[FFmpeg-devel] [PATCH]lavf/mov: Support alac extradata in wave atom
muken.the.vfrmaniac at gmail.com
Mon Aug 3 15:30:14 CEST 2015
2015-08-03 21:53 GMT+09:00 wm4 <nfxjfg at googlemail.com>:
> On Mon, 3 Aug 2015 10:32:02 +0000 (UTC)
> Carl Eugen Hoyos <cehoyos at ag.or.at> wrote:
> > Paul B Mahol <onemda <at> gmail.com> writes:
> > > > Attached patch fixes ticket #4747 for me,
> > > > I don't know how to detect that the wave
> > > > atom contains no frma / alac atom...
> > >
> > > And that is mentioned because?
> > The file from ticket #4747 does not contain an
> > alac atom which is required for alac decoding.
> > Instead it contains the necessary data in the
> > wave atom (where you normally would find the
> > alac atom). To distinguish, it would be
> > necessary to look "into" the wave atom to know
> > if it contains an alac atom or not.
> > I meanwhile found that the mpeg demuxer
> > sometimes seeks eight bytes back which I will
> > probably copy. The alternative, iiuc, would be
> > to copy the whole wave atom into a buffer and
> > call mov_read_default() from the wave_read()
> > function with that buffer.
> > But this should not be necessary if somebody
> > can confirm that it is ok to seek back x bytes
> > in a demuxer.
> > As an alternative the current version could
> > also be committed, I am not sure.
> What produced that file?
According to an udta atom, it is produced by Sound Grinder 3.4.6.
Does QuickTime play it? Why is there even a
> wave header?
wave atom != wave header
In QuickTime file formats, codec specific info of compressed audio are
stored in wave atom as container atom, and inserted between frma atom and
I doubt that QuickTime can play it.
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
More information about the ffmpeg-devel