The Codec class is a big class and objects of the Codec class are passed
around by value. That means that inlined operations would be duplicated
at many places, in particular inside STL.
By not inlining Codec methods, webrtc shrinks by 50 KB in
a Linux x64 clang build.
Total change: -54147 bytes
==========================
+2810 - Source: /home/bratell/src/chromium/src/third_party/libjingle/source/talk/media/base/codec.cc - (gained 2920, lost 110)
-1003 - Source: /home/bratell/src/chromium/src/third_party/libjingle/source/talk/media/base/codec.h - (gained 0, lost 1003)
-1129 - Source: /home/bratell/src/chromium/src/third_party/libjingle/source/talk/media/sctp/sctpdataengine.cc - (gained 1660, lost 2789)
-1190 - Source: /home/bratell/src/chromium/src/third_party/libjingle/source/talk/media/base/rtpdataengine.cc - (gained 1408, lost 2598)
-1747 - Source: /home/bratell/src/chromium/src/third_party/libjingle/source/talk/session/media/mediasession.cc - (gained 803, lost 2550)
-2141 - Source: /home/bratell/src/chromium/src/third_party/libjingle/source/talk/media/webrtc/webrtcvideoengine.cc - (gained 1679, lost 3820)
-2250 - Source: /home/bratell/src/chromium/src/third_party/libjingle/source/talk/app/webrtc/webrtcsdp.cc - (gained 1224, lost 3474)
-2927 - Source: /usr/include/c++/4.8/bits/stl_vector.h - (gained 0, lost 2927)
-3729 - Source: /home/bratell/src/chromium/src/third_party/libjingle/source/talk/media/webrtc/webrtcvideoengine2.cc - (gained 10925, lost 14654)
-6369 - Source: /usr/include/c++/4.8/bits/vector.tcc - (gained 0, lost 6369)
-10582 - Source: /usr/include/c++/4.8/bits/stl_heap.h - (gained 0, lost 10582)
-19324 - Source: /usr/include/c++/4.8/bits/stl_algo.h - (gained 743, lost 20067)
BUG=
R=juberti@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/40729005
Cr-Commit-Position: refs/heads/master@{#8436}
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8436 4adac7df-926f-26a2-2b94-8c16560cd09d
Fetching the current codec for sake of gathering stats, is frequently blocked since it's done by acquiring the same lock as is held while encoding frames. This can mean tens of milliseconds.
To improve this, I'm taking advantage of the fact that the codec information is set on the same thread as is used to query the information. This means that locking isn't needed for querying this information. I'm adding checks to make sure debug builds will crash if this isn't followed.
An alternative to this approach could be to add one more lock that is specifically used for the codec information variable. This would also decouple querying codec information from the encoder itself, but still requires a lock.
This patch depends on making ThreadChecker part of rtc_base_approved:
https://webrtc-codereview.appspot.com/40539004/
BUG=2822
R=mflodman@webrtc.org, pthatcher@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/37779004
Cr-Commit-Position: refs/heads/master@{#8435}
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8435 4adac7df-926f-26a2-2b94-8c16560cd09d
Instead of sending a black frame that's the size of the VideoFormat send
a black frame in the format we're already sending. This prevents
expensive encoder reconfiguration when the sending format is a different
resolution. This speeds up setting a null capturer (removing the
capturer) significantly as it doesn't entail an encoder reconfiguration.
R=mflodman@webrtc.org, pthatcher@webrtc.org
BUG=1788
Review URL: https://webrtc-codereview.appspot.com/39179004
Cr-Commit-Position: refs/heads/master@{#8405}
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8405 4adac7df-926f-26a2-2b94-8c16560cd09d
It’s possible to implement cricket::VideoFrame::CopyToBuffer using the virtual interface. This removes the need for subclasses to implement their own versions. This CL also fixes a bug in cricket::VideoFrame::CopyToPlanes which currently assumes that GetUPitch() == GetVPitch(), otherwise it may segfault.
I think this CL should land regardless, but the main purpose is to pave the way for for planned changes to I420VideoFrame. See https://review.webrtc.org/38879004.
R=fbarchard@google.com, perkj@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/39889004
Cr-Commit-Position: refs/heads/master@{#8403}
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8403 4adac7df-926f-26a2-2b94-8c16560cd09d
Modifies WebRtcVideoSendStream to use a default width/height of 16px.
This significantly reduces SetRemoteDescription time under
WebRtcVideoEngine2. Also preventing (expensive) reconfigurations due to
incoming frames when the channel is not sending yet.
Tests have been modified to generate a frame before expecting a certain
encoder size to have been configured.
Also adding tracing to WebRtcVideoSendStream::InputFrame as it can lead
to reconfigurations of the encoder which is expensive and it should show
up in chrome://tracing.
BUG=1788
R=stefan@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/42369004
Cr-Commit-Position: refs/heads/master@{#8381}
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8381 4adac7df-926f-26a2-2b94-8c16560cd09d
Removes buffer/texture path separation inside WebRtcVideoEngine and
DeliverTextureFrame(). This unifies frame delivery with
WebRtcVideoEngine2 which is expected to automagically work with texture
frames after this change.
BUG=1788
R=magjed@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/38069005
Cr-Commit-Position: refs/heads/master@{#8326}
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8326 4adac7df-926f-26a2-2b94-8c16560cd09d
This is a big refactoring of the existing C++/JNI/Java support for audio recording in native WebRTC:
- Removes unused code and old WEBRTC logging macros
- Now uses optimal sample rate and buffer size in Java AudioRecord (used hard-coded sample rate before)
- Makes code more inline with the implementation in Chrome
- Adds helper methods for JNI handling to improve readability
- Changes the threading model (high-prio audio thread now lives in Java-land and C++ only works as proxy)
- Adds basic thread checks
- Removes all locks in C++ land
- Removes all locks in Java
- Improves construction/destruction
- Additional cleanup
Tested using AppRTCDemo and WebRTCDemo APKs on N6, N5, N7, Samsung Galaxy S4 and
Samsung Galaxy S4 mini (which uses 44.1kHz as native sample rate).
BUG=NONE
R=magjed@webrtc.org, perkj@webrtc.org, pthatcher@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/33969004
Cr-Commit-Position: refs/heads/master@{#8325}
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8325 4adac7df-926f-26a2-2b94-8c16560cd09d
Removes AddRenderCallback from ViERenderer and implements
VideoReceiveStream on top of DeliverI420Frame like WebRtcVideoEngine
currently does today.
Also adds ::IsTextureSupported() to the VideoRenderer interface to
permit querying whether an external renderer supports texture rendering.
R=stefan@webrtc.orgTBR=mflodman@webrtc.org
BUG=1667
Review URL: https://webrtc-codereview.appspot.com/34169004
Cr-Commit-Position: refs/heads/master@{#8299}
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8299 4adac7df-926f-26a2-2b94-8c16560cd09d
The surrounding similar methods all used unsigned char, using unsigned int in
this case looks like an accident, especially since the function passes on the
value in question to a function expecting a uint8.
BUG=none
TEST=none
R=mflodman@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/40529004
Cr-Commit-Position: refs/heads/master@{#8228}
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8228 4adac7df-926f-26a2-2b94-8c16560cd09d
* Remove "dead" and "alive" variables.
* Remove critical section
* Skip synchronizing with the worker thread to verify startup (no need).
* Remove implementation of SetNotAlive()
* Always set thread name
* Add thread checks for correct usage.
Also added some TODOs for myself for the ThreadWrapper interface.
I'm removing the HasNoMonitorThread test since it is no longer relevant and ends up checking the wrong thing (ProcessThread - a generic thread type) in the wrong way (parsing a debug log) :) I think it served a purpose some years ago, but things have changed since.
BUG=2902
R=henrika@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/37069004
Cr-Commit-Position: refs/heads/master@{#8220}
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8220 4adac7df-926f-26a2-2b94-8c16560cd09d
This fixes a two year old TODO of deleting dead code :)
In cases where the _id or id_ member variable is being used for tracing,
I changed the member to at least be const.
It doesn't look like id's are that useful anymore so maybe the next step is to get rid of them.
BUG=
R=henrika@webrtc.org, perkj@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/37849004
Cr-Commit-Position: refs/heads/master@{#8201}
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8201 4adac7df-926f-26a2-2b94-8c16560cd09d
