These patches contain both a Ut Video encoder that creates
files that the official reference implementation likes, and
that in most cases decode just fine on the libavcodec decoder
as well (more on this later*), as well as FATE tests for the
encoder.

All colorspaces are supported (4:2:0/4:2:2 YCbCr, RGB, RGBA),
all known prediction modes are supported (none, left, median),
multi-slice encoding works, but is hardcoded to one slice
until slice-based multithreading is implemented.

This is my first bigger amount of code in C, and I welcome
all kinds of comments and reviews on it. The decoder's
tests all pass, and make checkheaders passes on utvideo.h.
I have not yet been able to test it with valgrind, as I
develop on Windows.

The tests are based on how Derek added tests for v410enc.

* libavcodec's decoder seems to have problems with the
  single symbol encoding mode, reference implementation
  decodes it just fine (md5 hash mismatch happens with
  avconv). Looking at the decoded picture via ffms2 in
  Aegisub shows that the output does seem to be
  alright, thus I am not wholly sure what exactly is
  throwing a fit. I might look into it after I finish
  multithreading on the encoder side of things.


Jan


_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to