These global arrays are shared amongst all AEC instances, and were at
serious risk of data races. A Chromium TSAN bot recently caught this.
Also move the function pointer selection for optimization to
create-time. (Ideally this would only be done once.)
BUG=chromium:404133,1503
R=bjornv@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/19069004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@6922 4adac7df-926f-26a2-2b94-8c16560cd09d
The modification only uses the unique part of the spectrum (as is done for the C and asm code). It passes
byte to byte conformance test, and the single function performance
(if not specified, the code is compiled by GCC 4.6) on different
platforms:
| run 100k times | cortex-a7 | cortex-a9 | cortex-a15 |
| use C as the base on each | (1.2Ghz) | (1.0Ghz) | (1.7Ghz) |
| CPU target | | | |
|----------------------------+-----------+-----------+------------|
| C | 100% | 100% | 100% |
| Neon asm | 18% | 14% | 19% |
| Neon inline asm | 31% | 25% | 27% |
| Neon intrinsic (GCC 4.6) | 33% | 27% | 42% |
| Neon intrinscis (GCC 4.8) | 17% | 14% | 19% |
| Neon intrinsics (LLVM 3.3) | 15% | 13% | 18% |
BUG=
R=andrew@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/13739004
Patch from Joe Yu <joe.yu@arm.com>.
git-svn-id: http://webrtc.googlecode.com/svn/trunk@6920 4adac7df-926f-26a2-2b94-8c16560cd09d
audio_processing did not compile when aec_untrusted_delay_for_testing=1 was set. The constant kDelayDiffOffsetSamples was declared only for Mac when WEBRTC_UNTRUSTED_DELAY was automatically turned on.
Moving the declaration outside the ifdef makes it build with the flag on for any platform.
BUG=3673
TESTED=locally and trybots
R=andrew@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/22049004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@6866 4adac7df-926f-26a2-2b94-8c16560cd09d
With GCC 4.9, the MIPS NDK toolchain has been changed to only support 16 spregs by default - the even-numbered ones. This has been changed to support the R6 MIPS architecture. While the old behaviour could be restored by adding "-modd-spreg", this would come with a performance hit because the kernel would emulate odd-numbered spregs and missing R2 instructions.
As a result of this change, the functions removed in this CL no longer compile as there are no longer enough spregs for the compiler to assign. So we are removing these functions and they will use the C implementation until the MIPS code is rewritten.
R=andrew@webrtc.org, ljubomir.papuga@gmail.com, pasko@chromium.org
Review URL: https://webrtc-codereview.appspot.com/16159005
Patch from Fabrice de Gans-Riberi <fdegans@chromium.org>.
git-svn-id: http://webrtc.googlecode.com/svn/trunk@6797 4adac7df-926f-26a2-2b94-8c16560cd09d
---
Fixes for re-enabling more MSVC level 4 warnings: webrtc/ edition
This contains fixes for the following sorts of issues:
* Possibly-uninitialized local variable
* Signedness mismatch
* Assignment inside conditional
This also contains a small number of other cleanups to nearby code. In
particular several warning-disables for MSVC are removed because they don't seem
to be necessary (either that warning is not enabled or the code does not trigger
it).
BUG=crbug.com/81439
TEST=none
R=henrika@webrtc.org, pkasting@chromium.org
Review URL: https://webrtc-codereview.appspot.com/18769004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@6667 4adac7df-926f-26a2-2b94-8c16560cd09d
Gives the user a possibility to dynamically change the history size. The main advantage is, for example, that you now can start with a wide delay range and over time decrease the search window to lower complexity.
Adds
- two new APIs.
- and updates unit tests.
- a history_size member variable to BinaryDelayEstimator.
- two help function re-allocating buffer memory.
One thing that makes this a little complicated is that you are allowed to have multiple delay estimators with the same reference, so changing the buffer sizes at one place will automatically give you a mismatch at other places.
BUG=3532, 3504
TESTED=trybots and manually
R=aluebs@webrtc.org, kwiberg@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/15879004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@6593 4adac7df-926f-26a2-2b94-8c16560cd09d
Was causing warnings in Chromium such as:
warning C4742: 'WebRtcAec_overDriveCurve' has different alignment in 'E:\src\buildbot\build\slave\fake_slave\build\src\third_party\webrtc\modules\audio_processing\aec\aec_core_sse2.c' and 'E:\src\buildbot\build\slave\fake_slave\build\src\third_party\webrtc\modules\audio_processing\aec\aec_core.c': 4 and 16
warning C4744: 'WebRtcAec_overDriveCurve' has different type in 'E:\src\buildbot\build\slave\fake_slave\build\src\third_party\webrtc\modules\audio_processing\aec\aec_core_sse2.c' and 'E:\src\buildbot\build\slave\fake_slave\build\src\third_party\webrtc\modules\audio_processing\aec\aec_core.c': 'array (260 bytes)' and '__declspec(align(16)) array (260 bytes)'
warning C4742: 'WebRtcAec_weightCurve' has different alignment in 'E:\src\buildbot\build\slave\fake_slave\build\src\third_party\webrtc\modules\audio_processing\aec\aec_core_sse2.c' and 'E:\src\buildbot\build\slave\fake_slave\build\src\third_party\webrtc\modules\audio_processing\aec\aec_core.c': 4 and 16
warning C4744: 'WebRtcAec_weightCurve' has different type in 'E:\src\buildbot\build\slave\fake_slave\build\src\third_party\webrtc\modules\audio_processing\aec\aec_core_sse2.c' and 'E:\src\buildbot\build\slave\fake_slave\build\src\third_party\webrtc\modules\audio_processing\aec\aec_core.c': 'array (260 bytes)' and '__declspec(align(16)) array (260 bytes)'
BUG=https://code.google.com/p/chromium/issues/detail?id=336620R=andrew@webrtc.org, cd@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/15869004
Patch from Sebastien Marchand <sebmarchand@chromium.org>.
git-svn-id: http://webrtc.googlecode.com/svn/trunk@6525 4adac7df-926f-26a2-2b94-8c16560cd09d
This should work as a foundation for all the work that is
left to do to make the parts of WebRTC that Chromium uses
to build with GN.
I implemented some the smaller modules myself in this CL.
The remaining work (TODO's in the .gn files) will be distributed
to various team members.
I'm adding myself to OWNERS files for BUILD.gn files in all the
directories where I'm adding a BUILD.gn file.
BUG=3441
TEST=
Successful compilation of WebRTC as standalone:
gn gen out/Default --args="build_with_chromium=false" && ninja -C out/Default
gn gen out/Default --args="build_with_chromium=false is_clang=true clang_use_chrome_plugins=false" && ninja -C out/Default
I built successfully from a Chromium checkout (with
https://codereview.chromium.org/321313006/ applied) using:
gn gen out/Default && ninja -C out/Default webrtc
R=brettw@chromium.org, niklas.enbom@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/13749004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@6523 4adac7df-926f-26a2-2b94-8c16560cd09d
- Disables ApmTest.EchoCancellationReportsCorrectDelays
This test relys completely on the structure of how reported system delays are handled in AEC. In addition it assumes a fix setup of delay logging buffers. This test should be refactored.
- Adds flag to turn off reported_delay in audioproc
Now it is feasible to turn on and off the use of reported system delays.
BUG=N/A
R=aluebs@webrtc.org, kwiberg@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/16749004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@6492 4adac7df-926f-26a2-2b94-8c16560cd09d
Several tests were disabled in r6325 and r6326. Also, see issue 3445. This CL fixes the remaining four of the audio_processing related ones. Affects the tests:
- SystemDelayTest.CorrectDelayAfterStableBufferBuildUp
- SystemDelayTest.CorrectDelayDuringDrift
- SystemDelayTest.ShouldRecoverAfterGlitch
- ApmTest.EchoCancellationReportsCorrectDelays
The tests assumes reported delays are used, which now is explicitly set.
BUG=3445
TESTED=trybots
R=aluebs@webrtc.org, kwiberg@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/19769004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@6489 4adac7df-926f-26a2-2b94-8c16560cd09d
