Re: Compression test on permutations

Sportman ha scritto:
On Apr 24, 1:58 am, erpy <i...@xxxxxxxxxxxxxxxx> wrote:
I suppose common compressors (Zip, Rar) would not compress the file that
much (each byte only occurs 256 times throughout the file, with equal
probability in each string, so LZW-like compressors shouldn't achieve a
high ratio, in theory). I will convert it in bytes and run these tests.
I get this:

Input:
65,536 bytes

Output:
Dark default = 66,666 bytes
ZIP max = 65,662 bytes
RAR max = 65,609 bytes
SBC -m3 = 65,589 bytes
GZIP default = 65,574 bytes
Durilca default = 63,028 bytes
PPMonstr default = 62,873 bytes
PAQ8L -2 or -3 = 61,591 bytes


Thanks Sportman, very interesting results indeed.
Impressive compressors those 3 that managed to take off up to ~4 Kb, considering they're "universal".

Best,
E.