Skip to content

Commit

Permalink
model: update hevc 10bit QP list
Browse files Browse the repository at this point in the history
Use a valid QP range accepted by the iHD driver for HEVC 10bit.

Unfortunately, there is still a bug in ffmpeg and vpl that
incorrectly forces QP in the wrong range...

See intel/vpl-gpu-rt#297

Thus, until the issue is fixed, the ffmpeg-qsv encoder code needs
to be modified, manually, to remove the QP range constraints so that
the requested values can propogate to vpl and the driver to properly
generate a trendline model.

Signed-off-by: U. Artie Eoff <[email protected]>
  • Loading branch information
uartie committed Mar 22, 2024
1 parent d409b2f commit 5f746fc
Showing 1 changed file with 3 additions and 0 deletions.
3 changes: 3 additions & 0 deletions model/encode/10bit/hevc.py
Original file line number Diff line number Diff line change
Expand Up @@ -35,6 +35,9 @@ def initvars(self, _):
@slash.parametrize(*TrendModelMixin.filter_spec(spec))
def test(self, case):
vars(self).update(case = case)
vars(self).update(
modelqps = [12, 13, 16, 18, 23, 31, 40, 42, 45, 48, 49, 51, 54, 57, 60, 61, 63]
)
vars(self).update(spec[case].copy())

# Some features/formats are only supported by VDENC (lowpower = 1).
Expand Down

0 comments on commit 5f746fc

Please sign in to comment.