[FFmpeg-devel] [PATCH 3/5] audioconvert: implement av_bprint_channel_layout().

[Clément Bœsch]

On Wed, Jun 06, 2012 at 04:23:56PM +0200, Nicolas George wrote:
> L'octidi 18 prairial, an CCXX, Clément Bœsch a écrit :
> > OK I did some random tests but it didn't work as smooth as I expected...
> > 
> > First I tried to generate a random sample with:
> > 
> >     ./ffmpeg -f lavfi -i "aevalsrc=sin(330*PI*2*t)[a];
> >                           aevalsrc=sin(440*PI*2*t)[b];
> >                           [a][b]amerge" -t 5 out.wav
> > 
> > But it crashed. Actually, it also crashes on master/HEAD.
> 
> Fixed by:
> http://ffmpeg.org/pipermail/ffmpeg-devel/2012-June/125541.html
> 
> > So I created two samples with sth like:
> > 
> >     ./ffmpeg -f lavfi -i aevalsrc='sin(330*PI*2*t)' -t 5 ch0.wav
> >              -f lavfi -i aevalsrc='sin(440*PI*2*t)' -t 5 ch1.wav
> > 
> > And then this:
> > 
> >     ./ffmpeg -i ch0.wav -i ch1.wav -filter_complex
> >         '[0:0][1:0]amerge[out0];
> >          [0:0][1:0]amerge[out1]' -map '[out0]'
> >                                  -map '[out1]' -y out.ogg
> > 
> > This actually crashes after your patchset (master/HEAD + the 5 patches of
> > this thread), but works with master/HEAD.
> 
> It works for me right now. Complete console output down below.
> 
> > With the previous out.ogg (generated from master/HEAD), in your amerge
> > branch I tried this:
> > 
> >     ./ffmpeg -i out.ogg -filter_complex
> >         '[0:0][0:1][1:0][1:1]amerge=inputs=4' -f null -
> > 
> > But it crashed again...
> 
> I get:
> 
> Invalid file index 1 in filtergraph description [0:0][0:1][1:0][1:1]amerge=inputs=4.
> 
> And indeed, your command line has only one input: 1:0 does not make sense.
> 

Ah yes IIRC I wanted to try two times -i out.ogg or something like that.

> > The only use case I'm interested is: take a video, and merge all the
> > streams into one. This is assuming all the streams are mono, OR stereo
> > pairs.
> 
> I am a little surprised here. I thought you wanted to implement -map_channel
> with that.
> 

No I didn't plan this, -map_channel was just meant to be a generic way of
doing such operation, but the feature I was actually interested in was
this "simple" task.

[...]

Anyway, if the problems above are fixed I don't have much to say on the
patch/patchset.

Thank you,

-- 
Clément B.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 490 bytes
Desc: not available
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20120606/3e6c35b2/attachment.asc>