Upstream commit: https://webrtc.googlesource.com/src/+/231ece13b7338e18146c2a6ab3191e93a2801a0c
Delete code that thinks H265 should have same QP as H264.
The QualityScalingExperiment class is used for two things:
1. Parsing a field trial that is not used in Chrome anymore but kept
around for backwards compat with other dependencies.
2. Provide default QP thresholds for VP8, VP9 and H264 in cases where
the encoder implementation is not providing their own QP thresholds
(which otherwise overrides these defaults).
The H264 thresholds were used as a placeholder for H265 not having any
QP thresholds of its own. Now manual QP/PSNR measurements, backed by
codec theory, has proven that H265 needs its own set of QP thresholds
separate from H264, so we delete this code path.
Any H265 encoder will now be forced to set `scaling_settings`, which is
already the case for other codecs that were added after the QP
experiment was sunset, namely AV1. (If it fails to do so, QP scaling is
disabled which would be a bug, see for example chromium:399587133.)
In Chrome, H265 sets its own QP thresholds as of this CL landing:
https://chromium-review.googlesource.com/c/chromium/src/+/6308082
Bug: chromium:391907171
Change-Id: I8ef2cd50f80bdf6019966ab07770a59b6e383d35
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/379401
Commit-Queue: Henrik Boström <hbos@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#44024}
Differential Revision: https://phabricator.services.mozilla.com/D244035
108009c984