The Data Compression News Blog

All about the most recent compression techniques, algorithms, patents, products, tools and events.

Subscribe

Posts: RSS Feed
Comments: RSS Feed

Sponsored Links

Recent Posts

  • Bijective BWT (6 Comments)

    David Scott has written a bijective BWT transform, which brings all the advantages of bijectiveness to BWT based compressors. Among other things, making BWT more suitable for compression-before-encryption and also give (slightly) better compression.

  • Asymmetric Binary System (113 Comments)

    Jarek Duda’s “Asymmetric Binary System” promises to be an alternate to arithmetic coding, having all the advantages, but being much simpler. Matt has coded a PAQ based compressor using ABS for back-end encoding. Update: Andrew Polar has written an alternate implementation of ABS.

  • Precomp: More Compression for your Compressed Files (3 Comments)

    So many of today’s files are already compressed (using old, outdated algorithms) that newer algorithms don’t even get a chance to touch them. Christian Schneider’s Precomp comes to rescue by undoing the harm.

  • On2 Technologies is Hiring

    There aren’t too many companies working on cutting edge codecs, and of those few this one is hiring. Best of luck.

  • China’s AVS Specifications Available (2 Comments)

    Its old news that China has developed their own Advanced Video Standard to avoid high licensing fees. English translation of the standard is now available, along with the IPR policy. Finally something technical that you can get your hands on to feed your appetite.

PAQ8C and PAQAR 4.5 released

Posted by Sachin Garg on 14th February 2006 | Permanent Link

Matt Mahoney announced that PAQ8C and PAQAR/PAQARCC 4.5 are now available.

I posted PAQ8C by Przemyslaw Skibinski and PAQAR/PAQARCC 4.5 by Alexander Ratushnyak to http://cs.fit.edu/~mmahoney/compression/

On the Calgary corpus, PAQAR 4.5 compresses about 2K smaller than PAQ8C, but PAQ8C is 3 times faster. Both are about a 20K improvement over paq8b. PAQARCC 4.5 is a version of PAQAR 4.5 optimized for the Calgary corpus. It compresses 4K smaller than PAQAR 4.5 and is the top compressor on my benchmark. All programs use dictionaries.

PAQ8C is based on PAQ8B. However, PAQAR 4.5 is based on the PAQ6 core. Future versions of PAQAR will probably be based on PAQ7/8, which replaces the gradient descent model mixer with a neural network.

Below are results for PAQ8C, PAQAR 4.5, and PAQARCC 4.5 each at 2 memory settings (116 and 418 MB for PAQ8C, 191 and 626 MB for the others). The times below are for a 2.2 GHz Athlon-64 in 32 bit mode (WinXP) and 1GB memory.

C:\res\calgary>dir/b | ..\paq8c\paq8c -4 ..\x8c
Enter names of files to compress, followed by blank line
111261 BIB: -> 83112 (text) -> 17160
768771 BOOK1: -> 493825 (text) -> 174567
610856 BOOK2: -> 348174 (text) -> 114098
102400 GEO: -> 44468
377109 NEWS: -> 249870 (text) -> 79313
21504 OBJ1: -> 8006
246814 OBJ2: -> 48510
53161 PAPER1: -> 32438 (text) -> 10568
82199 PAPER2: -> 46144 (text) -> 16082
513216 PIC: -> 23103
39611 PROGC: -> 28455 (text) -> 8456
71646 PROGL: -> 51708 (text) -> 9757
49379 PROGP: -> 40239 (text) -> 7084
93695 TRANS: -> 81099 (binary+text) -> 11357
3141622 -> 572723 (1.4584 bpc) in 119.20 sec (26.355 KB/sec)
Time 119.22 sec, memory 115973024 bytes

C:\res\calgary>dir/b | ..\paq8c\paq8c -6 ..\x8c6
Enter names of files to compress, followed by blank line
111261 BIB: -> 83112 (text) -> 17158
768771 BOOK1: -> 493825 (text) -> 174426
610856 BOOK2: -> 348174 (text) -> 113949
102400 GEO: -> 44441
377109 NEWS: -> 249870 (text) -> 79155
21504 OBJ1: -> 8059
246814 OBJ2: -> 48525
53161 PAPER1: -> 32438 (text) -> 10519
82199 PAPER2: -> 46144 (text) -> 16000
513216 PIC: -> 23089
39611 PROGC: -> 28455 (text) -> 8449
71646 PROGL: -> 51708 (text) -> 9810
49379 PROGP: -> 40239 (text) -> 7138
93695 TRANS: -> 81099 (binary+text) -> 11353
3141622 -> 572265 (1.4572 bpc) in 120.13 sec (26.153 KB/sec)
Time 120.16 sec, memory 417962912 bytes

C:\res\calgary>dir/b | ..\paqar45\paqar -5 ..\xar5
PAQAR (PAQ+Dictionary) v4.5 by M.Mahoney+A.Ratushnyak+P.Skibinski, 12.2.2006
Enter names of files to compress, followed by blank line or EOF.
- loading dictionary C:\res\paqar45\pasqda.dic
- loaded dictionary 103370/559168 words
BOOK2 610856 -> 111362+2109
BOOK1 768771 -> 171145+2913
NEWS 377109 -> 79172+469
PAPER1 53161 -> 10377+81
PAPER2 82199 -> 15849+157
BIB 111261 -> 17280
PROGC 39611 -> 8261
PROGL 71646 -> 9701
PROGP 49379 -> 7000
GEO 102400 -> 44340
OBJ1 21504 -> 7851
OBJ2 246814 -> 45937
PIC 513216 -> 24169
TRANS 93695 -> 11998
570374/3141622 in 373.34 sec. (1.4524 bpc, 18.16% at 8 KB/s)

C:\res\calgary>dir/b | ..\paqar45\paqar -7 ..\xar7
PAQAR (PAQ+Dictionary) v4.5 by M.Mahoney+A.Ratushnyak+P.Skibinski, 12.2.2006
Enter names of files to compress, followed by blank line or EOF.
- loading dictionary C:\res\paqar45\pasqda.dic
- loaded dictionary 103370/559168 words
BOOK2 610856 -> 111212+2109
BOOK1 768771 -> 170995+2913
NEWS 377109 -> 78979+469
PAPER1 53161 -> 10313+81
PAPER2 82199 -> 15791+157
BIB 111261 -> 17315
PROGC 39611 -> 8280
PROGL 71646 -> 9743
PROGP 49379 -> 7061
GEO 102400 -> 44300
OBJ1 21504 -> 7901
OBJ2 246814 -> 45977
PIC 513216 -> 24170
TRANS 93695 -> 11987
569956/3141622 in 369.38 sec. (1.4514 bpc, 18.14% at 9 KB/s)

C:\res\calgary>dir/b | ..\paqar45\paqarcc -5 ..\xcc5
PAQARCC (PAQ+Dictionary) v4.5 by M.Mahoney+A.Ratushnyak+P.Skibinski, 12.2.2006
Enter names of files to compress, followed by blank line or EOF.
- loading dictionary C:\res\paqar45\pasqda.dic
- loaded dictionary 103370/559168 words
BOOK2 610856 -> 111370+2109
BOOK1 768771 -> 171134+2913
NEWS 377109 -> 78702+469
PAPER1 53161 -> 10112+81
PAPER2 82199 -> 15773+157
BIB 111261 -> 17190
PROGC 39611 -> 8280
PROGL 71646 -> 9656
PROGP 49379 -> 6953
TRANS 93695 -> 10501
GEO 102400 -> 44330
OBJ1 21504 -> 7578
OBJ2 246814 -> 45404
PIC 513216 -> 23578
566495/3141622 in 372.28 sec. (1.4426 bpc, 18.03% at 8 KB/s)

C:\res\calgary>dir/b | ..\paqar45\paqarcc -7 ..\xcc7
PAQARCC (PAQ+Dictionary) v4.5 by M.Mahoney+A.Ratushnyak+P.Skibinski, 12.2.2006
Enter names of files to compress, followed by blank line or EOF.
- loading dictionary C:\res\paqar45\pasqda.dic
- loaded dictionary 103370/559168 words
BOOK2 610856 -> 111214+2109
BOOK1 768771 -> 170970+2913
NEWS 377109 -> 78513+469
PAPER1 53161 -> 10107+81
PAPER2 82199 -> 15694+157
BIB 111261 -> 17220
PROGC 39611 -> 8246
PROGL 71646 -> 9695
PROGP 49379 -> 6993
TRANS 93695 -> 10431
GEO 102400 -> 44300
OBJ1 21504 -> 7575
OBJ2 246814 -> 45210
PIC 513216 -> 23567
565669/3141622 in 443.01 sec. (1.4405 bpc, 18.01% at 7 KB/s)

– Matt Mahoney

Leave a Reply

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

This entry was posted on Tuesday, February 14th, 2006 at 6:53 am by Sachin Garg. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.