KR100215563B1 - The input data formatting method in run length decoder - Google Patents

Abstract

The present invention relates to an input data formatting method of a stemley decoder, and has a concise structure of the (run, level) data and other data so as to easily extract various data to be processed in addition to the (run, level) data in a line length decoder. Formed in a format, the line length decoder can easily extract various data to be processed in addition to the (run, level) data.

Description

A method of formatting input data of run 1ength decoder

1 is a schematic block diagram of a general video encoder,

2 is a schematic block diagram of a general video decoder,

3 illustrates a structure of data formatted by an input data formatting method of a line length decoder according to the present invention.

(a) shows the format of (run, level) data,

(b) shows the format of intra DC coefficient data,

(c) shows the format of macroblocked data without coefficients to be transmitted,

(d) shows the format of not coded blocks and end of block data.

Explanation of symbols for main parts of the drawings

1: discrete cosine transform unit 2: quantizer

3: line length encoder 4: variable length encoder

5 variable length decoder 6 line length decoder

7: reverse scanning unit 8: inverse quantizer

9: inverse discrete cosine transform unit

The present invention relates to a method for formatting data input to a line length decoder, and in particular, to facilitate the extraction of various data to be processed in addition to the (run, level) data in the line length decoder. An input data formatting method of a line length decoder for forming other data in a concise format.

In general, the old society is called an information society, and as the amount of information to be processed is increasing day by day, data must be compressed to effectively use the existing transmission band. In particular, in the case of digital video signals, since the amount of information is very large, it is necessary to compress the video data in order to more efficiently store, retrieve, and transmit the information.

1 is a block diagram schematically showing the configuration of a general video encoder, and is used in many standardized encoders such as H.261, MPEG-1, MPEG-2, and the like.

Referring to FIG. 1, the discrete cosine transforming unit 1 outputs a discrete cosine transform coefficient by performing discrete discrete conversion of an inter-frame difference image, for example, into a block of 8x8 pixels, in order to remove correlation between pixels. The quantizer 2 quantizes and outputs the discrete cosine transform coefficients of the inter-frame difference image output from the discrete constituent transform unit 1 at a predetermined quantization interval.

The DCT coefficients quantized by the quantizer 2 are converted into a one-dimensional data sequence through a zig-zag scanning process, input to a line length encoder 3, and the line length encoder 3 is a jig- The data output through the jag scanning process is made into a two-dimensional (run, level) consisting of a continuous number of zeros and a sequence of nonzero zeros.

The length coded data in the length coder 3 is variable length coded by the Huffman table in the variable length coder 4 and then output to a video buffer (not shown).

At this time, the DC coefficient and the AC coefficient among the DCT coefficients are distinguished and encoded in another method. Normally, the DC value of each block has much correlation with the DC value of the neighboring block. Therefore, the difference between the DC value of the previous block is obtained and the difference value is encoded. Obtain and encode the difference. The DC difference values thus obtained are encoded by one-dimensional variable length coding.

That is, the DC coefficient is variable length coded by dividing the DC size (dct-dc-size) and the DC difference (dct-dc-differential), but if the DC size (dct-dc-size) is 0, the DC size (dct Only the code of dc-size is transmitted, and if it is not 0, then DC value (dct-dc-differentia1) is transmitted by the number of bits of DC size (dct-dc-size).

In addition, AC reorders coefficients for more effective data compression by taking advantage of the fact that the AC coefficient near the DC coefficient is not zero in the DCT region and that the zero is more likely to be farther from DC. Align in one dimension via zig-zag scan. Here, zero-run and zero-coefficient values (leve1) are expressed in two dimensions (run, level).

For example, with a zig-zag scan, the aligned DCT coefficients, such as 30,2,0,0, -8.0,0,0,9, are passed through the line length encoder (0,30) , (2), (2, -8), (3,9)... It is expressed as

And, if the zig-zag scanned coefficients continue to end after some position, add an end of b1ock (EOB) sign indicating the end of the block.

In this manner, the line length coded data is variable length coded by the Huffman table.

In addition, when there is no coefficient to be transmitted in inter coding, this is called a masked macrob1ock. Information data indicating how many skipped macroblock blocks are continued, and information data indicating whether blocks belonging to each macroblock are coded are also coded. It is transmitted to a decoder through the video buffer (not shown).

On the other hand, the image data transmitted through the compression process as described above is restored to the original data in the image decoder, such image decoder is to implement the image encoder in reverse.

That is, FIG. 2 is a schematic block diagram of a general video decoder, comprising: a variable length decoder 5 for performing variable length decoding on encoded data on data; A line length decoder 6 for performing line length decoding on the image data output from the variable length decoder 5; An inverse scanning unit 7 which scans the data output from the line length decoder 6 in reverse and outputs an 8x8 frequency coefficient block; An inverse quantizer 8 for performing inverse quantization on the 8 × 8 frequency coefficient block output from the inverse scanning unit 7 and outputting the inverse quantizer; And an inverse discrete contribution converter 9 for outputting an 8x8 pixel block by cooling the DCT in reverse with respect to the 8x8 frequency coefficient block output from the inverse quantizer 8.

In the video decoder as described above, the variable length decoder 5 extracts the variable length coded D coefficients from the coded bit stream and outputs the result to the line length decoder 6 that performs variable length coding inversely. (6) performs line length decoding on the data output from the variable length decoder 5 and outputs it to the inverse scanning unit 7. FIG.

In addition, the inverse scanning unit 7 converts the one-dimensional DCT coefficients output from the line length decoder 6 into two dimensions again according to the scanning method.

As described above, the DCT coefficients output in two dimensions are inversely quantized by the inverse quantizer 8 to be restored to the actual DCT coefficients, and then inverse discrete coherence transformed by the inverse discrete cosine transform unit 9 to perform an 8 × 8 pixel block. Will be output as

At this time, the data input to the line length decoder 6 basically has a run and a level, and the run represents a length of zero and the level is zero after the run length. This is the following value.

However, the line length decoder not only processes the (run, level) symbol data as described above but also needs to perform data processing other than the (run, level) symbol data. There was a problem that could not handle the data of the exception.

Accordingly, the present invention is to solve the above-mentioned problems of the prior art, so that it is easy to extract various data to be processed in addition to the (run, level) symbol data in the line length decoder. It is an object of the present invention to provide a method for formatting input data of a line length decoder that forms data in a compact structure format.

The input data formatting method of the line length decoder according to the present invention for achieving the above object is, if the input data input to the line length decoder is (run, level) data, the level (leve1) data from the least significant first bit to 12th Up to a bit, run data is allocated up to the 18th bit of the 13th bit subheader, and an indicator is allocated to the 19th bit, which is the most significant bit.

If the input data input to the line length decoder is intra DC coefficient data, the DC difference value is allocated from the least significant first bit to the tenth bit, and the intra DC precision (intra -dc-precision is assigned to the 11th and 12th bits, the identifier (lD) is assigned to the 13th and 14th bits, and the DC size (dct-dc-size) is allocated from the 15th to 18th bits. And the indicator is assigned to the 19th bit, which is the most significant bit.

If the input data inputted to the line length decoder is a macroblock data having no coefficient to be transmitted, the number of skipped macrob1ock numbers to be transmitted is allocated from the least significant first bit to the seventh bit. The identifier is assigned to the 13th bit and the 14th bit, and the indicator is assigned to the 19th bit, which is the most significant bit.

If the input data input to the line length decoder is an uncoded b1ock or end of b1ock data, the not coded block number is set from the least significant first bit to the third. It assigns up to bits, assigns an identifier to the 13th bit and the 14th bit, and assigns an indicator to the 19th bit that is a high-order bit.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

In the embodiment of the present invention, the following four cases are treated as exceptions.

(1) intra DC cofficient

(2) a skipped macroblock with no coefficients to transmit

(3) not coded block

(4) end of block

In the case of (1), the DC size (dct-dc-size) and the DC difference value (dct-dc-differentia1) are not necessary (run, level) data to process the intra DC coefficient.

In the case of (2), since a macroblock having no coefficient to be transmitted is generated in macroblock units, it cannot be expressed as a run limited to a maximum size of one block.

Similarly, in the case of (3) above, an uncoded block can be one or more blocks, and thus cannot be expressed as a run.

In the case of (4), the block end (EOB) means that there are no more coefficients in the block, and since the rest of the block is set to 0, it can be expressed as (run, level) data. It was.

In order to handle exceptions as above, the format of the (run, level) data must be defined.

Therefore, in the input data formatting method of the line length decoder according to the present invention, if the input data input to the line length decoder is (run, level) data, the level (leve1) data is allocated to the 12th bit of the least significant first bit The run data is allocated from the 13th bit to the 18th bit, and the indicator (IND) is allocated to the 19th bit, which is the most significant bit.

In the format of the (run, level) data formed by the above method, as shown in FIG. 3 (a), 12 bits from the first bit to the 12th bit are loaded with the level (leve1) data and the 13th bit. 6 bits from 18 to 18 carry run data, and 19 first bits have a structure of 0 indices (IND).

In this case, when the indicator lND is 0, the (run, level) data is processed by performing a general line length decoding process, and when the indicator (IND) is 1, it means that the (run, level) data should be exception processed. .

On the other hand, if the input data input to the line length decoder is intra DC coefficient data, the DC difference (dct-dc-differential) value is allocated from the least significant first bit to the tenth bit, and the intra DC precision (intra -dc-precision (PREC) is assigned to the 11th and 12th bits, the identifier (ID) is assigned to the 13th and 14th bits, and the DC size (dct-dc-size) from the 15th bit to l8 Allocate even the victim bit and assign the indicator (IND) to the most significant bit, the 19th bit.

In the format of the intra DC coefficient data formed by the above method, as shown in FIG. 3 (b), DC difference (dct-dc-differential) data is loaded on 10 bits from the 1st bit to the 10th bit. In the 2nd bit of the 11th and 12th bits, DC long-precision (PREC) is carried.In the 13th bit, 14th bit, the 2nd bit carries an identifier (ID). DC size (dct-dc-size) data is carried on 4 bits up to the 18th bit, and an indicator (IND) of 1 is loaded on the 19th bit.

At this time, the identifier (1D) has a value of OO or 11 to inform that the current input value is a DC coefficient.

In addition, when the identifier ID is 11, as in the case of OO, it indicates that the current input value is a DC coefficient and also indicates that it is the first DC coefficient of the slice. In this case, the ID 11 indicates that the slice has been started or the ability to create intra-dc precision (PREC) is not given by a separate external signal, but the level (1eve1). It is designed according to the format of).

If the input data inputted to the line length decoder is a macroblock data having no coefficient to be transmitted, the number of skipped macroblock numbers to be transmitted is allocated from the lowest 1st bit to the 7th bit. The identifier ID is allocated to the 13th bit and the 14th bit, and the indicator lND is assigned to the 19th bit, which is the most significant bit. At this time, the 8th to 12th bits and the 15th to 18th bits are reserved bits.

The format of the macroblock data without coefficients to be transmitted formed by the above method is the number of blocks without coefficients to be transmitted in 7 bits from the 1st bit to the 7th bit, as shown in FIG. (skipped macrob1ock number: SKN) is loaded, 8 to 12 bits are not used as reserved bits, 2 bits of the 13th and 14th bits have an identifier (lD), and 15 to 18th bits are used. Instead of being used as a reserved bit, the 19th bit has a structure in which an indicator IND of 1 is loaded.

In this case, the number of macroblocks (SKN) without coefficients to be transmitted is generated in units of macroblocks, and the maximum number of macroblocks without coefficients to be transmitted is 120 because the maximum number of slices must be considered. Thus, we assign a hit to the number of macroblocks that have no coefficient to transmit.

The identifier lD has a value of 1, indicating that the current input is data of a macroblock having no coefficient to be transmitted. At this time, if SKN = 0, skip to the next macro block, and if SKN = 1, two macro blocks are skipped.

On the other hand, if the input data input to the line length decoder is not coded b1ock or end of b1ock (EOB) data, the least significant first bit is used as the not coded b1ock number. To bit 3 to bit 3, bit 4 to bit 12 should not be used as reserved bits, assign an identifier (ID) to bit 13 and bit 14, and bit 15 to bit 18 also reserved. Instead of being used as a bit, the indicator lND is assigned to the 19th bit, which is the most significant bit.

The format of the uncoded block and block end data formed by the above method is the number of uncoded blocks in three bits from the first to third bits, as shown in FIG. coded b1ocknumber (NCB) is carried, an identifier (ID) is carried in two bits of the thirteenth and fourteenth bits, and an indicator (1ND) of one is carried in the nineteenth bit. Bits 4 through 12 and bits 15 through 18 are not used as reserved bits.

At this time, the number of uncoded blocks (NCB) is generated in units of blocks, and up to one macroblock (four luminance blocks and two chrominance signal blocks) must be taken into account, so the number of uncoded blocks is up to six. Therefore, 3 bits are allocated to the number of uncoded blocks.

The identifier lD has a value of 10 indicating that the current input is uncoded block data.

Also, in the case of block end (EOB) data, it is processed in the same way as the uncoded block data, and only the number (NCB) of the uncoded blocks has a value of OOO.

Thus, if the number of uncoded blocks (NCB) is zero, the next block is skipped. If the number of uncoded blocks (NCB) is 1, two jumps are made.

According to the present invention as described above, the input data formatting method of the line length decoder may be performed by a variable length decoder, and the formatting may be performed by forming a separate formatting unit.

According to the present invention as described above, not only is it easy to extract various data to be processed in addition to the (run, level) data in the line length decoder, but also intra-dc-precision and slice start are input data. By including in the above two signals do not need a separate external signal.

And, it is necessary to know whether the current macroblock is intra or non intra, and this information can be obtained as an identifier (ID). That is, if the identifier ID is OO, 11, the current input data is an intra DC coefficient, and thus the current macroblock is intra.

In addition, the number of macroblocks without coefficients to be transmitted (SKN) and the number of uncoded blocks (NCB) can be stored in the same counter, so that the number of macroblocks without coefficients to be transmitted (SKN) and uncoded There is no need for a counter for each of the number of blocks (NCB).

Claims (3)

A method of formatting input data input to a line length decoder into 19-bit data according to its contents. If the input data is (run, level) data, the level (leve1) data from the least significant first bit to the twelfth bit. Assigns run data from the 13th to 18th bits, assigns an indicator to the 19th bit, which is the most significant bit, and if the input data is intra DC coefficient data, the DC difference ( Allocate the dct-dc-differentia1 value from the least significant 1st bit to the 10th bit, assign intra-dc-precision to the 11th and 12th bits, and assign the identifier (lD) to the 13th bit. And 14th bit, DC size (dct-dc-size) from 15th to 18th bit, and indicator to 19th bit, the most significant bit. How to format it. The method of claim 1, wherein the indicator indicates a normal (run, level) data format, and a value of 1 indicates a data format for exception processing. 2. The method of claim 1, wherein the input data formatting method comprises: if the input data is skipped macrob1ock data having no coefficients to be transmitted, the number of the skipped macrob1ock numbers having no coefficients to be transmitted is determined from the least significant first 7 bits. The first bit, the identifier to the 13th bit and the 14th bit, the indicator to the most significant bit, the 19th bit, and the input data is not coded block or end of block. b1ock) data, assigns the not coded block number from the least significant 1st bit to the 3rd bit, assigns the identifier to the 13th and 14th bits, and assigns the indicator to the 19th most significant bit. An input data formatting method of a line length decoder characterized by being assigned to bits.