WO1996008872A1 - Data compressing process and system - Google Patents
Data compressing process and system Download PDFInfo
- Publication number
- WO1996008872A1 WO1996008872A1 PCT/EP1995/003595 EP9503595W WO9608872A1 WO 1996008872 A1 WO1996008872 A1 WO 1996008872A1 EP 9503595 W EP9503595 W EP 9503595W WO 9608872 A1 WO9608872 A1 WO 9608872A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- data
- compression
- compressed
- data stream
- groups
- Prior art date
Links
Classifications
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
- H03M7/00—Conversion of a code where information is represented by a given sequence or number of digits to a code where the same, similar or subset of information is represented by a different sequence or number of digits
- H03M7/30—Compression; Expansion; Suppression of unnecessary data, e.g. redundancy reduction
- H03M7/3084—Compression; Expansion; Suppression of unnecessary data, e.g. redundancy reduction using adaptive string matching, e.g. the Lempel-Ziv method
- H03M7/3088—Compression; Expansion; Suppression of unnecessary data, e.g. redundancy reduction using adaptive string matching, e.g. the Lempel-Ziv method employing the use of a dictionary, e.g. LZ78
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
- H03M7/00—Conversion of a code where information is represented by a given sequence or number of digits to a code where the same, similar or subset of information is represented by a different sequence or number of digits
- H03M7/30—Compression; Expansion; Suppression of unnecessary data, e.g. redundancy reduction
- H03M7/40—Conversion to or from variable length codes, e.g. Shannon-Fano code, Huffman code, Morse code
- H03M7/42—Conversion to or from variable length codes, e.g. Shannon-Fano code, Huffman code, Morse code using table look-up for the coding or decoding process, e.g. using read-only memory
Definitions
- the present invention relates to methods for compressing data and corresponding devices therefor, the data to be compressed being analog and digital voice and image data and data provided by computers.
- Data compression methods are used to eliminate or to reduce redundancies in data streams to be transmitted or to be stored, in order to be able to store the data streams with a smaller storage space requirement or to be able to transmit them with lower bandwidths or at a higher speed.
- the LZW algorithm is based on the fact that a data stream to be transmitted is examined for the presence of elementary data groups contained therein and the data groups found are assigned a place in a tree representing the individual data groups.
- the data groups found can then be transmitted instead of their actual information content by means of codes which represent their position within the tree, as a result of which good compression results can be achieved, in particular in the case of data streams with constantly repeating patterns .
- SPARE BLADE (RULE 26) scarcely, it is necessary to renew the "learned tree" from time to time in order to adjust or adapt to a change in the structure of the data to be transmitted compared to the structure as it was during the learning.
- the object of the present invention is to provide methods for more efficient data compression as well as suitable devices for this.
- One aspect of the present invention is based on the insight that hitherto conventional compression method therefore does not work efficiently because they capture essentially only camerazei ⁇ term variations in the data streams using the compression algorithms used and taken into account for the compression be ⁇ .
- hitherto conventional compression method therefore does not work efficiently because they capture essentially only camerazei ⁇ term variations in the data streams using the compression algorithms used and taken into account for the compression be ⁇ .
- Basic repetitions and Structures within a certain type of data are therefore not sufficiently taken into account for the compression.
- the compression therefore takes place in two stages:
- a first stage in which the structures, correlations or patterns learned over a longer period of time, which occur within a certain type of data, are taken into account for the compression of a data stream currently pending, and
- the method mentioned is particularly advantageous because the compression in the second stage, which, as mentioned, can be implemented by conventional compression methods, is improved in its efficiency, since it receives as input data pre-compressed data that come from data groups with small input lengths exist with respect to which, for example the LZW algorithm, can work in a more efficient way, ie closer to its theoretically realizable optimal compression capabilities.
- frequency distributions of elementary data groups within the data stream to be compressed are preferably taken into account, as they were determined over a longer period of time for the corresponding data type.
- the case er ⁇ mediated elementary data groups that space even over long Zeit ⁇ constantly recur, such as certain Spring ⁇ pattern, certain combinations of numbers or certain Page down ⁇ cuts, their frequency will preferably after ordered so that the most frequently occurring data groups talking by ent short codes and less frequently occurring data groups, but data groups still taken into account in the pre-compression, are represented by correspondingly longer codes.
- a further aspect of the present invention relates to a compression method which can be carried out in one stage and which is based on the fact that for a data stream to be compressed, elementary data groups which are frequently repeated over a longer period of time are “learned” and stored in a memory become.
- the data stream to be compressed at the moment is examined for the presence of these learned elementary data groups and correspondingly learned data groups are then replaced by codes which represent the position of the learned data groups within the memory or within a table, so that the data stream compressed in this way when retrieved from a memory or on the receiver side can be decompressed again by looking up in the table.
- the codes used preferably correspond to the addresses of the learned and replaced data groups within the table, and this address agreement can also be made in such a way that only address differences between two successive replaced data groups are used for coding.
- the method according to claim 1 can preferably again be combined with a second compression stage, in accordance with the first aspect of the present invention.
- the figure shows a function group 1 surrounded by dashed lines, which is used for compressing analog or digital voice or image data or for compressing data that are provided by computers.
- a second function group 2 with dashed lines serves for further compression of the pre-compressed data supplied by the first stage.
- the reference number 3 denotes a transmission device for transmitting compressed data, this transmission group also comprising the two compression stages 1 and 2.
- the present invention also relates to each of the function groups mentioned alone and both to a compression method as described with reference to function group 1 and to a compression method as described with reference to function groups 1 and 2 is written independently.
- the supplied voice, image or computer data are fed to a data analysis device 4 in which individual repeating elementary data groups can be recognized.
- a certain number of different, most frequently recurring data groups are stored in the memory device 5.
- the most frequently occurring data groups are preferably stored within a table in the order of their frequencies. Can group learning these common data ⁇ where done being compressed data stream or at a previously analyzed data stream with similar builds character ⁇ istics.
- the current data stream to be compressed likewise passes through the data analysis device 4 in order to thereby recognize already learned data groups within the data stream and to be able to replace these corresponding data groups in the first compression stage 6 with codes.
- the addresses of the most frequent data groups stored in the memory 5 are preferably used as codes, so that a correspondingly compressed data stream later allows the data stream to be decompressed with the aid of the known association between addresses and stored elementary data groups.
- the differences between the addresses of two successive replaced data groups are preferably used as codes. This means that there is no absolute address assignment, but a relative address assignment. This usually leads to the possibility of using shorter codes, i.e. To be able to use codes with fewer bit numbers.
- the previously learned and stored in the memory 5 known data groups can be arranged in frequency, so that frequent data groups are assigned shorter addresses than less frequent data groups.
- Tests have shown, for example, that only around 4000 different data groups occur in data that are transferred in transactions between banks. With such a small number of different data groups, it is even possible to encode all data groups that occur, which results in a particularly high compression of the data.
- the compressed data can be stored in a memory, which results in a considerable reduction in the storage space required compared to storage in a non-compressed form.
- the data can also be transmitted to a receiver, as a result of which the bandwidth required for the transmission is reduced or the achievable transmission speed increased.
- the compressed data can then be decompressed with the aid of the known allocation scheme between codes and learned most common data groups.
- the corresponding assignment table must of course be known to the recipient, which can take place at the beginning of the transmission or can only occur sporadically if the structure and characteristics of the data to be transmitted have actually changed considerably compared to the time of learning.
- Further compression is preferably carried out after the first compression stage, this compression being able to be based on known compression methods, for example the LZW algorithm or the Huffman algorithm.
- this further compression stage which is formed overall by the function block 2, compression can take place, in which shorter-term variations in the data to be compressed are taken into account. For example, in the event that this stage carries out compression in accordance with the LZW algorithm and works in a synchronous environment, only the patterns or characteristics recognizable within a frame are taken into account.
- the data to be compressed is first analyzed in a data analysis device 7.
- Learned data groups are in turn stored in a further memory 8, this being the structure of the learned tree if an LZW algorithm is used.
- the actual compression then takes place in unit 9, building on the results provided by units 7 and 8.
- the learned data stored in the second memory are discarded from time to time in order to achieve an adaptation to the data which is actually to be compressed.
- the fact that the compression level 2 pre-primed from level 1 If data contains data, stage 2 is able to work with higher efficiency than would be the case if the input data were supplied to this second stage in an uncompressed form.
- Both the functional group 1 and the functional group 2 can be contained in a transmission device 3, this transmission device also additionally containing a transmission device 10 with which the compressed data can be transmitted to a receiver.
- the transmission device 10 can be a device known from modem technology.
- a particular advantage of the present invention is that high compression is achieved without certain information contents being lost.
- the compressed data to be finally saved or transmitted can then be converted into the original data by appropriate decompression and / or decoding operations.
- the data to be compressed can also be image or sound data as well as image or sound signals (also analog).
- the lossless compression provided by the present invention leads, particularly in the video field, to advantages over conventional methods in which compression is always associated with loss of information.
- the compression according to the invention in particular the two-stage compression, is applied to video signals, areas or patterns within a video image that overlap
- REPLACEMENT BUTT do not change a certain period of time more, recorded within the first compression stage, and instead of recording or transmitting the signals representing the corresponding areas or patterns, only corresponding codes are recorded or transmitted which are used when playing back the recorded video signals or then be converted back into the original image content at the receiver.
- a corresponding system will therefore preferably learn from successive images which image contents occur frequently and can therefore be replaced by corresponding table references instead of being transmitted individually.
- the corresponding system will thus independently adapt its table content, which is required for the first compression, corresponding to the images currently being processed, it also being possible, of course, to use previously learned table contents, which are typical for certain images.
- the table contents created for the first compression can also be updated if certain frequently repeated data or signal patterns are recognized within the second compression level, which patterns are then preferably taken into account in the first compression level by means of corresponding table references.
- Examples of picture elements or patterns which are preferably compressed in the first stage are format-related repetitions of signals (PAL; SECAM), blanking frequencies, camera-specific resolution patterns (blank, black, white switching, reference matching pattern) and others repeating image patterns, such as unchanged background of an image, frequency frequencies of image point sequences or gradients thereof.
- PAL format-related repetitions of signals
- blanking frequencies such as unchanged background of an image, frequency frequencies of image point sequences or gradients thereof.
- camera-specific resolution patterns such as unchanged background of an image, frequency frequencies of image point sequences or gradients thereof.
- both medium-specific features and repetitive picture elements can be compressed by the first compression stage.
- the compression according to the invention in particular image data, can already be used during signal acquisition. This means that it is not necessary to first record and temporarily store all of the image information in order to subsequently carry out a compression, but rather the sensor device (eg CCD, diode, tube sensors) performing the detection can at least compress the first stage.
- the sensor device eg CCD, diode, tube sensors
- a sensor device which has been improved in accordance with the invention in this way can, for example, examine different scanning lines for repeating image patterns and, in the event that corresponding repeating image patterns are present, assign certain codes to these image patterns, so that instead of the corresponding patterns, only corresponding codes are output.
- Such a sensor device which already provides compressed output data, appears to be particularly advantageous for space applications.
- intermediate storage of very large amounts of data may be prohibited, so that storing and / or transferring already compressed data is advantageous.
- the compression according to the invention ensures that the original image can then be reproduced true to the original, so that even the smallest image changes are not lost.
- the method according to the invention not only is the information transmitted for a black background that the corresponding image area is a black background, but minor noise-related changes in the background are also evaluated and made available for later reconstruction. Noise information, which can be helpful for later image correction in the recorded object, is thus not lost, but is also recorded or transmitted. Of course, such small changes can again form their own patterns, which in turn can then be appropriately compressed.
- speech applications also form a preferred field of application of the present invention.
- the language of a person also contains long-term patterns that can be learned and stored in a table. Such long-term patterns relate to speech coloring, the tone, the pitch, etc. These patterns can be learned for a specific speaker and stored in a table. If the corresponding language is then stored or transmitted, these characteristics are no longer transmitted as part of a spectrum, but only as codes. The same naturally also applies to music, so that the existing Generally speaking, the invention is also particularly suitable for sound applications.
- a special and interesting application of the present invention thus also arises for audio and video CDs, on which information can be stored in a compressed form according to the present invention.
- the code assignment used for the first compression level i.e. the table can, for example, also be stored on the CD, so that the device which is intended to reproduce the stored information can carry out a corresponding decompression on the basis of the code assignment likewise stored.
Abstract
Description
Claims
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US08/646,358 US6169820B1 (en) | 1995-09-12 | 1995-09-12 | Data compression process and system for compressing data |
EP95932731A EP0728381A1 (en) | 1994-09-12 | 1995-09-12 | Data compressing process and system |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
DE19944432436 DE4432436C2 (en) | 1994-09-12 | 1994-09-12 | Data compression method and device for compressing data |
DEP4432436.7 | 1994-09-12 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO1996008872A1 true WO1996008872A1 (en) | 1996-03-21 |
Family
ID=6528016
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/EP1995/003595 WO1996008872A1 (en) | 1994-09-12 | 1995-09-12 | Data compressing process and system |
Country Status (3)
Country | Link |
---|---|
EP (1) | EP0728381A1 (en) |
DE (1) | DE4432436C2 (en) |
WO (1) | WO1996008872A1 (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE19616531A1 (en) * | 1996-04-25 | 1997-11-06 | Rolf Nagel | Data compression memory storage process for data processing |
DE19706268A1 (en) * | 1997-02-18 | 1998-08-20 | Christian Wenz | Data-format specific compression process for video and sound data |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4813056A (en) * | 1987-12-08 | 1989-03-14 | General Electric Company | Modified statistical coding of digital signals |
US4839724A (en) * | 1986-02-18 | 1989-06-13 | Deutsche AG fur Unterhaltungselectronic OHG | Process for the digital transmission of information |
EP0582907A2 (en) * | 1992-08-10 | 1994-02-16 | Stac Electronics, Inc. | Data compression apparatus and method using matching string searching and Huffman encoding |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4843389A (en) * | 1986-12-04 | 1989-06-27 | International Business Machines Corp. | Text compression and expansion method and apparatus |
US5049881A (en) * | 1990-06-18 | 1991-09-17 | Intersecting Concepts, Inc. | Apparatus and method for very high data rate-compression incorporating lossless data compression and expansion utilizing a hashing technique |
US5339076A (en) * | 1992-04-27 | 1994-08-16 | Integrated Information Technology | Data compression using content addressable memory |
-
1994
- 1994-09-12 DE DE19944432436 patent/DE4432436C2/en not_active Expired - Fee Related
-
1995
- 1995-09-12 WO PCT/EP1995/003595 patent/WO1996008872A1/en not_active Application Discontinuation
- 1995-09-12 EP EP95932731A patent/EP0728381A1/en not_active Withdrawn
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4839724A (en) * | 1986-02-18 | 1989-06-13 | Deutsche AG fur Unterhaltungselectronic OHG | Process for the digital transmission of information |
US4813056A (en) * | 1987-12-08 | 1989-03-14 | General Electric Company | Modified statistical coding of digital signals |
EP0582907A2 (en) * | 1992-08-10 | 1994-02-16 | Stac Electronics, Inc. | Data compression apparatus and method using matching string searching and Huffman encoding |
Also Published As
Publication number | Publication date |
---|---|
DE4432436C2 (en) | 1997-04-03 |
EP0728381A1 (en) | 1996-08-28 |
DE4432436A1 (en) | 1996-03-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
DE3940682C2 (en) | Coding device and system, consisting of a coding device and a decoding device for digital image data | |
DE69333714T2 (en) | Memory for use in decoding variable length words and decoders and methods for use in digital communication systems | |
DE69720559T2 (en) | Image encoding method with variable length codes | |
DE3912605B4 (en) | Digital coding method | |
DE69933058T2 (en) | Lossless coding and decoding system | |
DE69935811T3 (en) | Frequency domain audio decoding with entropy code mode change | |
DE69734645T2 (en) | DATA PROCESSING OF A BITSTROM SIGNAL | |
DE19544761C2 (en) | Method of compressing an entered symbol | |
DE69918172T2 (en) | ENTROPY CODING FROM VARIABLE TO VARIABLE LENGTH | |
DE69723959T2 (en) | DATA COMPRESSION AND DECOMPRESSION BY RICE ENCODERS / DECODERS | |
DE69628972T2 (en) | MPEG audio decoder | |
DE2031646C2 (en) | Method for compressing image transmission data | |
DE3736193C2 (en) | ||
DE19606178A1 (en) | Compression and decompression of pixel values | |
DE4313175A1 (en) | Audio signal recording using semiconductor memory - over-writing highest hierarchical levels of data already in memory with new data of lower levels when memory is full | |
DE602004006716T2 (en) | PROCESSING PICTURES USING A LIMITED NUMBER OF BITS | |
DE3239607A1 (en) | METHOD AND CIRCUIT ARRANGEMENT FOR CODING OR. DECODING A BINARY DIGITAL SIGNAL | |
EP3624456A1 (en) | Method for compression and decompression of image data | |
DE69724208T2 (en) | Method and device for improving the performance of vector quantization | |
DE112010004844T5 (en) | Video encoding using pixel data streams | |
EP0867074B1 (en) | Method of compressing an analogue signal | |
DE69937761T2 (en) | Arithmetic coding / decoding of a digital information signal | |
EP0027233B1 (en) | Coding method for analog signals | |
DE2818891C2 (en) | ||
EP0728381A1 (en) | Data compressing process and system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AK | Designated states |
Kind code of ref document: A1 Designated state(s): JP US |
|
AL | Designated countries for regional patents |
Kind code of ref document: A1 Designated state(s): AT BE CH DE DK ES FR GB GR IE IT LU MC NL PT SE |
|
WWE | Wipo information: entry into national phase |
Ref document number: 1995932731 Country of ref document: EP |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
WWE | Wipo information: entry into national phase |
Ref document number: 08646358 Country of ref document: US |
|
WWP | Wipo information: published in national office |
Ref document number: 1995932731 Country of ref document: EP |
|
WWW | Wipo information: withdrawn in national office |
Ref document number: 1995932731 Country of ref document: EP |