[FFmpeg-devel] [PATCH] Fix sample_aspect_ratio computation for stereo matroska content.

[wm4]

On Mon, 30 Nov 2015 18:31:56 +0000
Aaron Colwell <acolwell at google.com> wrote:

> Hi,
> Can I get a review for this please? It has been a week since I sent out
> this patch and I haven't gotten any real feedback about the changes. I've
> attached a new copy that has been rebased to HEAD. If I need to be doing
> something different, please let me know so I can make progress on this.
> 
> Thanks,
> Aaron
> 
> On Wed, Nov 25, 2015 at 4:40 PM Aaron Colwell <acolwell at google.com> wrote:
> 
> > On Tue, Nov 24, 2015 at 10:43 AM Kirill Gavrilov <gavr.mail at gmail.com>
> > wrote:
> >  
> >> On Mon, Nov 23, 2015 at 11:37 PM, Aaron Colwell <acolwell at google.com>
> >> wrote:
> >>  
> >> > matroskaenc.c applies divisors to the display width/height when  
> >> generating  
> >> > stereo content. This patch adds the corresponding multipliers to
> >> > matroskadec.c
> >> > so that the original sample aspect ratio can be recovered.  
> >>
> >>
> >> Just to link this patch with Matroska Specification notes:
> >> http://www.matroska.org/technical/specs/notes.html#3D
> >>  
> >> > The pixel count of the track (PixelWidth/PixelHeight) should be the raw
> >> > amount of pixels (for example 3840x1080 for full HD side by side)
> >> > and the DisplayWidth/Height in pixels should be the amount of pixels for
> >> > one plane (1920x1080 for that full HD stream).
> >> >  
> >>  
> >
> > Is this intended to be a call to action or are you just adding information
> > to this particular thread? I'd like to get the change reviewed so I can
> > determine whether I need to make any other changes.
> >
> > This change basically allows 'ffmpeg -i blah.mkv foo.mkv ; ffprobe
> > foo.mkv' to show that foo.mkv has the same SAR and DAR as blah.mkv when
> > blah.mkv contains stereo content. Without this patch the SAR & DAR get
> > smaller based on whichever dimension is halved by the stereo layout.
> >
> > Hope this helps,
> > Aaron
> >
> >  

Are you sure matroskadec is wrong and not matroskaenc?

I suspect the change to matroskadec would make the behavior
inconsistent with vf_stereo3d in libavfilter. (But I'm not sure.)