[FFmpeg-devel] [PATCH] avdevice/v4l2: add limited support for multiplanar API

[Anton Khirnov]

Quoting Ramiro Polla (2024-06-28 13:47:13)
> On Thu, Jun 27, 2024 at 6:06 PM Anton Khirnov <anton at khirnov.net> wrote:
> > Quoting Ramiro Polla (2024-06-27 16:13:24)
> > > Is anyone fundamentally opposed to this approach to implement limited
> > > support for multiplanar API? I figure it could still be useful even
> > > when full multiplanar API support is implemented, because multiple
> > > memory buffers per packet will likely mean memcpy()s, which can be
> > > avoided if the plane count is one.
> >
> > Ideally this should be turned into a lavfi source, then multiple planes
> > can be exported individually.
> 
> Good idea. Is there any kind of roadmap on how this conversion from
> input device to source filter should take place, or is this something
> that still lacks consensus?

AFAIK there is a wide consensus that this is something that should be
done in general, though of course that does not directly translate into
volunteers actually doing the conversion.

> I assume we want to keep the functionality from the command line (so
> that "-i /dev/video0" would still use the lavfi source internally
> somehow). Would the code be shared between libraries, or would it be
> moved to lavfi and a wrapper written in lavd instead?

V4L2 is trickier than most devices, since it can also produce compressed
video. So at least that part of the functionaity will have to stay as a
demuxer.

-- 
Anton Khirnov