Reason for revert:
Perf test broke as it made assumptions that quality scaling was turned off. This turns out not to be the case. Fixed by turning quality scaling off for the tests.
Original issue's description:
> Revert of Drop frames until specified bitrate is achieved. (patchset #12 id:240001 of https://codereview.webrtc.org/2630333002/ )
>
> Reason for revert:
> due to failures on perf tests (not on perf stats, but fails running due to dcheck failures), see e.g., https://build.chromium.org/p/client.webrtc.perf/builders/Android32%20Tests%20(K%20Nexus5)
>
> Original issue's description:
> > Drop frames until specified bitrate is achieved.
> >
> > This CL fixes a regression introduced with the new quality scaler
> > where the video would no longer start in a scaled mode. This CL adds
> > code that compares incoming captured frames to the target bitrate,
> > and if they are found to be too large, they are dropped and sinkWants
> > set to a lower resolution. The number of dropped frames should be low
> > (0-4 in most cases) and should not introduce a noticeable delay, or
> > at least should be preferrable to having the first 2-4 seconds of video
> > have very low quality.
> >
> > BUG=webrtc:6953
> >
> > Review-Url: https://codereview.webrtc.org/2630333002
> > Cr-Commit-Position: refs/heads/master@{#16391}
> > Committed: 83399caec5
>
> TBR=perkj@webrtc.org,sprang@webrtc.org,stefan@webrtc.org,kthelgason@webrtc.org
> # Skipping CQ checks because original CL landed less than 1 days ago.
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=webrtc:6953
>
> Review-Url: https://codereview.webrtc.org/2666303002
> Cr-Commit-Position: refs/heads/master@{#16395}
> Committed: 35fc2aa82fTBR=perkj@webrtc.org,sprang@webrtc.org,stefan@webrtc.org,minyue@webrtc.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=webrtc:6953
Review-Url: https://codereview.webrtc.org/2675223002
Cr-Commit-Position: refs/heads/master@{#16473}
Refactor how |codec_id| is set, remove outdated TODO, update comments
with new bugs IDs.
BUG=webrtc:7061
Review-Url: https://codereview.webrtc.org/2670343002
Cr-Commit-Position: refs/heads/master@{#16467}
This tag is supposed to be temporary and removed when all Chromium tests
have been migrated to JUnit4.
BUG=webrtc:7123,chromium:640116
NOTRY=True
Review-Url: https://codereview.webrtc.org/2683583002
Cr-Commit-Position: refs/heads/master@{#16465}
This was causing some apps that include WebRTC to be rejected from the
app store.
BUG=webrtc:6382
Review-Url: https://codereview.webrtc.org/2679913002
Cr-Commit-Position: refs/heads/master@{#16462}
That would simplify their usage in tests where perfomance is not critical.
BUG=None
Review-Url: https://codereview.webrtc.org/2675713005
Cr-Commit-Position: refs/heads/master@{#16461}
We can then drop the CongestionController and RemoteBitrateEstimator
completely from the receive streams.
BUG=webrtc:6847
Review-Url: https://codereview.webrtc.org/2669463006
Cr-Commit-Position: refs/heads/master@{#16459}
Reason for revert:
Will try to reland FlexFEC tests, since these do not seem to be flaky on the buildbots.
Original issue's description:
> Revert of Improve and re-enable FEC end-to-end tests. (patchset #3 id:40001 of https://codereview.webrtc.org/2675573004/ )
>
> Reason for revert:
> Ulpfec tests are still flaky on buildbots.
>
> Original issue's description:
> > Improve and re-enable FEC end-to-end tests.
> >
> > These tests got flaky under the new jitter buffer.
> >
> > Enhancements:
> > - Use send-side BWE.
> > - Let BWE ramp up before applying packet loss.
> > - Improve packet loss simulation for ULPFEC.
> > - Add delay to fake network pipe for FlexFEC.
> > (Not added for ULPFEC, since this makes those flaky...?)
> > - Add FlexFEC+NACK test, using RTX instead of "raw retransmits".
> > - Tighter checks of received packets' payload types and SSRCs.
> >
> > TESTED=
> > $ ninja -C out/Debug video_engine_tests && third_party/gtest-parallel/gtest-parallel -r 1000 out/Debug/video_engine_tests --gtest_filter="*EndToEnd*Ulpfec*:*EndToEnd*Flexfec*"
> > ninja: Entering directory `out/Debug'
> > ninja: no work to do.
> > [12000/12000] TestWithNewVideoJitterBuffer/EndToEndTest.RecoversWithFlexfecAndNack/1 (14935 ms)
> >
> > BUG=webrtc:7047
> >
> > Review-Url: https://codereview.webrtc.org/2675573004
> > Cr-Commit-Position: refs/heads/master@{#16449}
> > Committed: d40b0f39e0
>
> TBR=stefan@webrtc.org,nisse@webrtc.org
> # Skipping CQ checks because original CL landed less than 1 days ago.
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=webrtc:7047
>
> Review-Url: https://codereview.webrtc.org/2672373002
> Cr-Commit-Position: refs/heads/master@{#16450}
> Committed: fd8d2654d7TBR=stefan@webrtc.org,nisse@webrtc.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=webrtc:7047
Review-Url: https://codereview.webrtc.org/2675283003
Cr-Commit-Position: refs/heads/master@{#16452}
Also fixes a bug where RTCP transport feedback was sent even though RTCP was disabled.
May affect perf numbers since the behavior of the send-side BWE differs a lot from the recv-side BWE.
BUG=webrtc:7111
Review-Url: https://codereview.webrtc.org/2669413003
Cr-Commit-Position: refs/heads/master@{#16451}
Reason for revert:
Ulpfec tests are still flaky on buildbots.
Original issue's description:
> Improve and re-enable FEC end-to-end tests.
>
> These tests got flaky under the new jitter buffer.
>
> Enhancements:
> - Use send-side BWE.
> - Let BWE ramp up before applying packet loss.
> - Improve packet loss simulation for ULPFEC.
> - Add delay to fake network pipe for FlexFEC.
> (Not added for ULPFEC, since this makes those flaky...?)
> - Add FlexFEC+NACK test, using RTX instead of "raw retransmits".
> - Tighter checks of received packets' payload types and SSRCs.
>
> TESTED=
> $ ninja -C out/Debug video_engine_tests && third_party/gtest-parallel/gtest-parallel -r 1000 out/Debug/video_engine_tests --gtest_filter="*EndToEnd*Ulpfec*:*EndToEnd*Flexfec*"
> ninja: Entering directory `out/Debug'
> ninja: no work to do.
> [12000/12000] TestWithNewVideoJitterBuffer/EndToEndTest.RecoversWithFlexfecAndNack/1 (14935 ms)
>
> BUG=webrtc:7047
>
> Review-Url: https://codereview.webrtc.org/2675573004
> Cr-Commit-Position: refs/heads/master@{#16449}
> Committed: d40b0f39e0TBR=stefan@webrtc.org,nisse@webrtc.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=webrtc:7047
Review-Url: https://codereview.webrtc.org/2672373002
Cr-Commit-Position: refs/heads/master@{#16450}
These tests got flaky under the new jitter buffer.
Enhancements:
- Use send-side BWE.
- Let BWE ramp up before applying packet loss.
- Improve packet loss simulation for ULPFEC.
- Add delay to fake network pipe for FlexFEC.
(Not added for ULPFEC, since this makes those flaky...?)
- Add FlexFEC+NACK test, using RTX instead of "raw retransmits".
- Tighter checks of received packets' payload types and SSRCs.
TESTED=
$ ninja -C out/Debug video_engine_tests && third_party/gtest-parallel/gtest-parallel -r 1000 out/Debug/video_engine_tests --gtest_filter="*EndToEnd*Ulpfec*:*EndToEnd*Flexfec*"
ninja: Entering directory `out/Debug'
ninja: no work to do.
[12000/12000] TestWithNewVideoJitterBuffer/EndToEndTest.RecoversWithFlexfecAndNack/1 (14935 ms)
BUG=webrtc:7047
Review-Url: https://codereview.webrtc.org/2675573004
Cr-Commit-Position: refs/heads/master@{#16449}
Intervals when video is paused is no longer included in the stats:
"WebRTC.Video.BitrateSentInKbps"
"WebRTC.Video.MediaBitrateSentInKbps"
"WebRTC.Video.PaddingBitrateSentInKbps"
"WebRTC.Video.RetransmittedBitrateSentInKbps"
"WebRTC.Video.RtxBitrateSentInKbps"
"WebRTC.Video.FecBitrateSentInKbps"
BUG=webrtc:5283
Review-Url: https://codereview.webrtc.org/2536613002
Cr-Commit-Position: refs/heads/master@{#16447}
Improves possibility to debug the demo application using adb.
As an example, now allows 'adb run as' which is useful when profiling CPU. It enables us to do profiling on non-rooted devices (excluding details).
BUG=NONE
Review-Url: https://codereview.webrtc.org/2676983003
Cr-Commit-Position: refs/heads/master@{#16446}
AEC3 and AEC2 are separate submodules in APM. This CL ensures that AEC3
deactivates AEC2 if both are active at the same time.
BUG=webrtc:6018
Review-Url: https://codereview.webrtc.org/2675863004
Cr-Commit-Position: refs/heads/master@{#16443}
These structs will be used for ORTC objects (and their WebRTC
equivalents).
This CL also introduces some minor changes to the existing implemented
structs:
- max_bitrate_bps uses rtc::Optional instead of "-1 means unset"
- "mime_type" turned into "name"/"kind" (which can be used to form the
MIME type string, if needed).
- clock_rate and channels changed to rtc::Optional, since they will
need to be for RtpSender.send().
- Renamed "channels" to "num_channels" (the ORTC name, which I prefer).
BUG=webrtc:7013, webrtc:7112
Review-Url: https://codereview.webrtc.org/2651883010
Cr-Commit-Position: refs/heads/master@{#16437}
Update libwertc AudioRtpSender::SetAudioSend with WEBRTC_WEBKIT_BUILD
This only introduces the WEBRTC_WEBKIT BUILD, inspired by WEBRTC_CHROMIUM_BUILD
macro. It is only defined by Webkit libwebrtc build system.
https://trac.webkit.org/changeset/210977
BUG=webrtc:7039
Review-Url: https://codereview.webrtc.org/2651273003
Cr-Commit-Position: refs/heads/master@{#16432}
Fix: Order of assignments is now correct, after being incorrect
due to an incorrect merge between
https://codereview.webrtc.org/2617373002/ and
https://codereview.webrtc.org/2589713003.
Improvement: Set parameters in more places, allowing for
correct reconfiguration. Add TODOs to point of minor issues
with current configuration.
TESTED=By locally patching an application using this code.
BUG=webrtc:5654
Review-Url: https://codereview.webrtc.org/2660403004
Cr-Commit-Position: refs/heads/master@{#16431}
The timer is only used for logging purposes and we can save CPU by
checking the volume less often.
BUG=webrtc:7096
Review-Url: https://codereview.webrtc.org/2669323003
Cr-Commit-Position: refs/heads/master@{#16430}
We have seen that PostTask can consume some CPU and the way we used it
before (logging only) in the ADB is not worth the cost we see when
profiling.
This CL simply moves frequent (trivial) stat updates from the task queue
to the native threads to avoid calling PostTask in each callback.
The reason for doing so before was to avoid locks but we can live without
them since races are benign here.
BUG=webrtc:7096
Review-Url: https://codereview.webrtc.org/2663383004
Cr-Commit-Position: refs/heads/master@{#16429}
This CL replaces the previously-hardcoded DTLS timeout value of 50 ms with one that is picked dynamically as 2x ICE RTT (clamped between 50 and 3000 ms to keep it reasonable).
No tests yet - I'm working on these, but sending out this CL early to get feedback. I've tested it manually and on slow networks it results in successful DTLS setup in a single attempt - instead of clogging up the slow link with multiple retransmissions as done previosly.
BUG=webrtc:7099
Review-Url: https://codereview.webrtc.org/2670113002
Cr-Commit-Position: refs/heads/master@{#16426}
As it is, the test fails to compile on some downstream compilers with the following error:
webrtc/modules/audio_coding/neteq/neteq_impl_unittest.cc:316:25: error: lambda capture 'kPayloadLength' is not required to be captured for this use [-Werror,-Wunused-lambda-capture]
.WillOnce(Invoke([kPayloadLength, kFirstSequenceNumber, kFirstTimestamp,
^
webrtc/modules/audio_coding/neteq/neteq_impl_unittest.cc:316:41: error: lambda capture 'kFirstSequenceNumber' is not required to be captured for this use [-Werror,-Wunused-lambda-capture]
.WillOnce(Invoke([kPayloadLength, kFirstSequenceNumber, kFirstTimestamp,
^
webrtc/modules/audio_coding/neteq/neteq_impl_unittest.cc:316:63: error: lambda capture 'kFirstTimestamp' is not required to be captured for this use [-Werror,-Wunused-lambda-capture]
.WillOnce(Invoke([kPayloadLength, kFirstSequenceNumber, kFirstTimestamp,
^
webrtc/modules/audio_coding/neteq/neteq_impl_unittest.cc:317:25: error: lambda capture 'kFirstReceiveTime' is not required to be captured for this use [-Werror,-Wunused-lambda-capture]
kFirstReceiveTime](const SdpAudioFormat& format,
BUG=webrtc:7107
Review-Url: https://codereview.webrtc.org/2672823002
Cr-Commit-Position: refs/heads/master@{#16422}
If an application sets a non-null value in RTCConfiguration.iceCheckMinInterval, we do not sent STUN pings more often than that. This is useful for bandwidth constrained scenarios.
This CL also increases the maximum STUN ping timeout to 60 seconds up from its previous value of 5 (which meant that a ping response received 5 seconds later would not be counted), and allows the RTT estimate to go up to 60 seconds from its previous limit of 3. RTTs above 3 seconds are possible on mobile links. (webrtc:7109)
This CL was originally written by pthatcher@, I am just submitting it after a minor cleanup.
BUG=webrtc:7082, webrtc:7109
Review-Url: https://codereview.webrtc.org/2670053002
Cr-Commit-Position: refs/heads/master@{#16421}
Since we can't know when a H264 frame really starts we instead group all packets together by timestamp when a frame seems to be complete (only in the case of H264).
BUG=webrtc:5514
Review-Url: https://codereview.webrtc.org/2675693002
Cr-Commit-Position: refs/heads/master@{#16419}
