Skip to content

Check the 10-bit video roundtrip gain explicitly instead of loosening the bound #92

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
hhoppe wants to merge 1 commit into
base: main
Choose a base branch
from
+11 −6
Open

Check the 10-bit video roundtrip gain explicitly instead of loosening the bound #92

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
17 changes: 11 additions & 6 deletions mediapy_test.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -648,13 +648,18 @@ def test_video_read_write_10bit(self):
)
new_video = media.read_video(path, dtype=np.uint16, output_format='gray')
self.assertEqual(new_video.dtype, np.uint16)
# FFmpeg <= 7.1 converts 16-bit <> 10-bit by a plain shift of 6 in both
# directions, so the round trip is exact. FFmpeg 8.0 still encodes with
# the shift but decodes by 65535/1020, taking the legacy peak 255 << 2 as
# white, so the round trip gains 65535/65280 = 257/256 and input values
# above 65280 saturate.
video_float = video.astype(np.float64)
gain = np.vdot(video_float, new_video) / np.vdot(video_float, video_float)
self.assertAlmostEqual(gain, 1.0, delta=0.01)
value_1_of_10bit_encoded_in_16bits = 64
# This factor may need to be raised in some recent versions of ffmpeg due
# to scaling issues on 10-bit <> 16-bit conversions.
factor = 2.5
self._check_similar(
video, new_video, max_rms=value_1_of_10bit_encoded_in_16bits * factor
)
factor = 0.8
max_rms = value_1_of_10bit_encoded_in_16bits * factor
self._check_similar(video, new_video / gain, max_rms=max_rms)

def test_video_read_write_vp9(self):
video = media.moving_circle((256, 256), num_images=4, dtype=np.uint8)
Expand Down
Loading