if we don't receive npt time mapping from the rtsp server (i.e. live stream)
Change-Id: I5147d665bd90c9a303ad6ffdafbf770f930f917c
related-to-bug: 5660357
No clients can signal a format change on either audio or video track (or both)
and a time discontinuity (timestamps changed) independantly.
Change-Id: I3e6cf4e7c260e85759879d61a9b517f68431c22f
related-to-bug: 5553055
a) one of the two decoders has a pending discontinuity
b) the renderer holds on to all output buffers for that decoder
c) the renderer is paused
if all three conditions are met the decoder won't ask for more input data
and therefore never see the discontinuity.
To avoid this we briefly resume the renderer just before shutting down.
Change-Id: I9e08af2a1eb4298d1cd00497d6aa33f4ad184e9a
related-to-bug: 5655016
This change moves the ANativeWindow connect and disconnect logic from
MediaPlayer to MediaPlayerService::Client.
Bug: 5502654
Change-Id: Ifc43b98b01ad8f35d62d7ece43110724ec7fda3d
When decoding a file for the SoundPool, do not
reject the entire file in case of error but
return what was decoded so far instead.
Change-Id: Iff199a1b6a4c8e064e42a0dfe0704e0ae36a27fd
- Drastically cut down the number of times we supply the AudioSink with data
by estimating the time until the sink would run out of data and then scheduling
a refill in advance of that.
- Use a dedicated looper for video decoders since they are currently taking
too long to return from OMX_FillThisBuffer (bug 5325201)
- Revise thread priorities for the OMX dispatcher and software codecs, instead
of running them at ANDROID_PRIORITY_AUDIO, they now only run at
ANDROID_PRIORITY_FOREGROUND
- Since threads created by pthread_create inherit all of the parent threads
attributes including thread priority, briefly reset thread priority to
ANDROID_PRIORITY_FOREGROUND before instantiating OMX components and then
restore it.
Change-Id: If9332a3a20dad5485333d68c11de0d2d5d3fffc3
this change intends to support its very limited case and signals an error in
all other cases of unexpected PID changes that we cannot recover from.
Change-Id: Icbfdf9fe7461969e2a8781ed416f54d891dd789a
those that have a queue. This ensures that the player doesn't observe discontinuities
that don't match up across streams.
Also, make sure output buffers arriving from the decoder to be rendered are sent
back to the decoder if we started flushing.
Finally, don't parse TS packets for streams we don't support. And don't allocate
memory for them.
Change-Id: I708e0de4cba8110a62e4c8ceb1e5702430d5d2bb
Codec errors (and codec not found errors) now trigger a controlled shutdown
of playback and signal errors to the MediaPlayer client.
Change-Id: I2ee23ff2a1422d05a1a21e50ecb87d7c7ab958cc
- Atribute network activity to uid calling the mediaplayer
- Enables logging of chromium network stack in logcat
Change-Id: I2d28c8392248a056b3cee305dd4d4475ebba4337
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
