These tests have been causing a large number of false alerts, which is
tons of work for the perf sheriff. It's infeasible to test every single
permutation, so we must choose carefully. This CL reduces the number
of RESULT lines from the test from ~450 to ~50. I attempted to choose
interesting permutations, but you probably know better what's
interesting...
https://chromeperf.appspot.com/report?
sid=a7193c96f708018848ca07ad9c78ac657cadab3c70b3939c42bd7d70a092d61a
also suggests most of the metrics have enormous standard deviations,
so maybe you could look into how stable the metrics really are
and remove/stabilize the ones that aren't?
BUG=chromium:666725
Review-Url: https://codereview.webrtc.org/2529393006
Cr-Commit-Position: refs/heads/master@{#15329}
And delete the method CongestionController::packet_router.
BUG=None
Review-Url: https://codereview.webrtc.org/2516983004
Cr-Commit-Position: refs/heads/master@{#15323}
RTCIceCandidatePairStats.transport_id is set to the related
RTCTransportStats' id.
Unittest for RTCIceCandidatePairStats is updated to do EXPECT_EQ
between actual and an expected hardcoded dictionary. The previous way of
testing, ExpectReportContainsCandidatePair, is removed.
(ExpectReportContainsCandidate still exist, we might want to replace
this by EXPECT_EQ testing in a follow up.)
Unittest for RTCTransportStats is similarly updated and
ExpectReportContainsTransportStats is removed. A bug was uncovered where
the "rtcp_connection_info.best_connection = true" case was not tested
(a copy of rtcp_connection_info was used in the test, modifying that had
no affect on the test) - fixed.
rtcstats_integrationtest.cc updated to take transport_id into account.
In order to reuse an updated version of expected_rt[c]p_transport in the
unittest, timestamps are ignored by RTCStats::operator==.
BUG=chromium:627816, chromium:653873, chromium:653873, webrtc:6755
Review-Url: https://codereview.webrtc.org/2527113002
Cr-Commit-Position: refs/heads/master@{#15316}
JitterBufferDelayInMs is used for the WebRTC-NewVideoJitterBuffer finch
experiment, and therefore needs to be calculated.
BUG=webrtc:5514
Review-Url: https://codereview.webrtc.org/2534093003
Cr-Commit-Position: refs/heads/master@{#15313}
WEBRTC_VOICE_ENGINE_AGC, WEBRTC_VOICE_ENGINE_ECHO, and
WEBRTC_VOICE_ENGINE_NR are now gone.
BUG=webrtc:6506
Review-Url: https://codereview.webrtc.org/2530373002
Cr-Commit-Position: refs/heads/master@{#15309}
To support rotation in DirectX capturer, several other changes are also
required.
1. Removing AddRect in RotateDesktopFrame, this is a performance improvement.
DxgiOutputDuplicator creates a rotated DesktopRegion, which can be directly
add to updated_region.
2. DxgiOutputDuplicator::SourceRect() is not accurate, the rectangle in source
is controlled by |offset| or |rotation_| + |offset|, instead of desktop_rect().
3. The |region| in DxgiTexture::CopyFrom() is not accurate. It needs an
unrotated DesktopRegion which offsets by |offset| instead of desktop_rect(). To
avoid generating both rotated and unrotated updated_region, this parameter has
been removed. This impacts DxgiTextureStagning performance a little bit (1.5ms).
Refer to bug for details.
BUG=webrtc:6646
Review-Url: https://codereview.webrtc.org/2530303002
Cr-Commit-Position: refs/heads/master@{#15308}
It's at the same level as the other end-to-end tests that are also
disabled. It seems to be failing because it's not fast enough to keep up
with processing fake video frames.
BUG=webrtc:4387
TBR=kjellander@webrtc.org
NOTRY=True
Review-Url: https://codereview.webrtc.org/2538783002
Cr-Commit-Position: refs/heads/master@{#15302}
This CL adds WebRTC.Audio.ApmCaptureInputLevelAverage and
WebRTC.Audio.ApmCaptureInputLevelPeak. The metrics are updated once
every 10 seconds.
BUG=webrtc:6622
Review-Url: https://codereview.webrtc.org/2534473004
Cr-Commit-Position: refs/heads/master@{#15300}
This Cl is the first step of a 3-step process to trick the build system of a
webrtc dependency into not failing.
The second step will be to register this folder in the dependency
build system. It cannot be done first, because there must be a GN
dependency from some existing target to a target in the new folder.
The third step is to land https://codereview.webrtc.org/2424173003, in
which the webrtc/audio/utility is used.
When a new folder with a publicly visible header is added, the build
files of said build system must be manually updated to register the
new folder.
BUG=webrtc:6548
Review-Url: https://codereview.webrtc.org/2532403003
Cr-Commit-Position: refs/heads/master@{#15299}
This change rewrites RMSLevel, making it accept an ArrayView as input,
and modify the implementation somewhat. It also makes the class keep
track of the peak RMS in addition to the average RMS over the
measurement period.
New tests are added to cover the new functionality.
BUG=webrtc:6622
Review-Url: https://codereview.webrtc.org/2535523002
Cr-Commit-Position: refs/heads/master@{#15294}
Remove entries for headers that no longer exists in the webrtc/ dir.
BUG=webrtc:5878
NOTRY=True
Review-Url: https://codereview.webrtc.org/2540633002
Cr-Commit-Position: refs/heads/master@{#15292}
Issue: This API was calculating the file_header and frame_header offset only for the first frame which is not the right logic. We need to skip the file and frame header every time we extract a new frame.
Also added a unit test case which compares the extracted frame with the frame stored in text file.
NOPRESUBMIT=true
NOTRY=true
BUG=webrtc:6761
Review-Url: https://codereview.webrtc.org/2532963002
Cr-Commit-Position: refs/heads/master@{#15288}
This is an integration test using peerconnectiontestwrapper.h to set up
and end to end test using a real PeerConnection implementation. These
tests will complement rtcstatscollector_unittest.cc which collects all
stats using mocks.
The integration test is set up so that all stats types are returned by
GetStats and verifies that expected dictionary members are defined. The
test could in the future be updated to include sanity checks for the
values of members. There is a sanity check that references to other
stats dictionaries yield existing stats of the appropriate type, but
other than that members are only tested for if they are defined not.
StatsCallback of rtcstatscollector_unittest.cc is moved so that it can
be reused and renamed to RTCStatsObtainer.
TODO: Audio stream track stats members are missing in the test. Find out
if this is because of a real problem or because of testing without real
devices. Do this before closing crbug.com/627816.
BUG=chromium:627816
Review-Url: https://codereview.webrtc.org/2521663002
Cr-Commit-Position: refs/heads/master@{#15287}
This brings QualityScaler much more in line with OveruseFrameDetector.
The two classes are conceptually similar, and should be used in the
same way. The biggest changes in this CL are:
- Quality scaling is now only done in ViEEncoder and not in each
encoder implementation separately.
- QualityScaler now checks the average QP asynchronously, instead of
having to be polled on each frame.
- QualityScaler is no longer responsible for actually scaling the frames,
but has a callback to ViEEncoder that it uses to express it's desire
for lower resolution.
BUG=webrtc:6495
Review-Url: https://codereview.webrtc.org/2398963003
Cr-Commit-Position: refs/heads/master@{#15286}
Intervals when video is paused is no longer included in the stats:
"WebRTC.Video.InputFramesPerSecond"
"WebRTC.Video.SentFramesPerSecond"
BUG=webrtc:5283
Review-Url: https://codereview.webrtc.org/2536743002
Cr-Commit-Position: refs/heads/master@{#15285}
The actual time used in production code should honor the epoch time.
BUG=webrtc:6737
Review-Url: https://codereview.webrtc.org/2526433002
Cr-Commit-Position: refs/heads/master@{#15282}
