Results 1 to 3 of 3

Thread: Brute forcing Delta block size

  1. #1
    Member
    Join Date
    Sep 2007
    Location
    Denmark
    Posts
    873
    Thanks
    49
    Thanked 106 Times in 84 Posts

    Brute forcing Delta block size

    i was just trying to experiment with Delta.

    I'm achieving CD images with delta/rep/rzm combo and brute forcing the different combinations

    I always used delta -b768 because cd images tend to be around 600mb to 720mb.
    But lately i found out that this might not be optimalt.

    i tried with brute force -b256 -b384 -b512 & -b768
    and it seems that with these particular cd images the -b384 results in the smallest size (less then a fraction of a percent difference)

    being unaware of the inner workings of compression program (no programming skills) i got curios


    1: what does block size actually determine for delta ?
    2: Is there a guideline or rule of thumb for optimal block size or should i just brute force



    i'll try with rep also as i love doing brute force compression since it improves compression with no to very small decompression penalty.
    i need god compression ratio but fast decompression

  2. #2
    Member Skymmer's Avatar
    Join Date
    Mar 2009
    Location
    Russia
    Posts
    681
    Thanks
    38
    Thanked 168 Times in 84 Posts
    From my point of view there is no rule of a thumb for Delta. I think best value depends on data you're filtering. If you're packing CD images - then such CDs can contain only compressed data (like installer packages or BIK movies) or partialy or fully uncompressed data. I can draw a parallel here with FB parameter of LZMA for example. Its impossible to say which will give the best compression ratio. Of course you can test a terrabyte set and say: "Averagely the value of X is best" but for particular file it can be the best or even the worst.
    I made a small test with SYSTEMSHOCK2.iso image with Delta values starting from 64 and up to 576 with the step of 64 and then compressed with 7-zip v4.65 Ultra (non-solid).
    Original size of image is: 583 903 232 bytes

    Code:
    400 725 282   SYSTEMSHOCK2.iso
    385 967 654   Test_064.dlt
    385 969 675   Test_128.dlt
    386 000 521   Test_192.dlt
    385 938 793   Test_256.dlt
    385 994 856   Test_320.dlt
    385 986 249   Test_384.dlt
    386 020 179   Test_448.dlt
    385 991 889   Test_512.dlt
    385 985 304   Test_576.dlt
    So for this file the best value was 256 while 384 takes only 5th place.

  3. #3
    Member
    Join Date
    Sep 2007
    Location
    Denmark
    Posts
    873
    Thanks
    49
    Thanked 106 Times in 84 Posts
    Thank you for the feedback.

    seems I'll just go ahead and expand my compression to brute force different delta values as well.


    would love to see this feature included in FA under advanced/silly options

Similar Threads

  1. bsc, new block sorting compressor
    By Gribok in forum Data Compression
    Replies: 363
    Last Post: 14th June 2017, 21:55
  2. bzip2 dictionary size
    By Wladmir in forum Data Compression
    Replies: 3
    Last Post: 7th April 2010, 17:09
  3. Delta transformation
    By encode in forum Forum Archive
    Replies: 16
    Last Post: 4th January 2008, 12:13
  4. PackJpg brute force methofs
    By SvenBent in forum Forum Archive
    Replies: 19
    Last Post: 2nd October 2007, 10:11
  5. Block sorting for LZ compression
    By Bulat Ziganshin in forum Forum Archive
    Replies: 15
    Last Post: 14th April 2007, 16:37

Posting Permissions

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