Page 51 of 57 FirstFirst ... 414950515253 ... LastLast
Results 1,501 to 1,530 of 1708

Thread: paq8px

  1. #1501
    Member
    Join Date
    Feb 2016
    Location
    Luxembourg
    Posts
    521
    Thanks
    198
    Thanked 745 Times in 302 Posts

    paq8px_v172

    Code:
    Changes:
    - Slightly improved TextModel
    Quote Originally Posted by LucaBiondi View Post
    Do you know why Gotty and MPais do not post new versions since few days?
    Quote Originally Posted by Darek View Post
    According to Gotty and mpais lack of versions - I don't know - they posted lot's of changes recently, maybe they need to rest and they do it ( ) , maybe they work on something even more extraordinary, new changes, etc.
    I've been really busy lately, and will probably continue to be so in the coming weeks. I usually have quite a few uncommited changes in my internal version, so here's one of them so you guys can test it. I don't really have the time to polish and test other stuff, so for now this will have to do.
    Attached Files Attached Files

  2. The Following 6 Users Say Thank You to mpais For This Useful Post:

    Darek (3rd December 2018),Gotty (10th December 2018),LucaBiondi (12th December 2018),Mike (2nd December 2018),moisesmcardona (3rd December 2018),Stephan Busch (2nd December 2018)

  3. #1502
    Member
    Join Date
    Sep 2014
    Location
    Italy
    Posts
    34
    Thanks
    51
    Thanked 20 Times in 12 Posts
    Welcomeback mpais!

  4. #1503
    Member
    Join Date
    Dec 2008
    Location
    Poland, Warsaw
    Posts
    936
    Thanks
    556
    Thanked 375 Times in 280 Posts
    Scores of my testset for paq8px v172. Looks like textual (or mainly textual) files got about 0.22% of gain in average. Nice!
    I'll test enwik8 then.
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	paq8px_v172.jpg 
Views:	55 
Size:	798.4 KB 
ID:	6312  

  5. The Following 3 Users Say Thank You to Darek For This Useful Post:

    Gotty (10th December 2018),LucaBiondi (12th December 2018),mpais (9th December 2018)

  6. #1504
    Member
    Join Date
    Dec 2008
    Location
    Poland, Warsaw
    Posts
    936
    Thanks
    556
    Thanked 375 Times in 280 Posts
    Scores of paq8px v172 for 4 corpuses - nice gains across all tests! Minor improvements on MmaximumCompression = +0.06%, however other tests got 0.12-0.16% of gain.
    Silesia score is now at the level better than cmix v3 with precomp and is approaching to cmix v4 with precomp!

    And some scores for enwik8:
    16'936'332 - enwik8 -s7eta -Paq8px_v170
    16'493'371 - enwik8.drt -s7eta -Paq8px_v170
    16'518'653 - enwik8 -s9eta -Paq8px_v170
    16'123'717 - enwik8.drt -s9eta -Paq8px_v170
    134'268'613 - enwik9_1423 -s9eta -Paq8px_v170
    130'355'911 - enwik9_1423.drt -s9eta -Paq8px_v170

    16'932'016 - enwik8 -s7eta -Paq8px_v171
    16'488'568 - enwik8.drt -s7eta -Paq8px_v171
    16'520'179 - enwik8 -s9eta -Paq8px_v171
    16'124'799 - enwik8.drt -s9eta -Paq8px_v171

    16'879'890 - enwik8 -s7eta -Paq8px_v172 - 0.30% of improvement
    16'444'222 - enwik8.drt -s7eta -Paq8px_v172 - 0.27% of improvement
    16'471'210 - enwik8 -s9eta -Paq8px_v172 - 0.30% of improvement
    16'081'924 - enwik8.drt -s9eta -Paq8px_v172 - 0.27% of improvement
    Looks that enwk9 should get around 133'9xx'xxx bytes and enwik9.drt around 130'0xx'xxx bytes - both scores are on the 7'th and 6'th place on LTCB respectively!
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	4_Corpuses_paq8px_v172.jpg 
Views:	27 
Size:	2.94 MB 
ID:	6313  
    Last edited by Darek; 4th December 2018 at 12:18.

  7. The Following 2 Users Say Thank You to Darek For This Useful Post:

    Gotty (10th December 2018),LucaBiondi (12th December 2018)

  8. #1505
    Member
    Join Date
    Sep 2014
    Location
    Italy
    Posts
    34
    Thanks
    51
    Thanked 20 Times in 12 Posts
    My big testcase results: v171 vs v172

    Click image for larger version. 

Name:	Pax8px_171_172.png 
Views:	68 
Size:	71.2 KB 
ID:	6318

    PDF files gain about 1.7 KB
    XML files gain about 1 KB
    TXT files gain about 41 KB !
    ISO files gain about 3 KB

    MP3 files lose few hundred bytes


    Click image for larger version. 

Name:	Overall_fino_172.png 
Views:	60 
Size:	79.0 KB 
ID:	6319

    Click image for larger version. 

Name:	Elenco2.png 
Views:	79 
Size:	102.2 KB 
ID:	6320


    NEW Overall record!
    NEW PDF record!
    NEW DOC record!
    NEW MP4 record!
    NEW XML record! (first tiime below 420.000 byte)

    Thank Mpais!!
    Luca
    Last edited by LucaBiondi; 5th December 2018 at 23:45.

  9. The Following 2 Users Say Thank You to LucaBiondi For This Useful Post:

    Darek (5th December 2018),Gotty (10th December 2018)

  10. #1506
    Member
    Join Date
    Dec 2008
    Location
    Poland, Warsaw
    Posts
    936
    Thanks
    556
    Thanked 375 Times in 280 Posts
    @Luca - attached files/tables don't open.

  11. The Following User Says Thank You to Darek For This Useful Post:

    LucaBiondi (12th December 2018)

  12. #1507
    Member
    Join Date
    Sep 2014
    Location
    Italy
    Posts
    34
    Thanks
    51
    Thanked 20 Times in 12 Posts
    Sorry..my fault!

  13. #1508
    Member
    Join Date
    Dec 2008
    Location
    Poland, Warsaw
    Posts
    936
    Thanks
    556
    Thanked 375 Times in 280 Posts
    Some scores for enwik9 for v172:
    134'268'613 - enwik9_1423 -s9eta -Paq8px_v170
    130'355'911 - enwik9_1423.drt -s9eta -Paq8px_v170


    133'708'688 - enwik9_1423 -s9eta -Paq8px_v172 - 0.43% of gain - very nice improvement!
    129'893'797 - enwik9_1423.drt -s9eta -Paq8px_v172 - 0.36% of gain, first score for paq8px below 130'000'000 bytes! Full package to LTCB = 130'240'384 => only 55KB worse score than 5'th place!

  14. The Following 3 Users Say Thank You to Darek For This Useful Post:

    Gotty (10th December 2018),LucaBiondi (6th December 2018),mpais (9th December 2018)

  15. #1509
    Member
    Join Date
    Feb 2018
    Location
    Canton, China
    Posts
    2
    Thanks
    0
    Thanked 1 Time in 1 Post
    For multiple files, sendto this bat:
    Code:
    @echo off
    echo\>%~dp1L
    for %%a in (%*) do (
    echo %%~nxa
    )>> %~dp1L
    paq8px -9 @%~dp1L %~dp1
    del %~dp1L
    pause
    Last edited by CPU; 7th December 2018 at 20:19.

  16. The Following User Says Thank You to CPU For This Useful Post:

    Gotty (10th December 2018)

  17. #1510
    Member
    Join Date
    Feb 2016
    Location
    Luxembourg
    Posts
    521
    Thanks
    198
    Thanked 745 Times in 302 Posts

    paq8px_v173

    Code:
    Changes:
    - Slightly improved models:
     - recordModel
     - TextModel
    Just another small update, mostly based on the same idea as the last version.
    Attached Files Attached Files

  18. The Following 6 Users Say Thank You to mpais For This Useful Post:

    Darek (9th December 2018),Gotty (10th December 2018),kaitz (24th December 2018),LucaBiondi (12th December 2018),Mike (9th December 2018),moisesmcardona (13th December 2018)

  19. #1511
    Member
    Join Date
    Dec 2008
    Location
    Poland, Warsaw
    Posts
    936
    Thanks
    556
    Thanked 375 Times in 280 Posts
    Scores of my testset for paq8px v173 - another 11KB of gain of total score, good improvements for K.WAD, L.PAK, Q.WK3.
    Very good gain for M.DBF file => 0.45%! Textual files got in average 0.05% of gain.
    In total this version is slowly approaching to beat 10'900'000 edge.
    Total score of paq8px_v173 is very close to cmix_v13n version also.
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	paq8px_v173.jpg 
Views:	49 
Size:	798.7 KB 
ID:	6323  

  20. The Following 2 Users Say Thank You to Darek For This Useful Post:

    Gotty (10th December 2018),LucaBiondi (12th December 2018)

  21. #1512
    Member
    Join Date
    Dec 2008
    Location
    Poland, Warsaw
    Posts
    936
    Thanks
    556
    Thanked 375 Times in 280 Posts
    4 Corpuses test for paq8px_v173 - for smaller sets there are a quite nice gains (0.12% and 0.18%), however for Silesia there is 0.07% of gain, and 0.01% for MaximumCompression.
    And some enwik8 scores:

    16'879'890 - enwik8 -s7eta -Paq8px_v172
    16'444'222 - enwik8.drt -s7eta -Paq8px_v172
    16'471'210 - enwik8 -s9eta -Paq8px_v172
    16'081'924 - enwik8.drt -s9eta -Paq8px_v172


    16'870'161 - enwik8 -s7eta -Paq8px_v173 - 0.06% of gain
    16'439'008 - enwik8.drt -s7eta -Paq8px_v173 - 0.03% of gain
    16'461'020 - enwik8 -s9eta -Paq8px_v173 - 0.06% of gain
    16'077'318 - enwik8.drt -s9eta -Paq8px_v173 - 0.03% of gain

    My estimate of enwik9.drt is still a bit behind 5'th place of LTCB
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	4_Corpuses_paq8px_v173.jpg 
Views:	31 
Size:	3.05 MB 
ID:	6325  

  22. The Following User Says Thank You to Darek For This Useful Post:

    LucaBiondi (12th December 2018)

  23. #1513
    Member
    Join Date
    Sep 2014
    Location
    Italy
    Posts
    34
    Thanks
    51
    Thanked 20 Times in 12 Posts
    My big testcase results: v172 vs v173

    Click image for larger version. 

Name:	Pax8px_172_173.png 
Views:	60 
Size:	81.4 KB 
ID:	6328

    PDF files gain about 1 KB
    MP3 files gain near 3 KB
    TXT files gain near 7 KB
    ISO files gain near 5 KB

    Click image for larger version. 

Name:	elenco_fino_173.png 
Views:	63 
Size:	134.7 KB 
ID:	6329

    New overall record
    Best JPG record owned by v169b version
    New record for PDF files
    New record for DOC files
    New record for MP3 files
    New record for MP4 files
    New record for XML files
    New record for TXT files
    New record for BAK files
    Record for ZIP file owned by v168 version
    New record for EXE files
    New record for ISO files

    Bye
    Luca

  24. The Following User Says Thank You to LucaBiondi For This Useful Post:

    Darek (12th December 2018)

  25. #1514
    Member
    Join Date
    Feb 2016
    Location
    Luxembourg
    Posts
    521
    Thanks
    198
    Thanked 745 Times in 302 Posts

    paq8px_v174

    Code:
    Changes:
    - Additional mixer context for the following models:
     - SparseMatchModel
     - 8bpp images, both grayscale and indexed color
     - 24/32bpp images
    - Slightly improved recordModel
    - New RLE transform, currently used only on 8bpp Targa images
    Attached Files Attached Files

  26. The Following 6 Users Say Thank You to mpais For This Useful Post:

    Darek (6th January 2019),Gotty (6th January 2019),kaitz (6th January 2019),LucaBiondi (6th January 2019),Mike (6th January 2019),moisesmcardona (6th January 2019)

  27. #1515
    Member
    Join Date
    Dec 2008
    Location
    Poland, Warsaw
    Posts
    936
    Thanks
    556
    Thanked 375 Times in 280 Posts
    Here are my testset scores. In general - very good improvement!

    The biggest gain is for D.TGA file which is a 8bit Targa image - 88KB (almost 29%) of gain! Of course that means this version of paq8px break record for this file. New RLE transform is amazing!

    Fine gains on I.EXE file (2.77% and another new record!), K.WAD (0.24%) and L.PAK (0.42% and yet another record! The best score is for -7t option = 2'595'597 bytes - first compressor which goes below 2'600'000 bytes!)

    1.BMP image and 0.WAV audio files also got nice improvements.

    In total this version gain is about 113'000 bytes which means paq8px goes below 10'900'000 bytes for total test and it's close to break 10'800'000 bytes barrier!
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	paq8px_v174.jpg 
Views:	33 
Size:	628.8 KB 
ID:	6363  

  28. The Following 3 Users Say Thank You to Darek For This Useful Post:

    Gotty (6th January 2019),LucaBiondi (6th January 2019),mpais (8th January 2019)

  29. #1516
    Member
    Join Date
    Dec 2008
    Location
    Poland, Warsaw
    Posts
    936
    Thanks
    556
    Thanked 375 Times in 280 Posts
    And the scores of 4 Corpuses by paq8px_v174.
    Nice gain on Silesia = 0.18%. Other corpuses got 0.05-0.08% of gain.
    Silesia score is at the level of cmix v6 with precompressor!
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	paq8px_v174_4_Corpuses.jpg 
Views:	23 
Size:	2.34 MB 
ID:	6367  

  30. The Following 2 Users Say Thank You to Darek For This Useful Post:

    Gotty (8th January 2019),LucaBiondi (9th January 2019)

  31. #1517
    Member Gotty's Avatar
    Join Date
    Oct 2017
    Location
    Hungary
    Posts
    376
    Thanks
    252
    Thanked 260 Times in 140 Posts
    Wow! Nice!

  32. The Following User Says Thank You to Gotty For This Useful Post:

    LucaBiondi (9th January 2019)

  33. #1518
    Member
    Join Date
    Feb 2016
    Location
    Luxembourg
    Posts
    521
    Thanks
    198
    Thanked 745 Times in 302 Posts

    paq8px_v175

    Code:
    Changes:
    - New linearPredictionModel
    - Bug fix in Targa image detection
    Attached Files Attached Files

  34. The Following 4 Users Say Thank You to mpais For This Useful Post:

    Darek (8th January 2019),Gotty (8th January 2019),kaitz (11th January 2019),LucaBiondi (8th January 2019)

  35. #1519
    Member
    Join Date
    Dec 2008
    Location
    Poland, Warsaw
    Posts
    936
    Thanks
    556
    Thanked 375 Times in 280 Posts
    Scores of paq8px v175 for my testset: next very good improvement - only for particular files but quite big:
    0.WAV - audio file got 13'721 bytes of gain
    K.WAD - got another 21'034 bytes of gain
    L.PAK - 8;938 bytes of gain
    N.ADX - 273 bytes (0.34%)

    Total there is almost 44'000 byets of improvement in total - that means this time paq8px got a total score below 10'800'000 bytes and this score is only 0.34% behind latest cmix v16e! Great!
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	paq8px_v175.jpg 
Views:	33 
Size:	629.0 KB 
ID:	6369  

  36. The Following 3 Users Say Thank You to Darek For This Useful Post:

    Gotty (8th January 2019),LucaBiondi (8th January 2019),mpais (9th January 2019)

  37. #1520
    Programmer schnaader's Avatar
    Join Date
    May 2008
    Location
    Hessen, Germany
    Posts
    549
    Thanks
    205
    Thanked 181 Times in 86 Posts
    Recently stumbled upon a paper about compression of dithered images. Since the latest paq8px changes often involved 8-bit image improvements and the transform used in the paper ("pseudo distance transform", PDT) doesn't look too hard to implement, thought I could post it here and ask for your opinions. Dithered images are quite different and less compressible, so they would deserve a seperate model, but it seems detection would be hard. On the other hand, the other referenced papers and some of the other work of the main writer - unfortunately, most of it is paywalled without academic access - don't seem to differentiate between color-mapped (= images using palettes) and dithered, so the transform could be of general use for 8-bit images. Of course, it's not sure if paq models would benefit from it at all (or if some similar transform is already used in it ), but I guess it's worth a try?
    http://schnaader.info
    Damn kids. They're all alike.

  38. The Following User Says Thank You to schnaader For This Useful Post:

    LucaBiondi (8th January 2019)

  39. #1521
    Member
    Join Date
    Oct 2010
    Location
    Germany
    Posts
    278
    Thanks
    6
    Thanked 29 Times in 18 Posts
    Quote Originally Posted by mpais View Post
    Code:
    Changes:
    - New linearPredictionModel
    care to elaborate? I checked the OLS class and it looks the same, or did you changed the causal neighbourhood?

  40. The Following User Says Thank You to Sebastian For This Useful Post:

    LucaBiondi (8th January 2019)

  41. #1522
    Member
    Join Date
    Sep 2014
    Location
    Italy
    Posts
    34
    Thanks
    51
    Thanked 20 Times in 12 Posts
    Hi All,

    Here my big testcase results: v173 vs v174

    Click image for larger version. 

Name:	paq8px_173_174.png 
Views:	42 
Size:	98.9 KB 
ID:	6371
    New overall record!

    Gain: ISO 16 KB, PDF 2KB, BAK 1KB, EXE 2 KB
    Loose: TXT 6 KB


    Click image for larger version. 

Name:	paq_173_174_all.png 
Views:	37 
Size:	128.8 KB 
ID:	6372

    New overall Record!
    New record for PDF, DOC, MP3, MP4, BAK, EXE, ISO!


    Paq8px Rocks!!!

    Bye Luca

    OT: Me and my Son adjust a gameboy cartridge! https://www.youtube.com/watch?v=PLFHyM5FLwQ
    Last edited by LucaBiondi; 8th January 2019 at 23:36.

  42. The Following 2 Users Say Thank You to LucaBiondi For This Useful Post:

    Darek (8th January 2019),moisesmcardona (9th January 2019)

  43. #1523
    Member
    Join Date
    Jun 2009
    Location
    Puerto Rico
    Posts
    181
    Thanks
    68
    Thanked 15 Times in 11 Posts
    Found my old GUI. I redesigned it . Also had to modify paq8px screen printing a bit so the logging can work

    Click image for larger version. 

Name:	PAQCompress v0.1.png 
Views:	68 
Size:	42.3 KB 
ID:	6373

    PAQCompress repo: https://github.com/moisesmcardona/PAQCompress
    paq8px fork with stdout adjustments: https://github.com/moisesmcardona/paq8px

  44. The Following 3 Users Say Thank You to moisesmcardona For This Useful Post:

    CompressMaster (10th February 2019),Darek (9th January 2019),spark (9th January 2019)

  45. #1524
    Member
    Join Date
    Dec 2008
    Location
    Poland, Warsaw
    Posts
    936
    Thanks
    556
    Thanked 375 Times in 280 Posts
    4 Corpuses scores for paq8px v175. No bug changes, a bit better scores for Silesia and Maximum Compression.
    Other good score is that even for -4 option this version goes below 10'900'000 bytes = 10'874'467! Impressed!
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	paq8px_v175_4_Corpuses.jpg 
Views:	27 
Size:	2.98 MB 
ID:	6374  

  46. The Following 2 Users Say Thank You to Darek For This Useful Post:

    Gotty (9th January 2019),kaitz (11th January 2019)

  47. #1525
    Member
    Join Date
    Feb 2016
    Location
    Luxembourg
    Posts
    521
    Thanks
    198
    Thanked 745 Times in 302 Posts
    Quote Originally Posted by schnaader View Post
    Recently stumbled upon a paper about compression of dithered images. Since the latest paq8px changes often involved 8-bit image improvements and the transform used in the paper ("pseudo distance transform", PDT) doesn't look too hard to implement, thought I could post it here and ask for your opinions. Dithered images are quite different and less compressible, so they would deserve a seperate model, but it seems detection would be hard. On the other hand, the other referenced papers and some of the other work of the main writer - unfortunately, most of it is paywalled without academic access - don't seem to differentiate between color-mapped (= images using palettes) and dithered, so the transform could be of general use for 8-bit images. Of course, it's not sure if paq models would benefit from it at all (or if some similar transform is already used in it ), but I guess it's worth a try?
    I skimmed over the 2 papers, and it's clear that whilst it may be an interesting idea for the purpose they're trying to catter to, it isn't applicable to something like paq8px/cmix/etc, which are always going for much higher compression density. I managed to find some of the test images they used, and a few quick tests show that even the simple 8bpp indexed-color model we're using now gives about 40 to 80% higher density. Their transform is useful since they're comparing against formats with LZ-style backends (GIF, PNG), which are really ill suited for image compression, but are practical in terms of speed.
    Using the transform on paq8px would most likely hurt compression, since we rely on context modelling based on sparse pixel neighbourhoods, which would be affected by the transform trying to decorrelate them.

    Quote Originally Posted by Sebastian View Post
    care to elaborate? I checked the OLS class and it looks the same, or did you changed the causal neighbourhood?
    It's a new model in the context-mixing sense, i.e., a model that produces bitwise predictions, not a linear prediction model outputting whole symbol predictions per-se. It's useful on files where you'd usually use some sort of delta transform prior to encoding with an LZ/BWT/etc engine.
    It's currently using 5 linear predictors (3 ols and 2 lagrange) which provide the symbol predictions, which are then combined into bitwise predictions that, when mixed, give much better compression than any single predictor in isolation. It's the standard way of doing this in context-mixing, like we're using in the image models. Do you use something different in SAC?

  48. The Following User Says Thank You to mpais For This Useful Post:

    schnaader (9th January 2019)

  49. #1526
    Member
    Join Date
    Dec 2008
    Location
    Poland, Warsaw
    Posts
    936
    Thanks
    556
    Thanked 375 Times in 280 Posts
    @moisesmcardona - is it possible to define with this GUI custom options and use it for particular files? And to make batch compression for defined testset with these options?
    @mpais - is it possible to use new RLE transform to other image filetypes (24bpp, other image containers)?

  50. #1527
    Member
    Join Date
    Jun 2009
    Location
    Puerto Rico
    Posts
    181
    Thanks
    68
    Thanked 15 Times in 11 Posts
    Quote Originally Posted by Darek View Post
    @moisesmcardona - is it possible to define with this GUI custom options and use it for particular files? And to make batch compression for defined testset with these options?
    @mpais - is it possible to use new RLE transform to other image filetypes (24bpp, other image containers)?
    Currently, it supports the flags from paq8px. It does not generate batch files. It also takes care of generating the file list text file and pass it to paq8px to compress files and folders with their original names.

    What options would you like to see?

    Also, just did a compression test on one of my scanned PDF documents and paq8px v175 beated cmix by 4kb. Both using the english dictionary file.

  51. #1528
    Member Gotty's Avatar
    Join Date
    Oct 2017
    Location
    Hungary
    Posts
    376
    Thanks
    252
    Thanked 260 Times in 140 Posts
    Quote Originally Posted by moisesmcardona View Post
    Found my old GUI. I redesigned it . Also had to modify paq8px screen printing a bit so the logging can work

    PAQCompress repo: https://github.com/moisesmcardona/PAQCompress
    paq8px fork with stdout adjustments: https://github.com/moisesmcardona/paq8px
    I compiled your paq8px fork, and it looks like the percentages (the progress indicator) are printed line by line, which is not desirable, i suppose.

    Code:
    >paq8px.exe -0 english.exp
    paq8px archiver v175a (C) 2019, Matt Mahoney et al.
    
    Creating archive english.exp.paq8px175a in single file mode...
    
    Filename: english.exp (13054 bytes)
    Block segmentation:
     0           | text - eol       |     13054 bytes [0 - 13053]
    Compressing...
      0.00%
     34.66%
     69.33%
    -----------------------
    Total input size     : 13054
    Total archive size   : 11829
    
    Time 0.28 sec, used 419 MB (439978893 bytes) of memory
    Also, when redirecting the screen output to a log file (like "paq8px.exe -0 english.exp >log.txt") no progress indicator should be printed to the log file.
    Could you fix that?

  52. #1529
    Member
    Join Date
    Jun 2009
    Location
    Puerto Rico
    Posts
    181
    Thanks
    68
    Thanked 15 Times in 11 Posts
    Quote Originally Posted by Gotty View Post
    I compiled your paq8px fork, and it looks like the percentages (the progress indicator) are printed line by line, which is not desirable, i suppose.

    Code:
    >paq8px.exe -0 english.exp
    paq8px archiver v175a (C) 2019, Matt Mahoney et al.
    
    Creating archive english.exp.paq8px175a in single file mode...
    
    Filename: english.exp (13054 bytes)
    Block segmentation:
     0           | text - eol       |     13054 bytes [0 - 13053]
    Compressing...
      0.00%
     34.66%
     69.33%
    -----------------------
    Total input size     : 13054
    Total archive size   : 11829
    
    Time 0.28 sec, used 419 MB (439978893 bytes) of memory
    Also, when redirecting the screen output to a log file (like "paq8px.exe -0 english.exp >log.txt") no progress indicator should be printed to the log file.
    Could you fix that?
    Yes, I did the percent-per-line adjustment to output to the GUI in realtime, and also had to add additional flushing. The mods are only intended to be used by my GUI. For other general usage, the standard paq8px build should be used.

  53. #1530
    Member Gotty's Avatar
    Join Date
    Oct 2017
    Location
    Hungary
    Posts
    376
    Thanks
    252
    Thanked 260 Times in 140 Posts
    Quote Originally Posted by moisesmcardona View Post
    Also, just did a compression test on one of my scanned PDF documents and paq8px v175 beated cmix by 4kb. Both using the english dictionary file.
    I'm not familiar with testing cmix (can't run it on my pc), but I have a feeling that precomp+cmix is the way to go in this case.

Page 51 of 57 FirstFirst ... 414950515253 ... LastLast

Similar Threads

  1. FrontPAQ - GUI frontend for PAQ8PF and PAQ8PX
    By LovePimple in forum Download Area
    Replies: 26
    Last Post: 17th January 2019, 13:36
  2. Alternative paq8px builds
    By M4ST3R in forum Download Area
    Replies: 20
    Last Post: 25th June 2010, 16:19
  3. Optimized paq7asm.asm code not compatible with paq8px?
    By M4ST3R in forum Data Compression
    Replies: 7
    Last Post: 3rd June 2009, 15:34

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •