Background of the Invention
1. Field of the Invention
The present invention relates to a digital audio system and, more particularly, to a signal processor for providing both sub-band coding and decoding in a digital audio system.
Korean Patent Application No. 93-5900 is incorporated herein by reference for all purposes.
2. Brief Discussion of Related Art
FIG. 1 shows an encoder for performing data encoding in a conventional universal sub-band coding system developed through a joint effort by the CCETT in France, the IRT in Germany and Philips in the Netherlands. The encoding block of the sub-band coding system is composed of a scale sample generator 10, a quantizer 11, a scale factor generator 12 and a formatter 13.
The operation of the sub-band coding system utilizing the configuration discussed above will now be briefly described.
It will be noted that each of 32 sub-bands of a frame signal is represented by 12 samples, thus 384 sub-band samples are included in each frame. The sample with the highest amplitude of each sub-band is used to determine a scale factor for every respective sub-band. Then, the samples are scaled by the respective scale factor in order for them to be normalized with respect to the value representing the highest amplitude.
One part of the encoding algorithm will have to supply an allocation information table, e.g., information regarding word length for each sub-band in the frame. It will be noted that it is possible to generate the allocation information table in various ways, the simplest way being to use the allocation information table in which all entries are fixed. However, this information may also change with every frame.
The samples are then quantized to reduce the number of bits of the samples to the number specified for the sub-band in the allocation information table and, then, the synchronization and system information is added to produce a formatted frame.
FIG. 2 shows a decoding block of a general sub-band coding system employed in the decoder according to the CCETT/IRT/Philips technique. The decoding block of the sub-band coding system includes a deformatter 20, an allocation information RAM 21, a scale factor generator 22, a dequantizer 23, a multiplier 24, a decoder controller 25, a multiplication factor generator 26 and a demultiplexer 27.
The decoding operation of the sub-band coding system according to the above configuration will now be described. It will be noted that the decoding process is always performed in the same manner. The essential information is decoded using the scale factor and the sub-band samples. All other information is included so as to permit the decoding to be performed correctly. The decoding process is repeated for every frame. After the reception of the synchronization and system information, the allocation information is used in storing the scale factor array correctly in generator 22.
The scale factor array is an array with two columns, 6 bits wide, and 32 rows. Each column corresponds to one of the output channels and each row corresponds to one of the 32 sub-bands. The received scale factors are stored in this array. The location of a scale factor which has not been transferred becomes 1111111, indicating that the output samples for that sub-band are all "0".
In stereo and bilingual modes, the scale factors for the two channels are stored in the two columns of the array. In mono mode, the same scale factor is stored in both columns of the array.
It will be noted that the scale factors are used as an index into a fixed table. This table contains the multiplication factors by which the sub-band samples have to be multiplied before they are sent to join transform implemented in the encoder. These scale factors are then used for both encoding and decoding operations.
In other words, the encoding block or circuit is independent of and separated from the decoding block or circuit in the conventional general sub-band coding system, making the configuration of the player/recorder very complex.
Summary of the Invention
The principal object of the present invention is to provide a signal processor for a sub-band coding system whose circuit configuration is simplified by combining the encoding and decoding function employed by a conventional general sub-band coding system in one single signal processor block.
This and other objects, features and advantages according to the present invention are provided by an encoding and decoding system including:
selecting means for receiving various data, i.e., sample data, allocation information, synchronization and system information at a time of encoding, changing such data into appropriate forms and outputting same;
storing means for temporarily storing the various data received from the selecting means;
operating means for scaling and quantizing sample data and for dequantizing and descaling encoded data;
classifying means for outputting as encoded data operated data, allocation information input to the operating means and system information, in accordance with the frame format of a general sub-band coding system, and for receiving and classifying the encoded data again into system information, allocation information and operated data; and
a control signal generating means for generating control signals for controlling encoding and decoding accomplished cooperatively the selecting, storing, operating and classifying means.
Brief Description of the Drawings
The above objects and advantages of the present invention will become more apparent by describing in detail a preferred embodiment thereof with reference to the attached drawings in which:
FIG. 1 is a block diagram of an encoder for performing data encoding in a conventional general sub-band coding system;
FIG. 2 is a block diagram of an decoder for performing data decoding in a conventional general sub-band coding system;
FIG. 3 is a high level block diagram of a circuit for performing data encoding and decoding in a general sub-band coding system according to the present invention;
FIG. 4 is a low level block diagram of an embodiment of the multiplexer shown in FIG. 3;
FIG. 5 is a low level block diagram of an embodiment of the operator shown in FIG. 3;
FIG. 6 is a detailed block diagram of an embodiment of the maximum absolute value calculator shown in FIG. 5;
FIG. 7 is an operational flowchart which is useful in understanding the process for obtaining the index value from the maximum absolute value;
FIG. 8 is a table showing quantized coefficients;
FIG. 9 is a table showing dequantized coefficients;
FIG. 10 is a table showing K values; and
FIG. 11 is a table showing index values and scale factors.
Detailed Description of the Preferred Embodiments
Referring to the accompanying drawings, a general encoding and decoding process will first be described before explaining the signal processor for a sub-band coding system according to the present invention. More specifically, the encoding and decoding method known as the Masking pattern adapted Universal Sub-band Integrated Coding And Multiplexing (MUSICAM) regime, which was developed through the cooperative efforts of CCETT, IRT, Philips and Matsushita, will be described.
The encoding process begins after the collection of 12 sample data for each of 32 sub-bands with a step whereby the maximum absolute value .vertline.S.sub.max .vertline. is obtained and then SF.sub.i is obtained. When S represents sample data, S' represents encoded data, and SF.sub.i represents the scale factor index, the quantities are related according to equation (1) as follows: ##EQU1##
Next, the values in equation (1) are converted into decibel (dB) values by taking the log of both sides of equation (1), as follows:
Since the value of SF.sub.i (dB) has been already obtained in the first step, if S is converted into the value of S(dB), the value of S'(dB) can be obtained by an arithmetic operation.
Then, S' is obtained by finding the antilog of S'(dB), and S' is quantized, in accordance with the following equation to become S".
It will be appreciated that the value of a coefficient pair (A,B) is subject to change in accordance with the allocation information and a quantization coefficient q.
Preferably, if the value of the coefficient (A, B) is expressed as a binary digit, S" of the coefficient (A, B) is obtained by obtaining any of the coded bits expressed as binary digits, inverting the coded bits and then processing the data with a multiplier and an adder.
The quantization process is advantageously finalized by taking the number of bits, starting from the most significant bit (MSB) of the value S", equal to the number of the bits corresponding to the allocation information.
Finally, the scale factor index, allocation information, quantized data and system information obtained from the above steps are adjusted in accordance with the selected frame format of the sub-band coding system and then output as 32 bit data.
Preferably, the decoding process starts with a step for receiving encoded data from which the quantized data S" are dequantized. The operation to dequantize the data is performed according to the following relations: ##EQU2## where C, D and K are coefficients, and C equals the sum of 1 plus K, i.e., C=1+K. The dequantized coefficient table of FIG. 9, which will be discussed in greater detail below, indicates the binary expressions of D and K and coefficients thereof.
Preferably, descaling is then performed in order to convert the dequantized result back into the original data value. Since the data is divided by SF.sub.i at the time of scaling during the encoding process, the data is multiplied by SF.sub.i at the time of descaling. That is to say,
Taking the log of both sides of equation (7) yields dB values as follows:
It will be appreciated that if the scale factor index value among the encoded data is converted into a dB value, the value of SF.sub.i (dB) in equation (8) becomes a dB value of an integer form. When the original value of S is obtained by applying an antilogarithmic expression to the S(dB) value obtained by addition (refer to equation (2)), the descaling operation is finalized. Then, the sample data S obtained by equation (8) is output in the same order as that of the data input at the time of encoding.
Now, the signal processor for performing the above encoding and decoding processes in the sub-band coding system according to the present invention will be explained.
FIG. 3 is a block diagram of the encoding and decoding system according to the present invention, wherein the system preferably includes a multiplexer 100, a storing means 101, e.g., a memory device, an operator 103, a formatter and deformatter 102, acting as a classifying means, all of which are mutually interconnected, and a controller 104, which advantageously is connected to control the respective components. The system configuration shown FIG. 3 advantageously performs encoding and decoding in one signal processing system, eliminating the need for providing or installing the respective separate circuits, in contrast to the conventional system.
The functions of the respective components are as follows. The multiplexer 100 receives various data, ie., sample data, allocation information, synchronization and system information, and modifies them into appropriate forms, e.g., serial to parallel conversion, and then outputs such modified data to other blocks, for performing the encoding operation.
The storing means 101 advantageously is used for temporarily storing the various data received from the multiplexer 100. It will be appreciated that the data are stored to permit the sample data to flow smoothly during scaling and quantizing and to prevent damage to the sample data. Preferably, at least the 384 samples of the sample data making up one frame should be stored. It will be noted that the storing means 101 also stores allocation information and scale factor index values.
The operator 103 includes a multiplier, an adder, a subtractor, a shift register and a comparator, for selectively scaling and quantizing the sample data or dequantizing and descaling the encoded data, as explained in greater detail below.
The formatter and deformatter 102, advantageously outputs the encoded data by adjusting the operated data and the allocation and system information input to the operator 103 in accordance with the frame format of the general sub-band coding system, or receives the encoded data and again classifies it into the system information, allocation information and operated data.
Preferably, controller 104 determines whether all data blocks are encoded or decoded and then generates the control signals appropriate for initiating the required functions. The controller 104 may advantageously be constructed to permit linkage with an external microprocessor.
FIG. 4 is a block diagram of an embodiment of the multiplexer shown in FIG. 3, wherein the multiplexer advantageously includes a register 110 for inputting allocation information in response to an enabling signal from the controller 104, a series/parallel register 111 receiving 12 sample data in series in response to the enabling signal and generating paralleled sample data, a parallel/series register 112 receiving the parallel output signals in response to the enabling signal and for generating serial signals, and a register 113 for receiving synchronization and system information in response to the enabling signal. Preferably, the multiplexer 100 receives 12 samples for each sub-band.
The information output by the multiplexer 100 is output to the storing means 101, formatter and deformatter 102 and operator 103 via interconnecting buses.
FIG. 5 is a block diagram of the operator 103 shown in FIG. 3, wherein the operator 103 advantageously includes a maximum absolute value calculator 200, a log means 201, an integer maker 202, an antilog means 203, an (A, B, D and K) coefficient generator 204, a multiplier 205, an adder 206, a buffer 207 and buses BUS1 and BUS2 for signal transmission.
During the encoding operation performed by the operator 103 shown in FIG. 5, the maximum absolute value calculator 200 calculates the maximum absolute value .vertline.S.sub.max .vertline. among 12 sample data for each sub-band. The log means 201 receives the maximum absolute value .vertline.S.sub.max .vertline. from the maximum absolute value calculator 200 and obtains the value SF.sub.i (dB). The integer maker 202 receives SF.sub.i (dB) and then obtains scale factor index value SF.sub.i. That is to say, the integer maker 202 includes the indices corresponding to the value of SF.sub.i (dB). Advantageously, the log means 201, e.g., a log circuit, also receives the respective sample data and then obtains the value of S(dB). The adder 206 performs the operation of the equation (2) and thereby obtains the value of S'(dB). The antilog means 203 receives the value of S'(dB) and then obtains the value of S'. The multiplier 205, adder 206 and coefficient generator 204 perform the operations of equation (3) to thereby generate S", which is the quantized value of S'. The buffer 207 stores the value of S" and later outputs it, starting from the uppermost bit and continuing for the number of bits corresponding to the allocation information.
Preferably, the decoding operation performed by the operator 103 shown in FIG. 5, starts with storing the value of S" in buffer 207. The multiplier 205, adder 206 and coefficient generator 204 perform the operations of equations (4)-(6) to thereby obtain the value of S' from the value of S". The log means 201 receives the value of S' and then obtains the value of S'(dB). The integer maker 202 receives the value of SF.sub.i and then obtains the value of SF.sub.i (dB). The adder 206 adds the values S'(dB) and SF.sub.i (dB), which antilog 203 receives as the value of S(dB) and then obtains the value of S.
It will be appreciated that both encoding and decoding processes can be advantageously performed using one single circuit, as shown in FIG. 5.
FIG. 6 is a block diagram of the maximum absolute value calculator shown in FIG. 5, wherein the maximum absolute value calculator 200 includes an absolute value calculator 300 receiving sample data S and calculating the absolute value thereof, a register 301 for storing and reproducing the output signal of the absolute value calculator 300 in response to a clock signal CLK ,a comparator 302 receiving and comparing the output signal of the register 301 with a previous maximum absolute value and then generating a flag signal if the output signal of the register 301 is greater than the previous maximum absolute value, and a register 303 for storing the output signal of the register 301 in response to the output signal of the comparator 302 and generating the current maximum absolute value. The maximum absolute value calculator having the above-described configuration outputs the maximum absolute value .vertline.S.sub.max .vertline. of the sample datas received thereby.
FIG. 7 is an operational flow chart for explaining the process for generating an index value from the maximum absolute value .vertline.S.sub.max .vertline.. The flow chart illustrates steps for obtaining the value of the digital bit rate (DBR) by calculating the logarithm of the maximum absolute value .vertline.S.sub.max .vertline. and converting it into a dB value during step S1, obtaining the value of a digital bit integer (DBI) by adopting an integer value to the DBR value obtained in step S1 during step S2, determining whether or not the DBI value obtained in step S2 is less than a -118 dB value during step S3, adopting the dB value as -118 DB if the DBI value is greater than -118 dB in step S3 during step S4, adding 2 to the DBI value if the DBI value is an even number and 1 if the DBI value is an odd number with respect to the case that the DBI value is less than -118 DB in step S3 during step S5 and providing the finally calculated dB value after performing the processes of steps S4 and S5 as the index value corresponding to the DB value during step S6. By performing such operations, the dB value of the maximum absolute value .vertline.S.sub.max .vertline. is advantageously calculated. The dB value preferably is between +6 and -118 and the values are spaced apart by 2dB. The operation shown in FIG. 7 can advantageously be performed by a microprocessor.
Hereinafter, an operating description of the system will be made using exemplary values to promote the understanding of the present invention, starting with the encoding process.
Assume that the maximum absolute value .vertline.S.sub.max .vertline. among 12 samples is 0.480000000 and one value among the sample data is 0.375000000.
First, the maximum absolute value .vertline.S.sub.max .vertline. is converted to a dB value to thereby obtain a signal value DBR and this value is made an integer value to then output a signal DBI. The operation for obtaining a scale factor index is shown in FIG. 7 and the index is searched by using the table shown in FIG. 11.
DBR=-6.3752
DBI=-7
Since the above value is an odd number, 1 is added. Then, the index becomes -6; that is, dB=-6.
Scaling is then performed, e.g., normalization is performed to produce a value between -1 and 1.
Then, a quantization step is performed by using the coefficient values A and B shown in the table of FIG. 8.
In the exemplary case where the quantization coefficient q equals 4, A is 0.9375 and B is -0.0625. Therefore,
where 0.6390 is expressed as 0.101000111 in binary digits.
Since q equals 4, if only 4 bits are taken starting from the most significant bit, S" becomes 0.625.
Hereinafter, the decoding process will be described. First, dequantization is performed. The operation of dequantization is performed after obtaining the values D and K by using the tables shown in FIGS. 9 and 10, respectively.
S"=0.625
S'=(S"+D)+K(S"+D)
In the exemplary case, where the quantization coefficient q equals 4, D is 0.125 and K is 0.066666667. Therefore, ##EQU4##
Preferably, descaling is then performed. As with the scaling operation, the descaling operation is also performed using the table shown in FIG. 11.
In other words, it will be appreciated that unencoded input sample data value 0.375 and decoded value 0.4 are close approximation values to one another. The encoding and decoding of the sub-band coding system advantageously is performed by such operations as described above. Therefore, the signal processor for the sub-band coding system according to the present invention provides an effective circuit configuration whereby only one single circuit configuration performs both operations, making separate configurations for encoding and decoding unnecessary and, thus, simplifying the circuitry of the overall system.
Other modifications and variations to the invention will be apparent to those skilled in the art from the foregoing disclosure and teachings. Thus, while only certain embodiments of the invention have been specifically described herein, it will be apparent that numerous modifications may be made thereto without departing from the spirit and scope of the invention.