Age | Commit message (Collapse) | Author |
|
This should clean up clangtidy warnings
Change-Id: Ifb5a986121b2d0bd71b9ad39a79dd46c63bdb998
|
|
this moves the framework to c++11 and changes *_TEST_CASE* to
_TEST_SUITE
BUG=webm:1695
Change-Id: I07f2c20850312a9c7e381b38353d2f9f45889cb1
|
|
Add control for delta q for uv. 0 by default.
Change-Id: Ib8ed160b1a5c9a61ba15985076f6c6f121477103
|
|
Test was running at speed 4, which is not used for real-time.
With this change all Datarate tests are now running at
(speed >= 5, 1 pass, real-time mode), which is what they
were intended for.
BUG=webm:1512
Change-Id: I47a721dadd24b73df722c44419df7cfc06c44226
|
|
this test is only useful for realtime mode testing given the number of
frames and that one-pass vod has never been a primary focus for
development.
BUG=webm:1512
Change-Id: I23208393a5fcc5bcf9b267fab4b0d1aad500918a
|
|
Condition the pre-encode buffer update based on
TS diff on temporal layers = 1 for now, as some
fix is needed for the case where #temporal_layers > 1.
Change-Id: I58163b956db415217e4687a31f8ba110545b09f5
|
|
Encode the next frame at max q.
For layers: post_encode_drop is only check on base
spatial layer, and if base is post-encoded-dropped,
then whole superframe is dropped.
Added API to guard postencode dropping. Turned off by default.
Added unittest.
BUG=b/112990050
Change-Id: I42fee279014aca616f7a4d9b582cb2bf5da2f2e7
|
|
|
|
BUG=b/112294545
Change-Id: I5be230c8969d69af3ad87068fdf3834ef1af11d9
|
|
Low speeds in good mode are too slow.
Move CBR large tests to non-'Large' ones such that they can run in
Jenkins per commit.
Change-Id: I1da73ca96ee89abcf3566d51ff52f1f2e904a048
|
|
Change-Id: Ifbd753183782e680a9ae77c55d75f4d9b3fb2477
|
|
Due to the change of clip.
Change-Id: Ibbe1a865df0837c349770287d148081230d10aaa
|
|
replace hantro_collage with niklas vga clip.
Change-Id: I79b89ce12823095a5ee75025b2ddce9e8ef1452a
|
|
aq mode 3 is never properly tested in non realtime mode.
BUG=webm:1553
Change-Id: I0663c9724ee57ba5c528a20b31ef8b6df0e03f6c
|
|
It caused failure on vp9 datarate tests for temporal layers.
Change-Id: Id6e260efa33b3b08070391a91a013efef2706fb5
|
|
Change-Id: I4e9c73d6d1d9ea560f04cc37aaf99d58ec2ab551
|
|
Change-Id: I4c4d31d013cb45e20918f4ef83ce32811d76e02b
|
|
Disable 8x8 blocks for higher resolutions,
reduce mv_thresh for 1/2 subpel motion, and
disable golden reference at superblock level
based on source sad and motion content.
~6% loss in RTC metrics over current speed 9.
Speedup about ~10% for high motion clip on linux.
Change-Id: I7ff8f81ac93ee8a90d5a1f4837c955d000bd75e7
|
|
|
|
Enable alt_ref and compound prediction at speed 5.
For 1 pass VBR mode, when lag > 0.
Gain for Live set: ~3% gain on average, several
clips have gains ~5-15%.
Encoder fps decrease ~5-10%, on desktop with 4 threads.
For now enable it only for resolutions <= 1280x720.
Change-Id: I25e3d61a2244a3a01962624052c5adf4837965c7
|
|
Set subpel search stop to 2 when motion vector is non zero.
10% speedup on 1 and 2 threads on Samsung Galaxy S8+.
Change-Id: I7323bb913000229cf60a37495bf88bcc51d0ac96
|
|
Change-Id: I1458bf25fadc23e8be5a9532a153d7129a53accf
|
|
Change-Id: Icfaf29e1ca847ba9e3748700c9e09383ce8d1f65
|
|
The control is set by log2 of number of threads (such that the number of
tiles is the same of number of threads).
Thus it should be log2(num_threads) instead of (num_threads >> 1).
Change-Id: I2ccec5557e660048dad3e561534e1c74fc8eec1f
|
|
in BasicRateTargetingVBRLagZero and
BasicRateTargetingVBRLagNonZeroFrameParDecOff after:
e0b28ad69 Add extra case to wq_err_divisor()
BUG=webm:1512
Change-Id: Id181613cc191ff2a2281deffe141efb982501edf
|
|
Add/Remove static to functions. Name change.
Change-Id: I5de3efc23cd151fe8e70fe67a7a11acfcfa707dc
|
|
As we add more tests to datarate_test.cc, it's growing bigger and hard
to find specific test.
Split it to vp8, vp9 and svc ones.
Change-Id: Ie8c302010cf304a95554bee19d87ddc90498d0fb
|