[FFmpeg-devel] Why is writing the colr atom not the default in the mov muxer?

Robert Krüger krueger at lesspain.de
Sun Mar 1 13:26:32 CET 2015


On Tue, Feb 24, 2015 at 11:55 AM, Michael Niedermayer <michaelni at gmx.at>
wrote:

> On Tue, Feb 24, 2015 at 08:00:07AM +0000, tim nicholson wrote:
> > On 21/02/15 01:34, Dave Rice wrote:
> > > Hi Robert, Kevin,
> > >
> > >> On Feb 20, 2015, at 9:56 AM, Robert Krüger <krueger at lesspain.de>
> wrote:
> > >>
> > >> Am Freitag, 20. Februar 2015 schrieb Kevin Wheatley :
> > >>
> > >>> On Fri, Feb 20, 2015 at 1:30 PM, Robert Krüger <krueger at lesspain.de
> > >>> <javascript:;>> wrote:
> > >>>> if I read the code correctly, the colr atom is only written in the
> mov
> > >>>> muxer if the flag write_colr is specified. Was that behaviour
> chosen to
> > >>>> have better backward compatibility or is there another reason not to
> > >>> write
> > >>>> this standard atom by default?
> > >>>
> > >>> I chose that way to preserve the older behaviour, as it can change
> how
> > >>> files will be interpreted.
> > >>>
> > >>> I assumed that but isn't the change then a change for the better
> (then
> > >> maybe requiring a version bump and an entry in the release notes)?
> After
> > >> all Apple muxers write it by default as well and not trusting the
> input
> > >> metadata seems to me like something that should be opt-out rather than
> > >> opt-in but that's just my 2c.
> > >
> > > That's also my two cents and I also wondered why users have to opt-in
> to a correctly written file. The QuickTime spec says that colr is required
> with some streams (such as raw uyvy422 and v210), see:
> https://developer.apple.com/library/mac/technotes/tn2162/_index.html#//apple_ref/doc/uid/DTS40013070-CH1-TNTAG9
> .
> > >
> > > I'd propose that colr be written by default. The interpretation may be
> different but, since a file with a colr atom is more self-descriptive, the
> interpretation is more likely to be correct.
> >
> > +1
>
> if someone posts a patch, ill apply it
>

Am I reading the committed patch incorrectly or is colr still not written
by default? I thought the argument against replacing the flag (Derek's
first patch) was withdrawn because there has not been a release that
contained the original flag.


More information about the ffmpeg-devel mailing list