2020.04.18 13:28 "Re: [Tiff] Updating zlib library", by Bob Friesenhahn

interesting. From my own perspective, I'd be more interested in *faster* alternatives to zlib than slightly better compression rates. Looking quickly at the links you provide, libdeflate advertize greater speeds, but I'd be interested in seing benchmarks.

Speed is definitely an important factor. I have always noticed that libtiff's own LZW compression is much faster than zlib's deflate.

For applications that use TIFF in several stages of complex processing, I now see ZSTD being used for intermediate steps due to its uncomparable speed * compression_ratio, and DEFLATE/LZW being used for the final user-facing product to maximize compatibility as ZSTD-compressed TIFF are presumably non-standard, outside of libtiff ecosystem.

Zstd is fast and remarkably flexible. It would be good if Adobe would wake up and officially bless Zstd as implemented in libtiff.

If absolute compression ratio is most important, then lossless WebP seems to be the leader (by far!) and in my quick test it was only 4X slower than zlib's deflate algorithm. Compression algorithms designed specifically for images will almost always beat generic algorithms when it comes to compression ratio. This another case where it would be good for Adobe to wake up and officially bless WebP compression as implemented in libtiff.

In today's world, small gains in absolute lossless compression levels are not worth very much. Decoding and encoding speeds are often much more important (especially decoding).

Bob

Bob Friesenhahn
bfriesen@simple.dallas.tx.us, http://www.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer, http://www.GraphicsMagick.org/
Public Key, http://www.simplesystems.org/users/bfriesen/public-key.txt