Results 1 to 18 of 18

Thread: TC 5.1dev1 is here!

  1. #1
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    3,982
    Thanks
    377
    Thanked 351 Times in 139 Posts
    TC 5.1dev1
    It's my pleasure introduce to you my new compressor. It uses my own variant of a ROLZ data compression algorithm. Note it's just DRAFT and lite-weight version of my new algoritm. Anyway, enjoy!

    Link:
    Download TC 5.1dev1 (26 KB)


  2. #2
    Moderator

    Join Date
    May 2008
    Location
    Tristan da Cunha
    Posts
    2,034
    Thanks
    0
    Thanked 4 Times in 4 Posts
    Thanks Ilia!

  3. #3
    Guest
    can you tell me about thih excelent algorithm

  4. #4
    Moderator

    Join Date
    May 2008
    Location
    Tristan da Cunha
    Posts
    2,034
    Thanks
    0
    Thanked 4 Times in 4 Posts
    I really like this engine!

    If this is a cut down version, then the original must be truly awesome!

  5. #5
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    3,982
    Thanks
    377
    Thanked 351 Times in 139 Posts
    can you tell me about thih excelent algorithm

    This engine can be classified as a fusion of the classic LZ77 algorithm and some context modeling techniques. Currently it's under development, so, in next version you can expect both speed and compression improvements.


  6. #6
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    3,982
    Thanks
    377
    Thanked 351 Times in 139 Posts
    I really like this engine!

    If this is a cut down version, then the original must be truly awesome!


    Thanks! Actually I release the lite version first, to see its performance, including decompression speed, on Benchmarks. If decompression speed is fast enough, probably, I'll release version with some additional features wich can increase compression in cost of decompression speed. Also note since here is an LZ algorithm, I can improve compression/speed of encoder without breaking compatibility - i.e. without any decoder modifications...


  7. #7
    Guest
    Why you tell this algorithm ROLz-like are not LZMA-like or LZCB?
    You used context-dependent base or context-dependent offset/sumbol model.

  8. #8
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    3,982
    Thanks
    377
    Thanked 351 Times in 139 Posts
    Why you tell this algorithm ROLz-like are not LZMA-like or LZCB?

    Since new TC is very similar to ROLZ (Reduced Offset Lempel Ziv). LZMA and LZCB are both utilizes slightly different ideas compared to TC.

    You used context-dependent base or context-dependent offset/sumbol model.

    I will not provide a detailed description of the algorithm, but as you can see from the name "Reduced Offset Lempel Ziv", TC uses some context modeling techniques to "reduce" number of offsets used.


  9. #9
    Guest
    And How match you can increased compression ratio width round speed

  10. #10
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    3,982
    Thanks
    377
    Thanked 351 Times in 139 Posts
    I have one word - noticeable! And using CM, difference in compression will be monstrous. Anyway, currently I'm just waiting for results of current version, to see what I've done.


  11. #11
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    3,982
    Thanks
    377
    Thanked 351 Times in 139 Posts
    Also note, in next versions I'll add some switches to choose compression mode (Fast, Normal, Max).


  12. #12
    Moderator

    Join Date
    May 2008
    Location
    Tristan da Cunha
    Posts
    2,034
    Thanks
    0
    Thanked 4 Times in 4 Posts
    Are you happy with the results now pubished at maximumcompression.com?

  13. #13
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    3,982
    Thanks
    377
    Thanked 351 Times in 139 Posts
    Are you happy with the results now pubished at maximumcompression.com?

    Well, no. But for first try-out version it's okay. Also in next version I will gain compression at any cost... I aim on ROLZ algorithm from mcomp. Okay, let's wait and see...


  14. #14
    Moderator

    Join Date
    May 2008
    Location
    Tristan da Cunha
    Posts
    2,034
    Thanks
    0
    Thanked 4 Times in 4 Posts
    Well, no. But for first try-out version it's okay. Also in next version I will gain compression at any cost... I aim on ROLZ algorithm from mcomp. Okay, let's wait and see...

    Excellent! There is still a chance that PIM could be in the top ten.

  15. #15
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    3,982
    Thanks
    377
    Thanked 351 Times in 139 Posts
    Okay, and what do you think about TC 5.1dev1? Actually, I won't decrease decompression speed - look at MFC, the decompression speed is not too high. The ROLZ like engines have potentially slower decompression speed than the baseline LZ77, but at the same time decompression speed is noticeable faster than compression, though. I can add some additional things like PPM literal coder etc. But decompression speed will be noticeable affected - x2...x4 times slower. So now I'll try to add some LZ related improvements to keep the decompression speed untouched...


  16. #16
    Moderator

    Join Date
    May 2008
    Location
    Tristan da Cunha
    Posts
    2,034
    Thanks
    0
    Thanked 4 Times in 4 Posts
    Okay, and what do you think about TC 5.1dev1?

    I like it but look forward to having the switches to choose compression mode (Fast, Normal, Max) that you mentioned above. With swiches and slight tweaking it could be one of the best around.

  17. #17
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    3,982
    Thanks
    377
    Thanked 351 Times in 139 Posts
    Good news! I add the flexible parsing to TC. This trick increases the compression without affecting decompression speed, furtermore decompressor stays untouched.

    Current TC uses the greedy parsing, wich is faster, but provides less compression. Next version will have the new, flexible parsing, later versions will have a few modes (at least):
    fast (greedy parsing)
    max (flexible parsing)


  18. #18
    Moderator

    Join Date
    May 2008
    Location
    Tristan da Cunha
    Posts
    2,034
    Thanks
    0
    Thanked 4 Times in 4 Posts
    Good news! I add the flexible parsing to TC. This trick increases the compression without affecting decompression speed, furtermore decompressor stays untouched.


Posting Permissions

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