CROSS-REFERENCES TO RELATED PATENTS AND APPLICATIONS
U.S. Pat. No. 3,821,494 granted June 28, 1974 to Besseyre.
U.S. Pat. application, entitled "Digital Voice Signaling With Digital Echo Detection and Voice Activity Compression Used to Cancel Echo", by P. Abramson and U. Appel, Ser. No. 560,423, filed Mar. 20, 1975 assigned to the assignee of the present application.
Application Ser. No. 537,211 by H. G. Markey et al, filed Dec. 30, 1974, entitled "Access Method and Station Apparatus for Compressed Handling of Digital Voice and Data Signals Relative to a High Speed TDMA Facility" and assigned to the assignee of this application.
BACKGROUND OF THE INVENTION
1. Field of the Invention
The invention relates to digital voice signal transmission circuits employing delta-modulation coding and decoding.
2. Statement of the Problem
In long haul transmission circuits for telephone communication the transmission channel may be intermittently interrupted, either for echo cancellation purposes or to permit time multiplex handling of other transmissions. It is well known, however, that total disconnection causes an unnatural and disconcerting reception effect of total silence giving the listener the impression that the connection has been broken.
In certain analog systems the problem is avoided by loss insertion. Instead of completely disconnecting the channel a high loss (e.g. 50 db) is switched into the transmission path. However, it is believed that this technique has not hitherto been applicable to a digital voice signal channel employing delta-modulation coding.
3. Environmental Background
The present invention is effective and convenient to use in the virtual connection TDM communication system described and claimed in the above cross-referenced patent application by Markey et al; the disclosure of said application being incorporated herein by this reference. In said system voice activity compression (VAC) processing and signaling circuits distinguish pauses in speech and momentarily "interrupt" a virtual transmission connection between a local delta-modulation coder serving a "talking" subscriber and a remote delta-modulation decoder serving a remote "listening" subscriber. This permits the channel used in the connection to be momentarily assigned to other virtual connections.
The delta-modulation coding circuits in the above system operate in concert with a "digital" integrator circuit receiving adaptive step-size integrand arguments. The integrator is used for companding and activity discrimination. In the code generation operation present and predicted digital representations of the analog (voice) input are recurrently compared. The predicted value is based on the previous binary sequence of delta-modulation code output. The code bit (difference) is quantized as a step-size integrand value proportioned to the latest representational value of the integration result. The step-size value is added digitally to or subtracted from the latest integration value according to the delta-modulation comparison result.
When the representational value of the step-size integrand falls in a range above a predetermined threshold activity level (e.g. -33 dbm) a voice activity indication is produced which is used by the VAC processing circuits to control allocation of a shared transmission channel for completing the virtual connection to the remote demodulator and listener. Since the step size is proportional to the representational value it can be used in the logical detection of threshold activity. If the step-size value is below the activity threshold the voice transmission channel "connection" is interrupted (allocated to another virtual connection). Information indicating the status of transmission channel utilization is recurrently transmitted via the VAC signaling facility (over another channel) for use in controlling demultiplexing.
Aspects of the variable step-size generation and activity detection are variously described in Vol. 16, No. 11 of the IBM Technical Disclosure Bulletin at Pages 3702-3705 and 3709-3710.
SUMMARY OF THE INVENTION
According to the present invention whenever the transmission channel for delta-modulated voice is momentarily interrupted (e.g. by operation of VAC processing circuits as explained above) information remotely indicating the interruption (e.g. the channel use information transferred via the VAC signaling facility) is used remotely to cause insertion of a special pseudo-random digital signal at the demodulator. The inserted signal comprises an essentially random bit sequence having the property that no more than n of any consecutive bits are equal; where n is a number associated with the "threshold" step size. The insertion effectively reduces the output of the remote demodulator to a low volume noise signal below the threshold level associated with the threshold step size.
A similar bit sequence may be inserted on the transmission side of an "unshared" delta-modulation channel to produce loss insertion effects in transmission and/or to "reset" an integrator associated with a delta-modulation coder.
The foregoing and other features, aspects, objects and advantages of the present invention will be appreciated from the following detailed description.
DETAILED DESCRIPTION OF THE DRAWINGS
FIG. 1 schematically illustrates the system orientation and utilization of the invention for delta-modulation transmission via a shared (interruptable) voice transmission channel;
FIG. 2 schematically illustrates the logic for generating the pseudo-noise "loss insertion" signal according to the invention for the exemplary case n = 3;
FIG. 3 illustrates timing (clock) signals for the logic circuit of FIG. 2; and
FIG. 4 illustrates the "transmit side" application of the subject invention for echo suppression relative to an unshared delta-modulation transmission channel.
Referring to FIG. 1 in a typical application for delta-modulation transmission over a shared channel, for instance in the system disclosed in the above-referenced and herewith incorporated disclosure by Markey et al., analog voice signals Vin are first delta-modulation coded in modulation circuits 1. A digital integrator 2 having a variable step-size argument provides companded digital representations tracking the volume intensity (magnitude) of Vin. The delta-modulation code is transmitted over virtually allocated channels of a transmission facility 3. In the system disclosed by Markey et al the transmission facility includes common control traffic-dependent switching circuits and a time division multiplex (TDM) multiple access satellite communication segment.
The coding section 4 of circuits 1 compares the digital output of analog-to-digital converting circuit 6 to a predicted value g(V) which is an uncompanded digital function of the previous coding history. The difference .DELTA.V is used both as the present delta-modulation code value (for signaling at L) and to control addition or subtraction of the step-size factor f(V) to the latest integration result. The representational value of f(V) is varied as the volume level represented by g(V) varies through predefined range limits.
At the remote demodulation station 7 the signal from the source line L is extracted from the shared (virtual) channel and passed via associated line L' to demodulation circuits 8 which provide analog signal Vout corresponding to Vin.
Voice Activity Compression (VAC) processing circuits 9, operating relative to circuits 2 at the transmitting station, determine when f(V) represents a predetermined noise (inactive) level and disconnect source line L from facility 3 (e.g. by transfer of schematically shown switch contact 10 to another virtual connecton, not shown). VAC mask information is sent over another channel to VAC receiving circuits 11 to indicate momentary "interruption" of the channel from L to L' (as suggested above the interrupted channel may then be allocated to another virtual connection not shown). Circuits 8 are connected (e.g. via switch circuit 12 shown schematically) to loss signal generator 13 which supplied binary signals in a pseudo-random sequence (long recurrence interval) subject to the logical constraint that not more than n of any consecutive bits have equal binary values; where n is an integer associated with the integrand step size used in the integration section 14 of demodulation circuit 8. Upon input to circuits 8 this sequence reduces the step size value progressively to a range below a threshold value and causes Vout to appear as an audibly reproducible low volume noise-like sound readily distinguishable from a dead line.
With the possible exception of circuits 13 all circuits in the foregoing environmental/applicational system are considered fully disclosed in the above-referenced Besseyre Patent, Markey et al. application and IBM Technical Disclosure Bulletin references.
FIG. 2 shows the logic 17, for constraining the pseudo-random sequence output of circuit 18 in accordance with the invention, for the case n = 3 (maximum number of consecutive equal bits is 3 for threshold activity). The pseudo-random input signal supplied by conventional pseudo-random signal generator 18, preferably a signal having a long recurrence time characteristic, operates flip-flop FF1. The alternately effective outputs Q and Q of FF1 are gated with respective outputs of a second flip-flop FF2 to condition successive inputs to 3-stage shift register 20. NOR circuit 21 and AND circuit 22 logically process outputs of register 20 and FF1 as described below and condition NOR circuit 23 to effectively cause inversion of the fourth bit of any sequence of four equal bits coincidentally manifested by the shift register and FF1. The extension for other values of n is considered obvious from the following discussion.
At clock times clk2 (FIG. 3) FF2 is preset to the complement of its "clear" state. In this state output Q of FF2 is enabling and Q is disabling. At clock times clk 1 register 20 is shifted and the state of FF1 is effectively shifted into the first (A) stage of register 20 if FF2 is in the preset state. FF1 is simultaneously reset to its cleared state (Q enabling). If input line 25 thereafter receives a signal representing a 1 bit the set (D) input of FF1 is pulsed to invert the state of FF1 (to Q enabling) before the next shift pulse clk1.
AND circuit 22 responds to the coincident presence of all 1's in register 20 and FF1 (representing a sequence of four consecutive 1 bit inputs at 25) to condition NOR circuit 23 to activate the "clear" input of FF2 setting FF2 to its clear state. This reversely conditions the logic between FF1 and register 20 to shift the complement of the state of FF1 (e.g. a 0) into register 20 at the next clk1.
Similarly NOR 22 responds to coincidence of all 0's in register 20 and FF1 (representing an input sequence of four consecutive 0 bits) to clear FF2 and invert the shift input from FF1 to register 20.
Thus, the fourth bit in any sequence of 4 equal bits received at 25 is effectively inverted in shift transit to register 20.
FIG. 4 indicates that the "loss" signal for reducing received ouput to low volume noise level can be inserted at the transmission side of an "unshared" transmission channel to produce the effect of noise level attenuation in the transmission path. The insertion may be used for echo suppression. Thus, echo control circuits 101 responsive to echo in the signal processed by delta-modulator circuits 4a may be used to control operation of contact L into the position for connecting output of loss generator circuit 17a (identical to circuit 17, FIG. 1) into the transmission path.
While the invention has been particularly shown and described with reference to preferred embodiments thereof, it will be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the spirit and scope of the invention.