CN101052972A - Multiple technique entropy coding system and method - Google Patents

Abstract

A system, method and computer program product having optimal matching to a known or measured probability distribution encodes data without the use of an excessively large lookup table. An encoder (201) constructed according to the present invention uses two or more different encoding methods in combination. In one embodiment, Huffman coding by table lookup is combined with computational generation, such as by using an exponential Golomb equation. The most commonly occurring elements are looked up in a small Huffman table, while the remaining elements are coded with the equation. In another embodiment, data is encoded using two or more equations. In yet another embodiment, data is encoded using multiple tables in conjunction with one or more equations.

Description

Multiple technique entropy coding system and method

Related application

The application requires the right of priority of following provisional application: the provisional application that No. the 60/612nd, 311, the U.S. Patent application of submitting on September 21st, 2004 that is entitled as RATE CONTROL WITH VARIABLE SUBBANDQUANTIZATION; The provisional application that No. the 60/612nd, 652, the U.S. Patent application of submitting on September 22nd, 2004 that is entitled as SPLIT TABLE ENTROPY CODING; The provisional application that No. the 60/612nd, 651, the U.S. Patent application of submitting on September 22nd, 2004 that is entitled as PERMUTATION PROCRASTINATION; That submitted on October 12nd, 2004 is entitled as MOBILEIMAGING APPLICATION, DEVICE ARCHITECTURE, the provisional application that No. the 60/618th, 558, the U.S. Patent application of ANDSERVICE PLATFORM ARCHITECTURE; That submitted on October 13rd, 2004 is entitled as VIDEOMONITORING APPLICATION, DEVICE ARCHITECTURES, the provisional application that No. the 60/618th, 938, the U.S. Patent application of ANDSYSTEM ARCHITECTURE; That submitted on February 16th, 2005 is entitled as MOBILE IMAGINGAPPLICATION, DEVICE ARCHITECTURE, the provisional application that No. the 60/654th, 058, the U.S. Patent application of AND SERVICEPLATFORM ARCHITECTURE AND SERVICES; Its full content is incorporated into this with for referencial use.

The application is the U.S. Patent application of submitting on September 16th, 2004 the 10/944th that is entitled as MULTIPLECODEC-IMAGER SYSTEM AND METHOD, the continuity of No. 437 part, its be now on May 19th, 2005 the disclosed U.S. disclose US2005/0104752 number; That submitted on April 17th, 2003 is entitled as SYSTEM, the U.S. Patent application the 10/418th of METHOD AND COMPUTER PROGRAM PRODUCT FOR IMAGEAND VIDEO TRANSCODING, the continuity of No. 649 part, its be now on November 6th, 2003 the disclosed U.S. disclose US2003/0206597 number; That submitted on April 17th, 2003 is entitled as WAVELETTRANSFORM SYSTEM, the U.S. Patent application the 10/418th of METHOD AND COMPUTER PROGRAMPRODUCT, the continuity of No. 363 part, its be now on October 23rd, 2003 the disclosed U.S. disclose US2003/0198395 number; The U.S. Patent application the 10/447th that is entitled as PILE-PROCESSING SYSTEM AND METHODFOR PARALLEL PROCESSORS that on May 28th, 2003 submitted to, the continuity of No. 455 part, its be now on Dec 11st, 2003 the disclosed U.S. disclose US2003/0229773 number; The U.S. Patent application the 10/447th that is entitled as CHROMATEMPORAL RATE REDUCTION AND HIGH-QUALITY PAUSESYSTEM AND METHOD that on May 28th, 2003 submitted to, the continuity of No. 514 part, its be now on Dec 25th, 2003 the disclosed U.S. disclose US2003/0235340 number; The U.S. Patent application the 10/955th that is entitled as SYSTEM AND METHOD FORTEMPORAL OUT-OF-ORDER COMPRESSION AND MULTI-SOURCE COMPRESSION RATE CONTROL that on September 29th, 2004 submitted to, the continuity of No. 240 part, its be now on May 19th, 2005 the disclosed U.S. disclose US2005/0105609 number; The U. S. application of submitting on September 20th, 2005 that is entitled as COMPRESSION RATE CONTROL SYSTEM AND METHOD WITHVARIABLE SUBBAND PROCESSING (procurator's document 74189-200301/US number) the _ _ _ _ number the part continuity; Its full content is incorporated into this with for referencial use.No. the 6th, 825,780, the United States Patent (USP) that is entitled as MULTIPLE CODEC-IMAGER SYSTEM AND METHOD that the application also publishes on November 30th, 2004; No. the 6th, 847,317, the United States Patent (USP) of publishing on January 25th, 2005 that is entitled as SYSTEMAND METHOD FOR A DYADIC-MONOTONIC (DM) CODEC; And on September 21st, 2005 U. S. application that is entitled as PERMUTATION PROCRASTINATION (procurator's document 74189-200501/US number) submitted to the _ _ _ _ number full content be incorporated into this with for referencial use.

Technical field

The present invention relates to data compression, and more specifically, the probability of occurrence that relates to based on them comes coded data element efficiently.

Background technology

The still image of Direct Digitalization and video need many " positions ".Therefore, usually in order to store, to transmit and other application and compressed image and video.Most image and video compressor are shared the basic structure with various variations.As shown in Figure 1, this basic structure comprises three levels: conversion stage, quantized level and entropy coding level.

Video " codec " (compressor/decompressor) is used for reducing the required data rate of data communication stream by carry out balance between picture quality, processor requirement (that is cost/power consumption) and compression ratio (that is result data speed).Current available compression method provides different compromise (trade-offs) scopes, and has produced a plurality of coding and decoding schemes, and wherein, each scheme is optimised to satisfy the needs of application-specific.

The purpose of conversion stage is to collect the energy (energy) of source picture or information with by utilizing local similar and pattern in picture or the sequence to be converted into the compressed format of maximum possible in the video compressor.Compressor reducer is designed to work well in " typical case " input, and neglects compression " at random " or " irrational " input to its fault that causes.

Many compression of images and video-frequency compression method such as MPEG-2 use discrete cosine transform (DCT), as conversion stage.

Some newer compression of images and video-frequency compression method such as the MPEG-4 structure use various wavelet transformations, as conversion stage.

Wavelet transformation comprise with one dimension or more than the form repeated application wavelet filter of one dimension to data set.For compression of images, can use 2D wavelet transformation (level is with vertical).For video data stream, can use 3D wavelet transformation (level, vertical and time).

Prior art Fig. 2 shows the compromise example 100 in the current available various compression algorithms.As shown in the figure, this compression algorithm comprises that they comprise different MPEG video distribution schemes based on the codec 102 of small echo with based on the codec 104 of DCT.

Different with codec algorithm based on DCT, 2D and 3D small echo because of its pleasing picture quality and flexibly ratio of compression spoken highly of, impel the JPEG council to adopt wavelet algorithm as its JPEG2000 still image standard.Unfortunately, with respect to the possibility of DCT, most small echo is implemented to use very complicated algorithm, needs the good treatment ability.In addition, small echo has proposed challenge especially for Time Compression, makes the 3D small echo especially difficult.

Owing to these reasons, with respect to the high power capacity industrial standard codec such as MPEG, small echo does not provide cost emulative advantage, therefore, only is used in compact applications.Therefore need a kind of 3D small echo of viable commercial to implement, its low-power consumption and low cost at three principal market stages is carried out optimization.

For example, miniature camera is more and more universal, and the advantage of its signal of digitized processing is tangible.For example, the fastest developing stage in cell phone market is the telephony phase with image and video clips function in some countries.Most of Digital Still Camera have the video clips function.At the mobile wireless mobile phone market, the transmission of these static images and video clip needs bigger device battery capacity.Existing video encoding standard and digital signal processor bring bigger pressure to battery.

Another new application is personal video recorder (PVR), and it allows spectators to suspend live TV and regularly changes program.These devices use the digital hard disk storer to come recording of video, and require the analog video from cable is carried out video compress.For the function as picture-in-picture and the record while seeing is provided, these unit need a plurality of video compression encoders.

Another application that is developing is the digital video recorder (DVR) that is used to monitor with security video.Each channel of input video to be stored also all needs compressed encoding.For utilize conveniently, digital network transmission structure flexibly, in the video camera of being everlasting, make video digitizer.Even have older complicated register structure, also use the multichannel condensing encoder.

Certainly, can make a profit from the compression scheme to low-power consumption and low-cost optimized viable commercial in other a large amount of market.

Entropy coding

The target of entropy coding (being also referred to as " source code " in the literature) normally generates short message from message or information source, its decoded subsequently time origination message, preferably and origination message identical.Typically, by source message is divided into " symbol " and one by one symbol ground processing messages realize, rather than by in the code book of super large, search bigger piece or even whole input message (for example, image or video GOP) finish.

The incoming symbol work of regular length and the entropy coder class that generates for the bit string of each variable-length are called " variable encoder group " in the literature.

Two kinds of typical methods of coded identification

Provide incoming symbol to be encoded, a kind of coding method is symbol is searched this symbol as index and in the table that is known as " code book ".The item that finds in code book is the output of the coding of this symbol.Typically, this code book is enough big, so that provide item for each possible symbol.

In some embodiments, the single random access of his-and-hers watches is very quick and efficient.Yet, in other embodiment, to random access of big table relatively slow (because cache memory loading) or cost higher relatively (because, such as the chip-scale memory cost in FPGA or ASIC).

Second typical scenario that is used for coded identification is its expression formula (normally bit string) is carried out some calculating operations, and it generates coding output as its result.Like this, need not big code book and can generate output.

In some embodiments, this calculating is quite quick and efficient.Yet, in other embodiment, need multistep calculating and relatively slow.

Demoder must determine to decode the back length of each variable-length bit string (that is coded word) of symbol.This is usually by finishing the coded word ordering (no coded word is the prefix of arbitrary other coded word) with " Huffman prefix feature ".

Distribute

Above-mentioned entropy coding comes work by utilizing the non-consistent probabilistic in these symbols.When symbol has high probability of occurrence when (meaning that it frequently occurs in message or source), use the short code word that it is encoded.When symbol has lower probability of occurrence when (meaning that it seldom occurs in message or source), use longer coded word that it is encoded.Like this, has the coding output of the long coded word of many short code words and minority usually than the input weak point.

By Shannon (C.E.Shannon, The Mathematical Theory ofCommunications, Bell System Technical Journal, July ﹠amp; October1948) optimized encoding of Miao Shuing has the length of each the output code word relevant with the probability of occurrence inverse logarithm of corresponding symbol in the input of its source.This can not obtain usually exactly, but the design of code device is as far as possible near it.

Therefore, in order to design effective entropy coding, the probability distribution of these symbols be known, record, approximate or supposition.

Distribute for some, the calculation of coding method can be finished with step seldom, and the coding that for other distribution, needs many steps to calculate.

In video compress work, the probability distribution of quantization parameter is unworkable sometimes.In other words, distributing is not the distribution with known quick calculation code, and the amounts required code book of probable value is excessive and be unsuitable for the available storage of searching.

Therefore, what need is and probability distribution optimization coupling known or that measure, and need not the encoding scheme of super large look-up table.

Summary of the invention

A kind of system, method and computer program that mates with probability distribution optimization known or that measure that have that is used to need not to use super large look-up table coded data disclosed.Scrambler constructed according to the invention is used in combination two or more different coding methods.

According to an aspect of the present invention, the Huffman coding searched of use table generates (for example, by using the index Golomb equation) and combines with the Accounting Legend Code word.In little Huffman table, search the element of the most frequent appearance, and remaining element is encoded with this equation.The Huffman that this scheme provides the use table to search encodes (promptly, with probability distribution optimization coupling known or that measure) encode (promptly with simple computation, the quick calculating that nothing is searched) advantage that advantage combines, and avoided the complete Huffman coding defective of (that is, needing to support greatly table).

According on the other hand, use two or more equation coded datas.Can not accurately be fit to can use different equations under the situation of data type at single equation, each equation is used for the different piece of data, to describe whole data probability distributions better.

According to more on the one hand, use a plurality of tables that combine with one or more equations to come coded data.A plurality of equations are used for a plurality of parts of data, and wherein, equation has accurately been described the probability distribution of data division.Showing spreadable these equations that has, to cover the gap in these equations, for example, is known place at non-quick calculation code.

Description of drawings

Fig. 1 shows the block scheme that is used for the compression/de-compression data according to an embodiment.

Fig. 2 shows the compromise example in current available various compression algorithms.

Embodiment

Fig. 1 shows the square frame Figure 200 that is used for the compression/de-compression data according to an embodiment.Comprise scrambler portion 201 and demoder portion 203 among this square frame Figure 200, they constitute " codec " together.Scrambler portion 201 comprises and is used for conversion module 202, quantizer 204 and the entropy coder 206 of packed data to store file 208 into.In order to carry out the decompression of this file 208, demoder portion 203 comprises and is used for decompressed data so that use entropy decoder 210, inverse DCT 212 and the inverse transform module 214 of (that is, checking the situation of video data etc.).In use, for the purpose of decorrelation, 202 pairs of a plurality of pixels of conversion module (under the situation of video data) are carried out reversible transformation, and it often is linear.Next, quantizer 204 is realized the quantification of transformed value, after this, and the entropy coding that entropy coder 206 is responsible for quantized transform coefficients.

Scrambler constructed according to the invention is used in combination two or more different coding methods.Using the negative exponent of big coefficient value (it has lower probability of occurrence in input source) and the little table of little coefficient value (the most frequent appearance has the maximum probability value in input source) comes well near some quantitation video DATA DISTRIBUTION.Therefore, according to an aspect, have only little table to use, any method in two kinds of technology (table or calculating) is used in simple computation method and selection.

According to a further aspect of the invention, select which kind of technology to be applied to which kind of data element or symbol can be simple size detection.In this example, treat that the symbol of entropy coding is always positive, its scope from 1 to 2 ¹⁵-1.Except the value zero.This symbol of easy detection is to determine that whether it is less than fixed constant.If so, then use the table of the identical size of this constant.Otherwise, use this computing method.

For little (high frequency) among this embodiment value, use the Huffman coded word in the look-up table.For big (low frequency) value, user's formula (for example, using index Golomb type equation) Accounting Legend Code word.This enforcement is symbol ground coding one by one, and does not preserve the historical record of the symbol that is encoded.The symbol of constant length is input to scrambler with 16, and output length changes to 16 (low frequency value) from 1 (high frequency value).

In two parts of scrambler each has Huffman prefix feature respectively.In other words, it is different with the beginning part of the coded word that is used for this another symbol of scrambler same section to be used for the coded word of symbol of a part of scrambler.Has the typical probability distribution scope that is used for many application, the combined code that is used for two parts of scrambler also has Huffman prefix feature, and making does not need extra marker bit to come coded word of instruction decoding device to finish wherein in the output stream of coded word and next coded word begins wherein.

Example algorithm 1

This example algorithm is accepted symbol S as input, 16 the positive nonzero integers of S for representing with binary mode.It generates bit string W as output.

Step 1

If S＞15 jump to step 3.

Step 2

In following given table 1, search S, with discovery value B and length L.

W is made up of the low order L position of B.

W is added in the output bit stream.Finish.

Step 3

Count from the significance bit in " 1 " position beginning logarithm S+8 that comprises of high order end.Call counting C.

Step 4

W comprises 2C-1 position: C-1 " 0 " position, follows C significance bit by S+8.

W is added in the output bit stream.Finish.

The table 1 that is used for example algorithm 1

Symbol	L	B	The output bit string
				1	1	1	1
2	3	2	010
				3	3	3	011
4	5	4	00100
				5	5	5	00101
6	5	6	00110
				7	5	7	00111

8	6	7	000100
				9	6	8	000101
10	6	9	000110
				11	6	10	000111
12	8	11	00001000
				13	8	12	00001001
14	8	13	00001010
				15	8	14	00001011

For comparison purposes, if do not use table 1, then down tabulation 2 provide should be by the output for providing less than 16 value of symbol of above step 3 and step 4 (calculating of coded word generates).By comparing these two tables as can be seen, to compare with the calculating generation method of table 2, the Huffman table method of use table 1 provides shorter coded word for the symbol of some more frequent appearance.

The table 2 that is used for example algorithm 1

Symbol	L	B	The output bit string
				1	1	1	1
2	3	2	010
				3	3	3	011
4	5	4	00100
				5	5	5	00101
6	5	6	00110
				7	5	7	00111
8	7	8	0001000
				9	7	9	0001001

10	7	10	0001010
				11	7	11	0001011
12	7	12	0001100
				13	7	13	0001101
14	7	14	0001110
				15	7	15	0001111

Performance

When the method for this example is carried out on some computer platforms, reached high performance target, this be because:

-it provides optional Huffman coding for most of average cases;

That part of-probability distribution measured for optimization coupling, its need can easily be suitable for the little table of limited storage;

-it uses very simple index-Golomb calculating of encoding for rare situation;

-no matter which type of symbol is operated all very rapid.

Various improvement can be applied to above-mentioned example of the present invention implements.For example, can revise entropy coder, the symbol of no symbol (the being positive) numeric character of symbol of encoding more than coding.In order to encode efficiently, each the L item in this table is added 1, sign bit is added to each B value, and comprise the list item of negative symbol.Under tabulate and 3 provide example.In this table, 0 symbol item allows directly to search faster.Because this 0 symbol item is obsolete pseudo-, so its content is inessential.

For situation, above simple algorithm is made an amendment slightly as algorithm 2.

Example algorithm 2

This algorithm is accepted symbol S as input, and S is 16 integers (not allowing null value) with binary representation.It generates bit string W as output, is used for adding to by turn the compression position flow that is generating.

Step 1

If the absolute value of S greater than 15, then jumps to step 3.

Step 2

Under tabulate and search S in 3, with the value of finding B and length L.

W is made up of the low order L position of B.

W is added in the output bit stream.Finish.

Step 3

Comprise that from high order end " 1 " position begins the significance bit in the absolute value of digital S+8 is counted.Be referred to as C.

Step 4

W comprises 2C position: C-1 " 0 " position, is thereafter C significance bit of the absolute value of S+8, is thereafter the sign bit of S again.

W is added in the output bit stream.Finish.

The table 3 that is used for example algorithm 2

Symbol	L	B	The output bit string
				-15	9	23	000010111
-14	9	21	000010101
				-13	9	19	000010011
-12	9	17	000010001
				-11	7	15	0001111
-10	7	13	0001101
				-9	7	11	0001011
-8	7	9	0001001
				-7	6	15	001111
-6	6	13	001101
				-5	6	11	001011
-4	6	9	001001
				-3	4	7	0111
-2	4	5	0101
				-1	2	3	11
0	0	0	(not usefulness)
				1	2	2	10
2	4	4	0100
				3	4	6	0110
4	6	8	001000
				5	6	10	001010
6	6	12	001100
				7	6	14	001110

8	7	8	0001000
				9	9	10	0001010
10	9	12	0001100
				11	9	14	0001110
12	9	16	000010000
				13	9	18	000010010
14	9	20	000010100
				15	8	22	000010110

In above example, the use table is searched the advantage of the Huffman coding of (with probability distribution optimization coupling known or that measure) and can (for example be encoded with simple computation, index-Golomb (the quick calculating that nothing is searched)) advantage combines, and has avoided the shortcoming of complete Huffman coding (greatly table).We have also described a kind of method, its for situation commonly used by sign bit being attached to the symbol data of the symbol of encoding quickly in the look-up table, and in can extra position not inducing one to export.

With the similar method of above-described method in, can the employing table search and calculate the various combinations of generation.For example, can use two different equations, each all is applied in the different subclass of the symbol of encoding.Use the advantage of this combination to be that single known equation can not be well be complementary with the probability distribution of specific data type, provide more approaching coupling but two or more equations are combined.Another advantage is that more common symbol can adopt simpler equation, to improve the bulk treatment speed of coding.

In a plurality of equations each all can be used the equation of identical universal class, perhaps can be dissimilar set of equations altogether.The example that some dissimilar equations maybe can be conveniently used in the processing in the system constructed according to the invention comprises index Golomb, Golomb, Golomb-Rice (or Rice-Golomb) coding and arithmetic coding/decoding device.Arithmetic coding/decoding device with the work of the non-mode of symbol one by one also comprises such as binary dullness (DM) codec (as the United States Patent (USP) of publishing on January 25th, 2005 the 6th that is entitled as System andMethod for a Dyadic-Monotonic (DM) Codec, described in 847, No. 317) and the subtype of CABAC.

In another embodiment, use a plurality of tables that combine with one or more equations to come coded data.Equation is used to a plurality of parts of data, and wherein, these equations have accurately been described the probability distribution of this data division.Table can be scattered with these equations, is known gap to cover wherein non-quick calculation code.

Its creatively confirmed when system be designed to provide or its when particular kind of relationship between the number of the number of the distinctive signs that expectation is used for encoding and look-up table distinctive signs is provided, can in scrambler, obtain the efficient that has superiority.When system so designs or the table that includes a plurality of distinctive signs (when it is illustrated in the entropy coding definite percentage of all (or frequency) of the symbol of typically handling) is provided, can obtain other advantages.

For example, have in the specific embodiment of scrambler of look-up table and at least one equational application in combination, can be found to be 0% to 50% the symbol that is expressed as typical case's distinctive signs to be encoded provides look-up table to have superiority.For the purpose of this disclosure, this percentage is called as " the other symbol percentage in list area (or TDSP) ".More specifically, this scope can be expressed as expectation and typically provide 0% to 30% of all distinctive signs of being used to encode, more specifically be 0% to 10%, even more specifically be 0% to 3%, even more specifically be 0% to 1%, even more specifically be 0% to 0.1%, and more specifically be 0% to 0.01%.

In addition, this table can creatively be designed to comprise a plurality of symbols, its comprise generally all (or frequency) of typically providing the symbol that is used to encode 100% to 50%.For the purpose of the disclosure, this percentage is called as " table set frequency percentage (or TAFP) ".More specifically, these a plurality of symbols are included as generally and typically provide 100% to 70% of the symbol that is used to encode all (or frequency), more specifically be 100% to 80%, even more specifically be 100% to 85%, more specifically be 100% to 90%, and more specifically be 100% to 95%.

For example, in a compression scheme according to the present invention, distinctive signs to be encoded add up to 32,766 symbols.Enumerated to be desirably in and had high-frequency 15 distinctive signs in the data that offer scrambler and offer look-up table.Therefore, these 15 symbols comprise about 0.046% of distinctive signs sum.Thereby this embodiment has about 0.046% the other symbol percentage in list area.Yet, can find, in typical operation, these 15 symbologies about 90% the set symbol that is used to encode that provides incident appears.Therefore, this example has about 90% table set frequency percentage.The relatively little table of such 15 symbols can be handled the coding (approx, 90% provide the symbol that is used to encode) of the symbol of overwhelming majority.

In addition, it is contemplated that within the scope of the invention, provide and use a plurality of searching or the coding techniques of other table, and wherein, calculate the other symbol percentage in this list area and this table is gathered frequency percentage based on the combination of a plurality of tables.

Additional feature of the present invention comprises the scrambler that the combination with at least one look-up table and at least one coding staff formula is provided, and wherein, has provided the particular combinations that other symbol percentage in list area and specific table are gathered frequency percentage.For example, favourable combination comprise have from 100% to 70% TAFP 0% to 5% TDSP.Favourable being combined in down shown in the tabulation 4 in addition.

Table 4

(TDSP) (percentage)	TAFP (approximate percentage)
0 to 40	100 to 50
		0 to 40	100 to 60
0 to 40	100 to 70
		0 to 40	100 to 80
0 to 5	100 to 50
		0 to 5	100 to 60
0 to 5	100 to 70
		0 to 5	100 to 80
0 to 5	100 to 90
		0 to 2	100 to 50

0 to 2	100 to 60
		0 to 2	100 to 70
0 to 2	100 to 80
		0 to 2	100 to 90
0 to 2	100 to 50
		0 to 0.5	100 to 60
0 to 0.5	100 to 70
		0 to 0.5	100 to 80
0 to 0.5	100 to 90
		0 to 0.5	100 to 95
0 to 0.1	100 to 50
		0 to 0.1	100 to 60
0 to 0.1	100 to 70
		0 to 0.1	100 to 80
0 to 0.1	100 to 90
		0 to 0.1	100 to 95
0 to 0.05	100 to 50
		0 to 0.05	100 to 60
0 to 0.05	100 to 70
		0 to 0.05	100 to 80
0 to 0.05	100 to 90
		0 to 0.05	100 to 95

Though more than be the complete description of the preferred embodiments of the present invention, can use different replacements, modification and analog.Therefore, the above should not be counted as the qualification to scope of the present invention, and scope of the present invention is defined by the appended claims.

Claims (4)

1. method of compressing data comprises:

Use the combination of at least two kinds of parallel coding techniquess that input traffic is carried out entropy coding, the different piece of each the processing said data stream in the described coding techniques.

2. method according to claim 1, wherein, described at least two kinds of coding techniquess comprise the Huffman look-up table and calculate generation.

3. method according to claim 2, wherein, described calculating is generated as index-Golomb type.

4. method according to claim 2, wherein, described input traffic comprises a series of symbols to be encoded, each symbol all has size, and wherein, use described Huffman look-up table to having encoding symbols, and use described calculating to generate having the encoding symbols of the size that is not less than described fixed constant less than the size of fixed constant.

Images