[FFmpeg-devel] [PATCH v2 0/1] avfilter/vf_vpp_qsv: apply 3D LUT from file

[Chen Yufei]

On Tue, Jan 23, 2024 at 10:00 AM Xiang, Haihao <haihao.xiang at intel.com> wrote:
>
> On Sa, 2024-01-20 at 23:14 +0800, Chen Yufei wrote:
> > This patch adds support for applying 3D LUT from file using oneVPL VPP.
> >
> > PATCH v1 uses VA-API to create LUT surface. Because oneVPL can't work with VA-
> > API on Windows,
> > this version now creates LUT in system memory (MFX_RESOURCE_SYSTEM_SURFACE)
> > and let oneVPL
> > copy LUT to video memory.
> >
> > Note: requires oneVPL-intel-gpu version >= 24.1.1 because this version
> > contains
> > a fix for creating LUT in video memory.
> > (For details, refer to
> > https://github.com/oneapi-src/oneVPL-intel-gpu/issues/307)
>
> Please bump a new runtime version, then you may check the runtime version for
> this feature.

What does "bump a new runtime version" mean? Could you please provide
more details?

I'm confused about the version number of libvpl and oneVPL-intel-gpu.

I looked at the implementation of `QSV_RUNTIME_VERSION_ATLEAST`, and
tried to print `mfxVersion` stored in `vpp->qsv`.
I can only get 2.10 which is the version of libvpl (2.10.1 actually)
on my system.

I've also looked at the build directory of oneVPL-intel-gpu on my
system, the compile commands defines following

    -DMEDIA_VERSION_STR="24.1.2" -DMFX_API_VERSION="2.10+"

Do you mean I should just use `QSV_RUNTIME_VERSION_ATLEAST(ver, 2,
10)` to check the runtime version?
Does this guarantee the underlying oneVPL-intel-gpu would be >= 24.1.1 then?

>
> Thanks
> Haihao
>
>
> >
> > Chen Yufei (1):
> >   avfilter/vf_vpp_qsv: apply 3D LUT from file.
> >
> >  libavfilter/Makefile     |   8 +-
> >  libavfilter/lut3d.c      | 669 +++++++++++++++++++++++++++++++++++++++
> >  libavfilter/lut3d.h      |  13 +
> >  libavfilter/vf_lut3d.c   | 590 +---------------------------------
> >  libavfilter/vf_vpp_qsv.c | 113 ++++++-
> >  5 files changed, 799 insertions(+), 594 deletions(-)
> >  create mode 100644 libavfilter/lut3d.c
> >
>

-- 
Best regards,
Chen Yufei