|
|
|
|
|
|
by klauspost
1376 days ago
|
|
|
There is no way to "rule some levels out". The levels only differ in the quality of the lz matches. Slower levels search further potential matches - you cannot see that from a compressed stream. You can compare the output from each level and that will be your guess, assuming of course that zlib was used in the first place. |
|
|
If you see a match that would not be visible from given compression level, or conversely if you don't see a match that should have been selected at given compression level, then it is clear that the level couldn't be used to make that compressed stream. I think preflate actually does this, maintaining multiple hash tables over the course.
That said, I see where my explanation was slightly off in hindsight (sorry!). A bitstream only composed of blocks with the static Huffman table can be produced in multiple ways. One such case is of course a very specific input that is more beneficial to not use dynamic Huffman tables (that said, zlib tallying is AFAIK greedy so this should be very rare). Another case is Z_FIXED, which was actually what I had in mind at that writing---I should have said compression parameters instead. But I stand by the claim that many parameter combinations can be excluded from partial or full decompression, and for the practical purpose this is enough to be classified as the "estimation".