Compress methods

• Introduction
  Compress method identifies the compression algorithm of compressed data which is following the header. Moreover, special compress method that is used when you store a empty directory to an archive, is included lha ver 2.55 or later. Since LHA does not have the identifier which means a start of header like ZIP, '-' on both ends of the compress methods are also has the role to show the position of a header.
  
  
  
• List of compress methods

  compress method	Threshold of LZSS	Max match length	Size of LZSS dictionary	Description
  -lzs-	2	17	2048	Compress method of LArc. It may be used by the old log of computer online services.
  -lz4-	-	-	-	Compress method of LArc. "-lz4-" means no compression.
  -lz5-	3	17	4096	Compress method of LArc. It may be used by the old log of computer online services. It can read and write, without bit stream.
  -lh0-	-	-	-	Compress method of LHA. "-lh0-" means no compression. "-lh0-" is used, if LHA fails the compression.(compressed size is exceeds the original size)
  -lh1-	3	60	4096	Compress method of LHA. "-lh1-" is old standard compress method. LZSS compression and dynamic huffman compression (LZSS uncompressed 1 byte/LZSS match length) and static huffman compression (high 6bit of LZSS match position)
  -lh2-	3	256	8192	Compress method of LHA. It was used to test the improvement of "-lh1-". According to Eriko Tachibana's document, "-lh2-" is used to compress the usage and self extract program in LHA. LZSS compression and dynamic huffman compression (LZSS uncompressed 1 byte/LZSS match length) and dynamic huffman compression (high 7bit of LZSS match position)
  -lh3-	3	256	8192	Compress method of LHA. It was used to test the improvement of "-lh1-". LZSS compression and static huffman compression (LZSS uncompressed 1 byte/LZSS match length) and static huffman compression (high 7bit of LZSS match position)
  -lh4-	3	256	4096	Compress method of LHA. "-lh4-" is used, if memory is not enough to "-lh5-". LZSS compression and static huffman compression (LZSS uncompressed 1 byte/LZSS match length) and static huffman compression (bit length of LZSS match position)
  -lh5-	3	256	8192	Compress method of LHA. "-lh5-" is standard compress method. LZSS compression and static huffman compression (LZSS uncompressed 1 byte/LZSS match length) and static huffman compression (bit length of LZSS match position)
  -lh6-	3	256	32768	Compress method of LHA. It is not recommended that you open an archive which contains the "-lh6-" to the public. LZSS compression and static huffman compression (LZSS uncompressed 1 byte/LZSS match length) and static huffman compression (bit length of LZSS match position)
  -lh7-	3	256	65536	Compress method of LHA. It is not recommended that you open an archive which contains the "-lh7-" to the public. LZSS compression and static huffman compression (LZSS uncompressed 1 byte/LZSS match length) and static huffman compression (bit length of LZSS match position)
  -lhd-	-	-	-	Compress method of LHA. "-lhd-" means no compression. "-lhd-" is used, when you store a empty direcotry to an archive.

  
  
• Other compress methods
  • Joe Jared proposed following compress methods.
    • -lh8-::dictionary size 64k
      
    • -lh9-::dictionary size 128k
      
    • -lha-::dictionary size 256k
      
    • -lhb-::dictionary size 512k
      
    • -lhc-::dictionary size 1024k
      
    • -lhe-::dictionary size 2048k
      
    
    
  • According to document of UNLHA32.DLL, following compress method is used for test.
    • -lhx-::dictionary size 128k-512k
      
    
    
  • According to Eriko Tachibana's document, following compress methods were used in LArc.
    
    • -lz2-::extended "-lzs-". In "-lz2-", you can change dictionary size and max match length.
    • -lz3-::Unknown for details.
      
    • -lz7-::Unknown for details.
      
    • -lz8-::Unknown for details.
      
    
    
  • According to Tatsu's web size, PMarc which is archiver that is created by Miyo uses following compress methods.
    • -pm0-::Unknown for details.
      
    • -pm2-::Unknown for details.
      
  
  
• Reference
  • Joe Jared's document
    http://www.osirusoft.com/joejared/lzhformat.html
    
    
  • Eriko Tachibana's document of LHA header(japanese)
    http://kuwa.omosiro.com/KGARC2/ARC/LHAHED15.ZIP
    
    
  • Satoshi Hiramatsu's header.tex which is attached to LHA for OS/2(japanese)
    http://www02.so-net.ne.jp/~hiramatu/lha/lha2-2.08.lzh
    
    
  • Tatsu's MSX intelligence office(japanese)
    http://member.nifty.ne.jp/Tatsu/TMR/MSX-TIPS.html