description by the author
I've named the compressor FBC, and it is a BWT compressor. It uses an alphabet reordering algorithm before the BWT stage. After the BWT, it arithmetic codes the data using a quickly adapting model with 14 bits of context: 11 bits of history and 3 bits for the position within the current byte. No MTF or SR transforms are used after the BWT stage, since they seemed to hurt the compression ratio. Version 1.1 automatically selects between the 32 and 64 bit versions of DivSufSort (Yuta Mori).
About v1.2: It fixes another decompression issue and also gives an optional flag to enable MTF. Some good parameters to test are with the default blocksize and 100000000 byte blocks, with and without the MTF stage.
About v1.3: It now has selectable EXE/delta filters and optional input file type detection. The compression engine is also now designed so that files produced with this version will also be decompressable by later versions.
About v1.3b: It improves filetype detection in automatic mode.
(Have info that should be added here? E-mail.)
|Ver||Rating||CPR||DPR||S.E.||R.E.||Ratio||C. kB/s||D. kB/s|