[FFmpeg-devel] [PATCH] Add YUV-based floating-point pixel formats.

[Ronald S. Bultje]

Hi,

On Sun, Jul 29, 2018 at 8:07 AM, Paul B Mahol <onemda at gmail.com> wrote:

> On 9/2/17, Andrew Sun <adsun701 at gmail.com> wrote:
> > This commit adds support for 32-bpc YUV-based floating-point pixel
> formats.
> > ---
> >  doc/APIchanges      |  2 ++
> >  libavutil/pixdesc.c | 75
> > +++++++++++++++++++++++++++++++++++++++++++++++++++++
> >  libavutil/pixfmt.h  | 11 ++++++++
> >  libavutil/version.h |  2 +-
> >  4 files changed, 89 insertions(+), 1 deletion(-)
> >
>
> All float formats need proper definition in nut container.


The nut maintainer can do that him/herself.

I agree with Michael we need swscale support, we've always required that
for new (non-hw) pix_fmt additions.

Ronald