daala encoding

[Antoine Martin]

​http://xiph.org/daala/

Should not be too hard to implement.

[Antoine Martin]

stub support for daala

[Antoine Martin]

work in progress patch

[Antoine Martin]

nearly there - as in, it no longer crashes all the time

[Antoine Martin]

latest daala patch: encoding seems to work, decoding does not

[Antoine Martin]

patch to the daala example player to log every ogg/daala call

[Antoine Martin]

shows the result of encoding an xterm session using daala

[Antoine Martin]

stream recorded on the server showing an xterm

[Antoine Martin]

With v6, we can encode using daala, and to make it easier to debug we can also record the output stream to a file (which should be generalized) as well as sending it to the client using:

XPRA_DAALA_SAVE_STREAM=basefilename XPRA_DAALA_DEBUG=1 xpra start --start-child=xterm :10

Then if I view the stream saved by the server using daala's example player I see:

So it looks like daala may have some serious encoding problems!

Unfortunately, the decoding side (client) does not work... despite doing exactly the same calls, in exactly the same order as the example daala client. This can be verified by applying the daala-example-player.diff patch to the player and running the daala xpra test with the ogg stream we have recorded:

;XPRA_DAALA_DEBUG=1 LD_LIBRARY_PATH=/usr/local/lib64  PYTHONPATH=. ./tests/xpra/codecs/test_daala.py xterm-daala.ogg

(for the record, everything also builds fine on OSX after installing ​check.sf.net)

[Antoine Martin]

output from the daala xpra test

[Antoine Martin]

output from the daala example player

[Antoine Martin]

I've asked on the mailing list and got no reply, although the ​daala source tree does see some updates, if the mailing list is any sign... this project does not care and neither should we.

[migration script]

this ticket has been moved to: ​https://github.com/Xpra-org/xpra/issues/454