Results 1 to 15 of 15

Thread: LZW, LZMW and LZAP comparison

Hybrid View

Previous Post Previous Post   Next Post Next Post
  1. #1
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    4,009
    Thanks
    399
    Thanked 397 Times in 152 Posts

    Arrow LZW, LZMW and LZAP comparison

    After testing LZW with a huge dictionary I decided to additionally test some LZW modifications: LZMW and LZAP. All modifications are identical - except dictionary handling - thus we may see the real difference of performance on various data types. So, LZ-codes (or dictionary pointers) are coded using flat codes, all files coded with no dictionary reset.
    book1
    LZW: 314,543 bytes
    LZMW: 317,227 bytes
    LZAP: 306,942 bytes

    calgary.tar
    LZW: 1,307,956 bytes
    LZMW: 1,243,481 bytes
    LZAP: 1,221,904 bytes

    world95.txt
    LZW: 947,113 bytes
    LZMW: 764,013 bytes
    LZAP: 751,613 bytes

    fp.log
    LZW: 2,009,356 bytes
    LZMW: 1,074,090 bytes
    LZAP: 1,147,196 bytes

    english.dic
    LZW: 1,607,939 bytes
    LZMW: 1,815,974 bytes
    LZAP: 1,646,121 bytes

    A10.jpg
    LZW: 990,614 bytes
    LZMW: 1,081,968 bytes
    LZAP: 1,030,322 bytes

    3200.txt
    LZW: 5,642,451 bytes
    LZMW: 5,607,396 bytes
    LZAP: 5,475,965 bytes

    pht.psd
    LZW: 4,880,558 bytes
    LZMW: 3,240,734 bytes
    LZAP: 2,931,192 bytes

  2. #2
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    4,009
    Thanks
    399
    Thanked 397 Times in 152 Posts
    Anyway, LZ77 is supreme. Especially if we talk about LZ77/LZSS with S&S or even Optimal Parsing. Even simple scheme with byte-aligned output (no BitIO, Huffman or Arithmetic coding) may easily beat LZW even on text files...

  3. #3
    Member
    Join Date
    May 2008
    Location
    brazil
    Posts
    163
    Thanks
    0
    Thanked 3 Times in 3 Posts
    Quote Originally Posted by encode View Post
    Anyway, LZ77 is supreme. Especially if we talk about LZ77/LZSS with S&S or even Optimal Parsing. Even simple scheme with byte-aligned output (no BitIO, Huffman or Arithmetic coding) may easily beat LZW even on text files...

    It's possilble to use Some LZT Techiniques to test new model of LZ algos?

    http://compgt.googlepages.com/lz77

  4. #4
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    4,009
    Thanks
    399
    Thanked 397 Times in 152 Posts
    Quote Originally Posted by Modeling for Text Compression
    3.4.9 LZT
    LZT [Tischer 1987] is based on LZC. The
    main difference is that once the dictionary
    is full, space is made for new phrases by
    discarding the least recently used phrase
    (LRU replacement). This is performed efficiently
    by maintaining phrases in a selforganizing
    list indexed by a hash table. The
    list is designed so that a phrase can be
    replaced in a small, bounded number of
    pointer operations. Because of the extra
    housekeeping, this algorithm is a little
    slower than LZC, but the more sophisticated
    choice of phrases in the dictionary
    means that it can achieve the same
    compression ratio with less memory.
    LZT also codes phrase numbers slightly
    more efficiently than LZC by using a
    slightly better method of phasing the binary
    encoding. (This improvement could
    also be applied to several other LZ algorithms.)
    A little extra effort is required of
    the encoder and decoder, but it is insignificant
    compared with the task of searching
    and maintaining the LRU list. The second
    variation of Miller and Wegman [1984] is
    an independent discovery of LZT.
    Like you see in my implementations listed above the dictionary reset even not performed. Well, theoretically, it is possible to implement LZW/LZMW/LZAP with such LRU dictionary...

  5. #5
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    4,009
    Thanks
    399
    Thanked 397 Times in 152 Posts
    Phasing the binary encoding or phased-in codes may be applied, although I never tested such trick...

  6. #6
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    4,009
    Thanks
    399
    Thanked 397 Times in 152 Posts
    Some paper for interested audience:
    Attached Files Attached Files

  7. #7
    Member
    Join Date
    Jul 2006
    Location
    US
    Posts
    39
    Thanks
    26
    Thanked 1 Time in 1 Post
    Quote Originally Posted by encode View Post
    After testing LZW with a huge dictionary I decided to additionally test some LZW modifications: LZMW and LZAP.
    Is this LZAP by Mike Goldman? If so, which version were you running?

  8. #8
    The Founder encode's Avatar
    Join Date
    May 2006
    Location
    Moscow, Russia
    Posts
    4,009
    Thanks
    399
    Thanked 397 Times in 152 Posts
    Quote Originally Posted by spark View Post
    Is this LZAP by Mike Goldman? If so, which version were you running?
    Nope. This is my own implementation of LZAP compression algorithm.

  9. #9
    Member
    Join Date
    Jun 2017
    Location
    Bangladesh
    Posts
    7
    Thanks
    0
    Thanked 0 Times in 0 Posts
    How many versions of LZAP available ? Can You give some link about latest LZAP ?

  10. #10
    Member nikkho's Avatar
    Join Date
    Jul 2011
    Location
    Spain
    Posts
    550
    Thanks
    222
    Thanked 165 Times in 106 Posts
    Quote Originally Posted by imransuet View Post
    How many versions of LZAP available ? Can You give some link about latest LZAP ?
    Please, do not duplicate posts: https://encode.su/threads/2799-What-...3561#post53561

Similar Threads

  1. Comparison of lossless PNG compression tools
    By Surfer in forum Data Compression
    Replies: 54
    Last Post: 19th September 2011, 22:58
  2. exe prefilter quick comparison
    By evg in forum Data Compression
    Replies: 7
    Last Post: 23rd May 2009, 16:20
  3. Comparison of the recent CM demo coders
    By Shelwien in forum Data Compression
    Replies: 38
    Last Post: 13th June 2008, 13:21
  4. Original paper on LZMW
    By encode in forum Forum Archive
    Replies: 3
    Last Post: 9th February 2008, 13:34
  5. LZW v0.1 is here!
    By encode in forum Forum Archive
    Replies: 20
    Last Post: 2nd February 2008, 13:46

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
  •