[FFmpeg-devel] [PATCH]Support muxing more qcelp samples in mov

[Carl Eugen Hoyos]

Paul B Mahol <onemda <at> gmail.com> writes:

> 
> On 6/9/13, Carl Eugen Hoyos <cehoyos <at> ag.or.at> wrote:
> > Paul B Mahol <onemda <at> gmail.com> writes:
> >
> >> I already said that adding subframe capability to decoder
> >> looks to be best way to solve issue(s).
> >
> > Could you explain how this could fix the muxer issue?
> > I honestly don't understand.
> 
> I don't know what patch actually really solves

I originally wrote that the patch fixes "muxing of qcelp 
in mov and playback with QuickTime" meaning that with 
the patch all qcelp files remuxed to mov play fine with 
QuickTime which does not work without the patch, if 
that was really unclear without additional explanation 
I am sorry but note that you told me to test this, so 
I assumed you knew (or at least expected) this to be 
broken.

> and why solution in patch is actually prefered.

As said, because I cannot implement another solution 
and I doubt anybody else will work on this corner-
case (to use an euphemism).

> Could you confirm that this patch does not increases
> packets size for non 4 rate packets.

Of course it increases the packet size (that is the 
single thing this patch does), but note that it will 
either not increase the overall file size (if input 
was mov) or only compared to files that simply do 
not play at all (for qclp input).

> Is patch actually tested with files with non-4 
> rates packet?

I know that I sent an untested patch last week (I 
actually tested it but it appears I found the single 
case where it worked when I tested) and I am sorry 
about it, but I typically test everything that I 
send.

(The patch has no effect on 4 rate, so one could 
argue I only tested it on non-4 rate files, but 
I also verified that 4 rate files still work with 
QuickTime.)

Carl Eugen