bilboed at gmail.com
Wed Dec 24 15:23:47 CET 2008
Managed to find some time to have a look at this, and I'm seeing
The provided example can be trimmed even more (by not adding the last
adder and source) and will cause a deadlock (yes !).
The problem I'm seeing is that the first 5s play fine, then the
following 5s are fine too (first source is mixed with second source),
but the it just hangs before switching to the 'second-source-only'.
Digging some more, the problem is that one of the two sources stops
too early, resulting with remaining data in one of the adder's
collectpad, and therefore blocking waiting for the other stream to
And that seems to be due to mpegaudioparse aborting too early,
thinking it has sent out enough data for the configured segment. Maybe
Sebastian has a better idea of what's going wrong there.
Since I can't reproduce the exact behaviour, maybe the best at this
point would be to open a bugzilla item along with full debug logs and
report of what behaviour is happening.
Regardless of segments and bugs, I wish you all a merry xmas and/or
On Tue, 2008-12-23 at 12:24 +0100, Thomas Vander Stichele wrote:
> Hi guys,
> I've been trying various approaches to figuring out what bug I'm running
> into when using effects with gnonlin, but I'm failing.
> I've boiled it down to a simple test case that you can run on south.mp3
> that should make it easy to reproduce the problem.
> The script tries to play the last 10 secs of that file, then starts a
> second copy of that file from the start 5 seconds after the first, doing
> an adder mix for 5 seconds. A third copy is supposed to start a few
> seconds before the end of the second.
> As you can tell, instead of playing the whole part where the second copy
> should be playing alone for 50 seconds, it completely drops the solo
> part and jumps to the next mix immediately.
> I have no idea what's causing this. When I don't add the effects, while
> it obviously doesn't mix correctly, the timing of each of the three
> copies is correct.
> Edward is telling me that Jokosher is cheating, and that their solution is not really the proper gnonlin way.
> Anyone know what the 'proper' gnonlin way is, or is it just not possible today to create a jukebox-type
> application with gnonlin ?
> gstreamer-devel mailing list
> gstreamer-devel at lists.sourceforge.net
More information about the gstreamer-devel