[FFmpeg-devel] GSoC 2009: FFmpeg is in

[Luca Abeni]

Ronald S. Bultje wrote:
[...]
>> After this first step, I believe that there are some bugs to fix (RTCP
>> RR) and features to improve (better recovery after lost packets, handling
>> of packets arriving in the wrong order, etc...)
> 
> I like all of these. As Luca Barbato said, we should add a list of all
> of these "specific tasks" to the wiki small tasks page.

I agree; I'll try to fill something in the wiki soon.


>  If the SoC
> doesn't choose our RT(S)P-project, I'll go through them one-by-one,
> and hopefully so will you. :-).

Of course I'll try to work on this myself... The problem is to find some 
time ;-) I keep saying that in few weeks I should have more spare time, 
and every week my todo list get longer. Anyway something changed and I 
hope I'll really have more time to work on these things.


> Maybe one of my biggest issues is that I don't really know what
> feature requests are unimplemented. Hell, I don't even know half of
> the payloads in gst/live555/nemesi, but we should try to port them
> anyway. Keeping notes, having sample streams for each of them etc.

Yes, test streams are a problem here... Hence I think we should start 
from the most popular streams. Such as youtube mobile (support for it is 
periodically asked)... H.263 support is almost there, but I do not know 
how difficult AMR support can be (never looked at it).

> would be a great help. Since we have ffserver and feng and are
> hopefully familiar with their codebase (a bit?), we can try to setup a
> testcase/-environment for each of these bugs/features, e.g. one where
> packets arrive in the wrong order and create a task to "fix playback
> of this stream".

This would be a good idea. Some time ago, I tried something similar for 
writing rtpenc regression tests (but I never had them working and this 
is just another item on my todo list ;-)


				Luca