[FFmpeg-devel] [PATCH 0/2] avfilter/vf_vpp_qsv: apply 3D LUT from file

[Chen Yufei]

First time modifing FFmpeg code and sending patch, please bear with me for
mistakes.

Following patches add support for applying 3D LUT from file using oneVPL VPP.

Limitation:
The new feature is only available when `CONFIG_VAAPI` is enabled, because of
my limited test environemnt and little experience with video processing.

1st patch moves some 3D LUT parse function from `vf_lut3d.c` to `lut3d.c` to
expose them for use in `vf_vpp_qsv.c`. I try to move only code that deals with
3D LUT table.

2nd patch applies 3D LUT from file in `vf_vpp_qsv.c`. I followed oneVPL's
document and sample code, but there are still things that are not clear to me.
The LUT value type used in VPP is `mfxU16` instead of float, I choose to
multiply float value with 65535 to convert to `mfxU16` and the result seems correct.

Test is done on an Arch Linux system (kernel 6.5.3) with Intel Arc A380 on a
Thunderbolt 3 GPU dock. I compared transcoding output with `vf_lut3d` and don't
see noticeable difference with my eyes.

`make fate` passes without error.

Chen Yufei (2):
  avfilter/vf_lut3d: expose 3D LUT file parse function.
  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 | 242 +++++++++++++-
 5 files changed, 926 insertions(+), 596 deletions(-)
 create mode 100644 libavfilter/lut3d.c

-- 
2.42.0