[FFmpeg-cvslog] r17881 - in trunk/libavcodec: elbg.c elbg.h roqvideo.h roqvideoenc.c

[Mike Melanson]

Carl Eugen Hoyos wrote:
> M?ns Rullg?rd <mans <at> mansr.com> writes:
> 
>>> Use FLG pseudo-random number generator in RoQ and ELBG
>> This broke the FATE idroq test.
> 
> May I suggest to move the test into the regression suite, where it belongs, imo?

We have already discussed this. A few issues:

1) It takes an absurdly long time to encode all 60 of the synthetic 
frames using a VQ algorithm. I don't want FATE to take that kind of hit 
when 8 or so frames proves the algorithm well enough.

2) FFmpeg's RoQ demuxer isn't capable of demuxing a RoQ file as short 
as, say, 8 frames that is also silent. This is due to some bad decisions 
on my part which I have not rectified (cf. issue 553).

3) I eventually want to split all of the regression suite's 
functionality into individual FATE tests anyway, rather than work the 
other way around.

-- 
     -Mike Melanson