[FFmpeg-devel] [PATCH 1/1] [RFC] ffprobe: report DAR even if SAR is undefined
cus at passwd.hu
Sun Apr 15 17:42:01 EEST 2018
On Sun, 15 Apr 2018, Timo Teräs wrote:
> Calculate DAR with assumed SAR 1:1 when SAR is undefined. Same
> assumption is done in ffplay to create the play window. Usually
> DAR is more useful metadata than SAR when e.g. choosing which
> media of multiple versions to use to fit the display.
> Normally undefined SAR means 1:1. E.g. in mov/mp4 files there's
> 'pasp' atom to explicitly define SAR. If that is not specified,
> the video codec's SAR should be used. Should it be also undefined,
> the SAR should be assumed to be 1:1. It makes sense to not change
> SAR in the demux info, so ffmpeg can make copies with matching
> atom structure and codec extra data. So the simplest way to
> report DAR is assume SAR 1:1 when undefined.
> Signed-off-by: Timo Teräs <timo.teras at iki.fi>
> This applies on top of the previous DAR/SAR reporting fix:
> For more details behind this patch, see:
> This seemed to be the most simple and logical fix. Alternatively,
> we could add a new 'effective_display_aspect_ratio' or add a flag
> to enable this assumption if the existing functionality should be
> kept unchanged. I felt this would be the most sensible thing to do.
> I am happy this if there's a more desirable approach to the issue.
> Hopefully this and the DAR/SAR reporting fix can be applied before
> next release tag. Thanks for considering!
I don't think it's good idea to generally assume 1:1 display_aspect_ratio
for every undefined sample aspect ratio. It depends heavily on your actual
use case. If MOV/MP4 specifies that 1:1 SAR should be used, then maybe you
should fix av_guess_sample_aspect_ratio instead, and return 1:1 there if
the format context is MOV/MP4. You may add a demuxer (AVFMT) flag to
signal that such behaviour should be used, and
av_guess_sample_aspect_ratio can check for that demuxer flag.
More information about the ffmpeg-devel