As can be seen here: Lossy Compressed Image Formats Study,
webp in terms of efficiency.
It would be interesting to see if it is fast, or at least if we can tune the speed.
The implementation used in the benchmarks above is here:
The layout of this project is very windowsy, in other words: bad and messy.
It seems that it produces some static libraries:
$ ls -la lib/libTLibEncoderStatic* -rw-rw-r--. 1 antoine antoine 828198 Dec 27 22:20 lib/libTLibEncoderStatic.a -rw-rw-r--. 1 antoine antoine 6025250 Dec 27 22:20 lib/libTLibEncoderStaticd.a
I assume the "d" is the debug version. The first one has over 3000 symbols in it, the second one over 11000...
The sample encoder app has 2400 lines of code dedicated to configuring the encoder - oh joy.
Is this for real!?
See also #784
Probably not worth the pain right now.
#784 is more likely to succeed.
this ticket has been moved to: https://github.com/Xpra-org/xpra/issues/616