Making a connection from MediaRecorder Native layer to the
SurfaceMediaSource for the purpose of encoding GL Frames. This will be
called from the java side inside the Mobile Filter Framework.
The mediarecorder native layer (client), when set the videosource to
option VIDEO_SOURCE_FRAMES, asks the StageFrightRecorder on the mediaserver
side to create a SurfaceMediaSource object and pass it back as a
sp<ISurfaceTexture> object. Using that, the client side will dequeue and
queue buffers. Connecting the GL Frames to the obtained
sp<ISurfaceTexture> is not part of this CL.
Related to bug id: 4529323
Change-Id: I651bec718dd5b935779e7d7a050b841c2d0b0fcd
The purpose of ICameraRecordingProxy and ICameraRecordingProxyListener is to
allow applications using the camera during recording.
Camera service allows only one client at a time. Since camcorder application
needs to own the camera to do things like zoom, the media recorder cannot
access the camera directly during recording. So ICameraRecordingProxy is a proxy
of ICamera, which allows the media recorder to start/stop the recording and
release recording frames. ICameraRecordingProxyListener is an interface that
allows the recorder to receive video frames during recording.
ICameraRecordingProxy
startRecording()
stopRecording()
releaseRecordingFrame()
ICameraRecordingProxyListener
dataCallbackTimestamp()
The camcorder app opens the camera and starts the preview. The app passes
ICamera and ICameraRecordingProxy to the media recorder by
MediaRecorder::setCamera(). The recorder uses ICamera to setup the camera in
MediaRecorder::start(). After setup, the recorder disconnects from camera
service. The recorder calls ICameraRecordingProxy::startRecording() and
passes a ICameraRecordingProxyListener to the app. The app connects back to
camera service and starts the recording. The app owns the camera and can do
things like zoom. The media recorder receives the video frames from the
listener and releases them by ICameraRecordingProxy::releaseRecordingFrame.
The recorder calls ICameraRecordingProxy::stopRecording() to stop the
recording.
The call sequences are as follows:
1. The app: Camera.unlock().
2. The app: MediaRecorder.setCamera().
3. Start recording
(1) The app: MediaRecorder.start().
(2) The recorder: ICamera.unlock() and ICamera.disconnect().
(3) The recorder: ICameraRecordingProxy.startRecording().
(4) The app: ICamera.reconnect().
(5) The app: ICamera.startRecording().
4. During recording
(1) The recorder: receive frames from ICameraRecordingProxyListener.dataCallbackTimestamp()
(2) The recorder: release frames by ICameraRecordingProxy.releaseRecordingFrame().
5. Stop recording
(1) The app: MediaRecorder.stop()
(2) The recorder: ICameraRecordingProxy.stopRecording().
(3) The app: ICamera.stopRecording().
bug:2644213
Change-Id: I15269397defc25cbbcae16abc071c8349c123122
Record and playback objects (resp AudioRecord and AudioTrack)
are created using a channel mask, but this information is lost
in the mixer because only the channel count is known to
AudioFlinger. A channel count can always be derived from a
channel mask.
The change consists in:
- disambiguiting variable names for channel masks and counts
- passing the mask information from the client to AudioFlinger
and the mixer.
- when using the DIRECT ouput, only verifying the format of
the track is compatible with the output's for PCM.
Change-Id: I50d87bfb7d7afcabdf5f12d4ab75ef3a54132c0e
through listener during video playback.
- Add OnTimedTextListener in the MediaPlayer
For feature request 800939.
Change-Id: I65072c27acb4c0037109a72be38c73e9f667420f
Instead of returning 0-filled buffers after EOS from AudioTrack we do this
work in AudioOutput instead. That way the EOS signal (0 frames returned)
is preserved in AudioCache which otherwise would lead to a heap size overflow
filling everything with zeroes.
Change-Id: I7e07429ba887957a4340dd4b21eef4bba76248cd
related-to-bug: 3514073
this is the first step in unifying surfacetexture and surface.
for this reason the header files were not moved, as most of them
will eventually go away.
NOTE: currently we keep libsurfaceflinger_client.so as an empty
library to workaround prebuilt binaries wrongly linking against
it.
Change-Id: I130f0de2428e8579033dc41394d093f4e1431a00
The problem is that when switching from A2DP to device speakers or headset,
The AudioTrack binder interface to AudioFlinger must be destroyed and restored
to accomodate new buffer size requirements. Current AudioTrack implementation
did not restore properly the PCM buffer write index which caused a mismatch between
the written frame count in the mediaplayer renderer and the AudioTrack. The renderer
could then believe the AudioTrack buffer was full and stop writing data preventing the
AudioTrack to reach a bufffer full condition and resume playback.
The rendered was also modified to refresh the AudioTrack frame count (buffer size)
inside the write loop in NuPlayer::Renderer::onDrainAudioQueue() as this count can change
from one write to the next.
Also modified AudioTrack::obtainBuffer() to check for track invalidated status before
querying for available space in the buffer. This avoids writing to the old track's
buffer until full before detecting the invalidated condition and create a new track.
Change-Id: I16a857e464e466880847f52f640820aa271539ad
- To track the currently used audio device
- The devices are separated as speaker and other audio devices
- Provide the collected data to battery application through pullBatteryData()
Change-Id: I374c755266b5ac6b1c6c630400f4daf901ea8acc
This change enables the use of a SurfaceTexture in place of a Surface
as the video sink for an android.media.MediaPlayer. The new API
MediaPlayer.setTexture is currently hidden.
This includes:
- New Java and C++ interfaces
- C++ plumbing and implementation (JNI, Binder)
- Stagefright AwesomePlayer and NuPlayer use ANativeWindow
(either Surface or SurfaceTextureClient)
Change-Id: I2b568bee143d9eaf3dfc6cc4533c1bebbd5afc51
This will fix the stop failure issue where we have to wait n * time_interval before a key frame can be received by the file writer, where
o n is the actual number of buffers advertised by the video encoder
o time_interval is the interval settings for timelapse video recording
specifying the time distance between neighboring input video frames
The fix includes two parts:
o OMXCodec will not submit all n buffers at one time, but instead submit one input
frame at one time if it become available.
o Timelapse camera source made available the first two input frames and do not skip
them so that the first compressed output frame data can be received regardless
the specified time_interval
bug - 3367659
Change-Id: Ia68cc2cb0d71aa7dc54540e9ad82fae911ad530b
- play audio-only streams again
- workaround for malformed streams that switch PIDs across bandwidths
- attempt to pick a different bandwidth stream if the previously chosen one appears
to be malformed/unsupported.
Change-Id: I426d0a40dc725aa242f619d4c9d048b69aca55c9
related-to-bug: 2368598
Due to the variable delays in the system, the recording sound can still be heard
if the audio track is started early. This happens even after we suppress a relative
long period of initial audio for a recording session. By starting the video track
before the audio track, we effectively play the recording sound earlier.
bug - 3371141
Change-Id: I2cccf77eae953d96827fc080c618b1d962445c26
