[FFmpeg-devel] [PATCH] miscelleneous NUT fixes

[Rich Felker]

On Sun, Feb 03, 2008 at 12:55:39PM +0100, Michael Niedermayer wrote:
> On Sun, Feb 03, 2008 at 01:11:33AM -0500, Rich Felker wrote:
> > On Sun, Feb 03, 2008 at 12:34:46AM +0100, Michael Niedermayer wrote:
> > > On Sat, Feb 02, 2008 at 02:19:25PM -0500, Rich Felker wrote:
> > > > On Sat, Feb 02, 2008 at 06:41:20PM +0100, Michael Niedermayer wrote:
> > > > > > 1. "Bleh, libnut muxed this ;)\n" please, I fixed this bug over a year ago :)
> > > > > 
> > > > > rejected, we do not remove workarounds needed for demuxing/decoding
> > > > 
> > > > While I agree with this principle, I don't think it applies to
> > > > under-development formats. The snow decoder does not support every
> > > > random historical version of snow, and likewise the nut demuxer should
> > > 
> > > I would support them if it is easy to do and would not cause a slowdown.
> > 
> > I reviewed the code in question and it's actually a bug. Samplerate
> > num=2 den=3 means 2 samples every 3 seconds, but with the bogus code
> > in place, it will be interpreted as 3 samples per second. Apparently
> > lavf does not support non-integer values for samples-per-second so it
> > should drop an "unsupported format" error in any case rather than
> > misinterpreting the data.
> 
> Maybe we should have disallowed non integer samplerates in nut ...

No, that creates a fundamental dependency on a physical unit
(seconds). Also what if one has a stream sampled at 44100 Hz but wants
to retime the video from NTSC to PAL or film to NTSC?? The conversion
factor does not divide 44100 evenly so a noninteger samplerate is
needed.

> > I stand by Oded's patch to remove the incorrect interpretation and
> 
> We have files which need this interpretation, we do not have ones which
> break due to it. Feel free to add a big comment to the code saying its
> not spec compliant if you like.

We have invalid files which should be patched or deleted.

> > hope lavf will eventually support non-integer samplerates.
> 
> Right after we add mpeg4 scalability, multiple object and most important so
> called face and body animation objects support. That also would make all
> the profile counters happy.

Um, all that's required for noninteger samplerates is converting the
samplerate field to a rational type instead of an integer type. This
is hardly comparable to implementing major high-complexity features.

Rich