JP2988825B2 - Variable length encoding method and apparatus using small memory - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and an apparatus for encoding a digital signal into a variable length code (VLC), and more particularly, to a method for encoding a signal source code having an original length of 18 bits. The variable length code lookup table used to encode the variable length code is stored in a 16-bit read-only-memory (ROM), and is necessary to store the variable length code lookup table. The present invention relates to a variable-length encoder based on a new method that saves memory space.

[0002]

2. Description of the Related Art A digital signal transmission system usually includes a quantizer, a predictor, an encoder, a buffer, a control circuit, and the like. The term "encoder" as used in such a field may refer to the entire transmission system or may refer only to the coding operation of a coding circuit within the transmission system. Here, this term means the latter.

Since the available bandwidth in the transmission channel is limited, it is desirable to compress the transmitted data as much as possible. Various methods have been developed to increase the efficiency of such data compression while maintaining the quality of the transmitted video.

One of such compression methods is variable length coding. Variable length coding is a statistical coding technique that assigns codewords to the source code to be coded, where short codewords are assigned to data that occurs frequently and data that is relatively infrequent A technique for assigning long code words.

[0005] According to the MPEG and CCITT standard recommendation, VLC has a maximum run length of 63 and -127.
And a code word of up to 14 bits representing a level in the range from to 127, and the four most significant bits (MSBs) representing the length of the code word.
The result is a VLC with a maximum length of 18 bits.
As used herein, the term "run-length" refers to the number of runs having consecutive zero values, and the non-zero value immediately following a consecutive zero is referred to as a level. The run length and level data are encoded using a look-up table as shown in Table 1. In Table 1, which shows only the combinations of run lengths and levels that are mainly used, the bit "s" at the end of the codeword is the sign of the level, where 0 represents a positive number and 1 represents a negative number. That is, if the non-zero value after the consecutive zeros is a positive number, a zero is placed at the “s” position.
Is satisfied, and if the value is a negative number, 1 is satisfied. As shown in Table 1, the code word has a maximum length of 14 bits including the sign of the level, and has a maximum length of 4 bits (not shown in Table 1) representing the length of the code word. 1
This forms an 8-bit VLC.

[0006] A look-up table as shown in Table 1
A memory with bits of codeword length data, eg, R
Stored in OM. The encoding process is performed by addressing a look-up table using run lengths and levels. The codeword and codeword length data are stored as contents of a look-up table and are sequentially output to the transmission channel at a constant data rate.

[0007]

[Table 1]

[0008]

However, in the variable length coding method described above, MPEG and CCITT are used.
Since a large-scale memory such as a 24-bit ROM is required to store an 18-bit VLC lookup table according to the standard recommendation, the system is expensive.

If a 16-bit ROM is used instead of a 24-bit ROM, the VLC look-up table must be read twice in order to read an 18-bit VLC. there were.

Accordingly, it is an object of the present invention to provide an encoding method and apparatus capable of optimizing a memory space required for storing VLC look-up table data when encoding a digital signal into VLC. .

Another object of the present invention is to provide MPEG and CC
An object of the present invention is to provide a method and apparatus for storing ITT standard VLC look-up table data in a 16-bit ROM.

[0012]

SUMMARY OF THE INVENTION It is an object of the present invention to provide a method of encoding a digital source code into a variable length code for transmission of the digital source code within a microprocessor. Converting the source code into a suitable format for accessing the variable length code table data stored in the 16-bit memory; and determining the length of the variable length code read from the variable length code table And transmitting the processed variable length code of the determined length.

[0013]

FIG. 1 shows a microprocessor 20 and R
A schematic diagram of the VLC encoder 100 including the OM 30 is shown.

Since the variable length encoding method can efficiently compress data without loss, the DCT (discrete cosine transform) is used in various digital video processing applications.
After that, the DC and AC transform coefficients obtained by quantization are used to encode and compress (Ming-Ti
ng Sun and other “An Entropy Coding System for DigitalHD
TV Applications ”, IEEE transactions on circuits an
d Systems for VideoTechnology, 1, No. 1 (March, 1991)).

In such a digital video processing application field, each block composed of 8 × 8 elements is converted from a spatial domain to a frequency domain mainly using DCT.
Thereafter, the 8 × 8 coefficient array obtained by the DCT operation is quantized to be an array of 8 × 8 quantized coefficients. Generally, the number of quantized coefficients having non-zero values is very small, so this compression method is relatively practical.

In the 8 × 8 array of quantized coefficients, the first left coefficient of the first upper rank is referred to as a DC coefficient, and the rest are referred to as AC coefficients.

The quantized DC coefficient is obtained by a differential pulse code modulation scheme DPCM (differential pulse code modulatio).
n) can be encoded without any loss.

The quantized AC coefficients are first scanned in a zigzag order to increase coding efficiency. Thereafter, the AC coefficients are encoded using the run length and level data. Using the zigzag scanning method, 0
This use is desired because it has fewer runs and can be encoded into shorter VLCs.

The video signal processor 10 receives a video signal as an input, processes the video signal as described above, and outputs a zigzag scanned and quantized AC coefficient to the microprocessor 20. Microprocessor 20
Reads the quantized AC coefficient from the video signal processor 10 as a signal source code.

Thereafter, the microprocessor 20 performs the function of determining the run length and level of the input source code and storing the code of the determined level therein. Thereafter, the microprocessor 20 addresses the VLC lookup table stored in the ROM 30 with the determined run length and level, and processes the VLC read from the VLC lookup table according to the present invention.

In a preferred embodiment of the present invention, a 16-bit ROM is used to store the VLC lookup table. As suggested by MPEG and CCITT, VLC can be up to 18 to represent up to 63 run lengths and levels ranging from -127 to 127.
It has a bit length. However, in the preferred embodiment of the present invention, the VLC is stored in a 16-bit ROM, so that the memory space can be reduced as described later with reference to FIG.

FIG. 2 shows that VLC is stored in ROM 3 according to the present invention.
The method of storing at 0 is shown. The table shown in FIG.
Does not include level code data. This code data is shown in FIG.
And a method described later with reference to FIG.

In FIG. 2, when n is an integer from 1 to 13, the n least significant bits LBS (l
The east significant bit includes a codeword having a length of n bits, where the four most significant bits (MSB) of part A or the three most significant bits of part B are codeword lengths.
n ". In the binary system, four bits can be used to represent up to 16 pieces of data (i.e., from" 0000 "to" 1111 "), but in the preferred embodiment of the present invention, each represents 1 to Only 13 data representing the codeword length of the VLC up to 13 are needed, so that the memory portion used to represent the codeword length is reduced, so that the look-up table is stored in the ROM 30. In a preferred embodiment of the present invention, a 4-bit storage capacity capable of storing or representing up to 16 data is shown in part A of FIG. In this case, only the twelve data are used, that is, '0000', '1101', '11
10 'and' 1111 'are not used. To reduce the space required for VLC in accordance with the present invention, as shown in Part A, 4 bits are allocated to represent a codeword equal to or shorter than 12 bits and B
As shown in the section, three bits, all ones, are allocated to represent a codeword having a length of 13 bits. As can be seen from part A of FIG. 2, since there is no code word length starting with three consecutive 1s, a code word having a length of up to 12 bits and a 13-bit length represented by '111' are used. There is no confusion between having a code word.

According to such a method, if the code word length is equal to or shorter than 12 bits,
The number of bits required to represent VLC is equal to or less than 16, and if the codeword is 13 bits long, the total number of bits required to represent VLC is 16
Becomes By such a method, MPEG and CCIT
The 18-bit VLC according to the T standard recommendation can be represented by 16 bits by processing the codeword code outside the ROM 30. As a result, all the information shown in Table 1 can be represented by only a 16-bit memory space as shown in FIG.

FIGS. 3 and 4 show a 16-bit ROM 30.
4 is a flowchart illustrating a method of encoding a signal source code into a VLC using VLC lookup table data stored in the VLC. In step S1, the microprocessor 20 reads the AC coefficients quantized by zigzag scanning from the video signal processor 10 as signal source codes. In step S2, the microprocessor 20 determines the run length and level of the source code to be encoded. In step S3, the code of the determined level is stored in a buffer in the microprocessor 20.

At step S4, the microprocessor 20 stores the ROM in the determined run length and level.
30 is accessed, and the 16-bit VL of the accessed address is accessed.
Read C.

At step S5, microprocessor 20 examines the three most significant bits MSB of VLC. If the three bits representing the codeword length are all 1, the codeword length is determined as 13 bits in step S7. If not, the codeword length is determined by decoding the four MSBs in step S8.

After the codeword length is determined, transmission of the codeword having the determined length begins. Step S9
, The microprocessor 20 performs the bit count (bit c
ount) is set to 1. In steps S10, S11 and S12, each codeword is transmitted to the transmission channel through a serial port built into microprocessor 20 until the bit count is equal to or greater than the length of the codeword determined in step S7 or S8. Bits are transmitted one by one, and the bit count is increased by one at a time.

After all the VLC data is sequentially transmitted in step S13, the code stored in step S3 is transmitted to the transmission channel. If the sign of the level is positive, 0 is transmitted, and if it is negative, 1 is transmitted.

[0030]

As described above, according to the present invention, a signal source code is encoded into VLC (MPEG and CCITT).
Since only 16 bits of memory are needed to store the 18-bit VLC look-up table (of the standard recommendation), memory space can be optimized and thus the cost of the system can be reduced.

[Brief description of the drawings]

FIG. 1 shows a VLC including a microprocessor and a ROM.
FIG. 2 is a schematic diagram of an encoder.

FIG. 2 shows a 16-bit VL stored in a ROM of FIG.
It is a figure showing the structure of C look-up table.

FIG. 3 is a flowchart illustrating a method of encoding a digital signal into a VLC.

FIG. 4 is a flowchart illustrating a method of encoding a digital signal into a VLC.

[Explanation of symbols]

 Reference Signs List 10 video signal processor 20 microprocessor 30 ROM 100 VLC encoder

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁶ , DB name) H03M 7/42

Claims (12)

(57) [Claims] For 1. A digital signal source-code transmission, a method for encoding the signal source code into a variable length code, the signal source code to source code within the microprocessor
Convert to a form that includes run length and level
Stored in a 16-bit memory corresponding to the source code
That accessing a variable length code table that assigns variable length codes, and determining a code length of variable length code read from said variable length code table, the variable
Determine if the length code has a code length of 13 bits
And the variable length code is a code of 13 bits.
Code length of the variable length code when it is determined that the code length is not
And transmitting the variable length code having the determined code length . 2. The variable length code table includes a set of variable length codes. If the variable length code is a variable length code having a maximum code length of 12 bits, the variable length code table indicates the variable length code. If the variable length code has a code length of 13 bits, it has 12 least significant bits and four most significant bits to represent the code length of the variable length code. The variable length according to claim 1, characterized in that it has 13 least significant bits to represent the 13 bits and three most significant bits all having a value of 1 to represent the 13 bits code length. Encoding method. 3. The code of the read variable length code.
The step of determining the length comprises three of the variable length codes.
That the configuration comprises inspecting the most significant bit
The variable length encoding method according to claim 1, wherein: 4. The variable of claim 2, wherein the converted source code includes a run length representing the number of consecutive zeros and a level representing a value immediately following the consecutive zeros. Length encoding method. 5. The method according to claim 1, wherein the variable length code is a 13-bit code.
Determining whether the variable length code has a code length includes checking if all three most significant bits are all ones, and determining the code length of the variable length code comprises :
3. The variable length encoding method according to claim 2, further comprising the step of determining the code length of the variable length code from the four most significant bits unless the three most significant bits are all ones. 6. The method according to claim 1, wherein said converting step comprises:
Temporarily storing a signal in the microprocessor
Hints, the transmission step, a variable length coding method according to claim 2, characterized in that it comprises the step of transmitting the code stored in said microprocessor. 7. An apparatus for encoding a digital source code into a variable length code, said apparatus comprising a plurality of 16 bit memory spaces for storing a variable length code table having a code length of 18 bits. (A) the source code is a run-length code of the source code.
And a variable length code stored in the memory means and corresponding to the signal source code.
Means for accessing a variable length code table to assign, to determine the code length of the variable length code read from said variable length code table (b), the variable length
How to determine if a code has a code length of 13 bits
And the variable length code is a code length of 13 bits
If it is determined that the code length is not
And those containing a means for the constant, the encoding apparatus characterized in that it comprises a means for transmitting said variable length code having a code length which is the determined (c). 8. The variable length code table includes a set of variable length codes. If the variable length code is a variable length code having a code length of up to 12 bits, the variable length code table indicates the variable length code table. Of 1
And the two least significant bits, if the code length of the variable length code <br/> and a four most significant bits for representing a variable length code having a code length of 13 bits 8. The variable length of claim 7, wherein the variable length has 13 least significant bits for representing it, and three most significant bits all having a value of 1 to represent the 13 bit code length. Encoding device. 9. The code of the read variable length code.
To determine the length, the three most significant
The code according to claim 7, wherein the code is checked.
Device. 10. The method of claim 8, wherein the converted source code includes a run length representing the number of consecutive zeros and a level representing a value immediately following the consecutive zeros. Variable length coding device. 11. The method according to claim 11, wherein the variable length code is a 13-bit code.
Wherein the means for determining whether the three most significant bits have a code length includes means for checking whether all three most significant bits are all ones, and the means for determining the code length of the variable length code comprises: 9. The variable length coding apparatus according to claim 8, further comprising means for determining a code length of the variable length code from the four most significant bits if not all ones. 12. The conversion means according to claim 1, wherein
9. The apparatus according to claim 8 , further comprising means for temporarily storing the code in the microprocessor, and wherein the transmitting means includes means for transmitting the code stored in the microprocessor.
3. The variable length encoding device according to claim 1.