Is now set to:
<= 320x240: 600kbps
<= 640x480: 1.7Mbps
<= 960x540: 2Mbps
> 960x540: 2.5Mbps
For QVGA and VGA, the qp was around 10 at the selected thresholds when running some tests. The change in qp declined above the selected bitrates.
BUG=
R=mflodman@webrtc.org, pbos@webrtc.org, stefan@webrtc.org
Review URL: https://codereview.webrtc.org/1297373003 .
Cr-Commit-Position: refs/heads/master@{#9878}
This CL contains major modifications of the audio output parts for WebRTC on iOS:
- general code cleanup
- improves thread handling (added thread checks, remove critical section, atomic ops etc.)
- reduces loopback latency of iPhone 6 from ~90ms to ~60ms ;-)
- improves selection of audio parameters on iOS
- reduces complexity by removing complex and redundant delay estimates
- now instead uses fixed delay estimates if for some reason the SW EAC must be used
- adds AudioFineBuffer to compensate for differences in native output buffer size and
the 10ms size used by WebRTC. Same class as is used today on Android and we have unit tests for
this class (the old code was buggy and we have several issue reports of crashes related to it)
Similar improvements will be done for the recording sid as well in a separate CL.
I will also add support for 48kHz in an upcoming CL since that will improve Opus performance.
BUG=webrtc:4796,webrtc:4817,webrtc:4954, webrtc:4212
TEST=AppRTC demo and iOS modules_unittests using --gtest_filter=AudioDevice*
R=pbos@webrtc.org, tkchin@webrtc.org
Review URL: https://codereview.webrtc.org/1254883002 .
Cr-Commit-Position: refs/heads/master@{#9875}
This generates incorrect -lX11 with use_x11=0 in our other build system,
which causes the standalone libjingle_media target to not build.
This patch should fix that. I could remove -lX11 completely, and
libjingle still links fine. So does Chrome if I do the corresponding
change there, so I think this change is safe to make.
BUG=None
Review URL: https://codereview.webrtc.org/1306243013
Cr-Commit-Position: refs/heads/master@{#9869}
Reason for revert:
It looks like this broke the FYI bots. I tried updating libjingle_nacl.gyp, but the IOS build still failed because in Chrome it's configured to use NSS. See https://codereview.chromium.org/1316863012/.
Original issue's description:
> purge nss files and dependencies
>
> BUG=webrtc:4497
>
> Committed: https://crrev.com/5647a2cf3db888195c928a1259d98f72f6ecbc15
> Cr-Commit-Position: refs/heads/master@{#9862}
TBR=tommi@webrtc.org,kjellander@webrtc.org,torbjorng@webrtc.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=webrtc:4497
Review URL: https://codereview.webrtc.org/1311843006
Cr-Commit-Position: refs/heads/master@{#9867}
Enumerating using android.hardware.camera2 is 10x faster than enumerating using android.hardware.camera, but they don't list exactly the same formats. android.hardware.camera2 support higher resolutions for some cameras, and also different framerates.
R=tommi@webrtc.org
Review URL: https://codereview.webrtc.org/1321893003 .
Cr-Commit-Position: refs/heads/master@{#9861}
The purpose with this CL is to remove some code bloat. A subtle change is that GL_TEXTURE_MIN_FILTER in MediaCodecVideoDecoder is changed from GL_NEAREST to GL_LINEAR. This may lead to slightly worse performance when the decoded video is rendered minified, but with better visual quality. After reading https://crbug.com/351458 and the fix https://codereview.chromium.org/713603002 I think this is the right choice.
BUG=webrtc:4742
R=hbos@webrtc.org, tommi@webrtc.org
Review URL: https://codereview.webrtc.org/1303373005 .
Cr-Commit-Position: refs/heads/master@{#9845}
of decoder factory class.
- Add new Peer connection factory method to initialize shared
EGL context.
This provides an option to use single peer connection factory
in the application and create peer connections from the same
factory and reinitialize shared EGL context for video
decoding HW acceleration.
R=wzh@webrtc.org
Review URL: https://codereview.webrtc.org/1304063011 .
Cr-Commit-Position: refs/heads/master@{#9838}
Enumerating camera capabilities in the deprecated android.hardware.Camera interface is really slow because of the need to open and release the camera. By making getSupportedFormats() an interface, we allow apps the opportunity to inject their own implementation, such as storing the supported formats offline in the device's internal storage. It will also be possible to add an implementation of getSupportedFormats() using the new android.hardware.Camera2 interface in a follow-up CL.
R=tommi@webrtc.org
Review URL: https://codereview.webrtc.org/1321903002 .
Cr-Commit-Position: refs/heads/master@{#9819}
This CL makes the Java render interface asynchronous by requiring every call to renderFrame() to be followed by an explicit renderFrameDone() call. In JNI, this is implemented with cricket::VideoFrame::Copy() before calling renderFrame(), and a corresponding call to delete in renderFrameDone(). This CL is primarily done to prepare for a new renderer implementation.
BUG=webrtc:4742, webrtc:4909
R=glaznev@webrtc.org
Review URL: https://codereview.webrtc.org/1313563002 .
Cr-Commit-Position: refs/heads/master@{#9814}
This is handled by Android itself and may result in GL errors
when trying to release shaders when Activity is destroyed.
R=wzh@webrtc.org
Review URL: https://codereview.webrtc.org/1322703004 .
Cr-Commit-Position: refs/heads/master@{#9811}
webrtc::VideoSource resolves the kMaxFrameRate constraint by capping the desired framerate to kMaxFrameRate. That framerate is then passed into cricket::VideoCapturer::GetBestCaptureFormat(). The default implementation will choose a format from the supported formats list. Instead, we should override this function in AndroidVideoCapturer to give VideoCapturerAndroid.java the opportunity to choose a suitable framerate range.
BUG=webrtc:4938
R=glaznev@webrtc.org
Review URL: https://codereview.webrtc.org/1308953004 .
Cr-Commit-Position: refs/heads/master@{#9805}
Why the replacements? Mainly two reasons:
1) RTCCertificate owns the identity and as long as things are referencing the identity there should be a scoped_refptr reference to the RTCCertificate. Handing out raw pointers is less memory safe.
2) With the latest RFC, an RTCCertificate should be sufficient for specifying a crypto cert and the code should be updated to use RTCCertificate instead of SSLIdentity directly.
This replace work is split up into multiple CLs. In this CL...
- WebRtcSessionDescriptionFactory is updated to use RTCCertificate over SSLIdentity.
- WebRtcSessionDescriptionFactory::SignalCertificateReady is connected to WebRtcSession::OnCertificateReady and WebRtcSession is updated to use RTCCertificate.
- The cricket::Transport and related classes are updated to use RTCCertificate. These are called from WebRtcSession::OnCertificateReady.
BUG=webrtc:4927
R=tommi@webrtc.org, torbjorng@webrtc.org
Review URL: https://codereview.webrtc.org/1312643004 .
Cr-Commit-Position: refs/heads/master@{#9794}
VideoFrameBuffer currently has two overloaded data() functions for pixel access, one for const and one for non-const. Unfortunately, it will default to the non-const version, even when 'const scoped_refptr<VideoFrameBuffer>&' is used. This is a problem, because many subclasses use RTC_NOTREACHED() in the non-const version.
This CL makes the non-const version of data() explicit with a different, longer function name MutableData().
R=tommi@webrtc.org
Review URL: https://codereview.webrtc.org/1304143003 .
Cr-Commit-Position: refs/heads/master@{#9787}
- Pull out part of WebRtcVoiceMediaChannel::SetRecvCodecs() into WebRtcVoiceMediaChannel::SetRecvCodecsInternal().
BUG=webrtc:4690
R=pbos@webrtc.org
Review URL: https://codereview.webrtc.org/1291343002 .
Cr-Commit-Position: refs/heads/master@{#9785}
The primary fix in this CL is to remove the dangling |thread_| pointer in AndroidVideoCapturerJni. That thread is not safe to use after Stop() has been called. Even after Stop() has been called, we must still be able to return late frames to Java in order to not leak them, so that path has been made thread safe instead. To make sure that we always return frames, the Java frame should be wrapped in a scoped_refptr as quickly as possible, so this CL moves the wrapping from AndroidVideoCapturer to AndroidVideoCapturerJni. This also removes the need for the interface function AndroidVideoCapturerDelegate::ReturnBuffer().
Some other minor changes are:
* Remove |valid_global_refs_| and all logic related to that. Now that rtc::Bind() captures method objects as scoped_refptr, the destructor of AndroidVideoCapturerJni will not be called before all frames are returned.
* Remove global ref |j_frame_observer_|. No need for this, we don’t call it and it is kept alive with standard Java memory management.
* Add helper function ShallowCenterCrop() for VideoFrameBuffers. This functionality already exists in the constructor of WrappedI420Buffer, but it’s more convenient to have it as a separate function.
BUG=webrtc:4742,webrtc:4909
R=glaznev@webrtc.org, tommi@webrtc.org
Review URL: https://codereview.webrtc.org/1307973002 .
Cr-Commit-Position: refs/heads/master@{#9784}
