compensation network 2058 and then applied to a voltage controlled oscillator 2060 to adjust the frequency and phase of its output in accordance with the error signal. The voltage controlled oscillator 2060 provides a six phase frequency and phase adjusted output which is coupled by line 2087 to control logic circuitry 2061 that drives the 3 phase switching inverter 2045. In this manner, the motor 2040 can be servo controlled so that an associated index position for the driven disc pack is locked to vertical sync that may be derived from either the station reference for playback or a video input signal in the event a recording is being performed.

Turning now to the schematic drawings and particularly FIG. 44B, when the drive motor 2040 is turned on in response to a motor run command on the input line 2065 from the disc drive control circuitry and after it has come up to speed, a signal from the disc drive control circuitry will appear on line 2066. This signal is gated through NAND gate 2067 to actuate a one-shot 2069 which has a time delay of about four seconds. Following the four second delay, a FF 2070 is clocked by the one-shot 2069 and provides a command on line 2071 that turns on the DC power supply 2046 (FIG. 27) providing the power for the switching inverter 2045. The output of the FF 2070 after gating with a power supply verification signal also is applied to line 2072 which triggers a one-shot 2073 that has a delay of about 50 milliseconds. After one-shot 2073 times out, it clocks an FF 2074 that provides a signal on line 2075 to short out a 50 ohm resistor that is in series with the inverter for the purpose of protecting it from transients during the switching period. The shorting signal is provided to the inverter 2045 over line 2068. A signal on line 2072' also provides the command to actuate the relay 2042 (FIG. 27) to change over from the power lines 2041 to the switching inverter 2045. The output line 2075 also extends to yet another one-shot 2076 for triggering it when a signal is placed on the line 2075 by the clocking of FF 2074. One-shot 2076 has a 40 millisecond delay and clocks an FF 2077 that provides a signal on line 2078 that causes the shorting out of a 10 ohm resistor which is series connected to the inverter 2045 (FIG. 27) and thereby performs the same protection function as is performed for the aforementioned 50 ohm resistor. The shorting signal is provided to the inverter 2045 over line 2078'.

Turning to FIG. 45A, the power line phase reference is detected and a representative signal is applied to a line 2080 that is connected to a voltage controlled oscillator 2081 which phase locks its output on line 2082 to the phase of the power line. During change over from the power line 2041 (FIG. 30) to the inverter 2045, the phase locked voltage controlled oscillator 2081 maintains the phase of the voltage drive to the motor provided by the inverter synchronous with the phase of the power line and no substantial disruption occurs. The outputs of the voltage controlled oscillator 2081 and 2060 (see FIG. 32B) are coupled through gating circuitry that selects the appropriate output for application to the following 3 phase logic 2061 in accordance with operating condition of the disc drive system. For example, the signal appearing on line 2082 is at a frequency of 720 Hz (12.times.60 Hz) which is gated through NAND gate 2083, NOR gate 2084 into a ring counter 2085 via line 2086. The ring counter 2085 provides 60 Hz square wave outputs on six lines 2087 that have a 30.degree. phase relationship between them and provides through the following 3 phase logic 2061 signals for phases A, B and C as indicated for driving the switching inverter 2045 shown in FIG. 27. The outputs of the 3 phase logic 2061 are sent to opto-isolaters and provide drive signals for the power switching inverter 2045. The NAND gate 2083 gates the output of the oscillator 2081 into the ring counter 2085 when a high signal is present on line 2090. When the line 2090 is low, inverter 2091 enables a NAND gate 2092 to gate through pulses from line 2093 which are provided by the voltage controlled oscillator 2060 (see FIG. 32B) at a frequency of 720 Hz.

Referring to FIG. 32B, the voltage controlled oscillator 2060 and frequency/phase detector 2054 are included within a single integrated circuit component which has the input reference signal present on line 2055 as well as the feedback signal on line 2053 for use by the detector 2054. An error output signal from the detector 2054 is coupled by line 2057 to a storage capacitor 2095, and through an impedance matching operational amplifier 2096, is coupled to the phase lead compensation networks 2058. The network 2058 conditions the error signal generated by detector 2054 for application to the oscillator 2060. The reference and feedback signals on lines 2055 and 2053 that are used by the frequency/phase detector 2054 are produced by circuitry shown in FIG. 32A which which are operatively associated with sector 000 (index) pulses applied to line 2100. The index pulses are shaped by a voltage translator 2101 to produce the narrow pulses on line 2053 at the correct voltage levels for application to the detector 2054. Similarly, the reference vertical pulses appear on line 2103 and are shaped by a voltage translator 2104 and are applied to a one-shot 2105 which cooperates with a following one-shot 2106 to inhibit a second pulse from occurring for a time period of about 8 milliseconds. The one-shot 2106 has its output coupled to line 2055 that provides the reference input to the detector 2054. The one-shot 2106 has a 5 microsecond period and its second output is coupled to control a switch 2107 to turn it on for 5 microseconds during every vertical pulse. This produces a 5 microsecond offset which improves the perfomance of the servo by removing jitter which is present when the sector 000 (index) pulse and reference vertical pulse are coincident. The line 2108 extends to the capacitor 2095 (FIG. 32B) in the phase comparator output line 2057 that controls the oscillator 2060. The one-shot 2106 has output line 2055 also connected to another one-shot 2110 which has a two millisecond period and produces an output on line 2111 which is differentiated by differentiator 2112 and applied to NAND gates 2113 and 2114 via inverter 2116 and line 2115. A one-shot 2117 triggered by the sector 000 (index) pulse produces a 4 millisecond window, i.e., a high level on line 2118 to the NAND gate 2113 as well as a low level on line 2119 to the NAND gate 2114. When the pulse appearing on line 2115 first falls within the 4 millisecond window generated by the one-shot 2117, indicating that the two signals are particularly close to being phase locked, then NAND gate 2113 will set a latch 2120, which activates a one-shot 2121 whose output on line 2122 is applied to NOR gate 2123. The NOR gate 2123 responds to close a switch 2124, which applies a voltage from the voltage divider 2125 onto the line 2108 to the capacitor 2095 (FIG. 32B) and thereby changes the time constant and gain characteristics of the control loop to speed up the locking procedure. The one-shot 2121 closes the switch 2124 for a period of about 10 milliseconds.

The output line 2055 from the one-shot 2106 also extends to the trigger input of a one-shot 2127 that has a 15 microsecond. A differentiator 2128 is coupled to the output of the one-shot 2127 and produces a narrow pulse on the trailing edge of the signal generated by the one-shot 2127. The narrow pulse is applied to one input of a NAND gate 2129, the other input of which is supplied by a one-shot 2131 that is triggered by the sector 000 (index) pulse from line 2053. The one-shot 2131 produces a 30 microsecond window which inhibits the pulse on line 2130 from passing the NAND gate 2129. When phase lock is achieved within plus or minus 15 microseconds, a one-shot 2132, which has a relatively long one second period, will time out producing a low signal on line 2133. This indicates that the servo is locked up, i.e., the motor is being timed with respect to reference vertical as is desired.