Current AudioTrack implementation enforces that the requested audio
buffer size is at least corresponding the audio latency.
This requirement is too strong and leads to problems with current
stagefright and AudioSink implementations when playing over output
streams with long latency.
Ultimately, the AudioSink design should be changed to specify a minimum
buffer size in time or frames units but not in buffer count units.
Change-Id: I8ba603956f92ac49143a8249572665aa548f2f0f
Conflicts:
media/libmedia/AudioTrack.cpp
Squashed merge from master-tungsten of the following changes:
commit 73d09e18c4557e583a1684d44d598a1a02fd0cf2
Author: John Grossman <johngro@google.com>
Date: Mon Jun 20 13:57:44 2011 -0700
Remove TungstenMisc and rename LinearTransform
Change-Id: Ie8aa3e24e09fdbf6ef8996c26deb9c5640e20d1b
commit 3114aabe76ad733b59929d87e49c68229f5ae2e8
Author: John Grossman <johngro@google.com>
Date: Fri Jun 3 10:47:16 2011 -0700
Name changes and spelling fixes.
+ Replace the term TungstenTime with the Eugene-approved term CommonTime.
+ Fix a spelling error in a comment I noticed.
Change-Id: I8c10d618206826d16055f78c7724e24443bb03fd
commit cbf2903ab6893b6e662514e2f6d670e268a419df
Author: John Grossman <johngro@google.com>
Date: Fri Apr 15 09:27:54 2011 -0700
Migrate Tungsten code from the HC-Tungsten to the Master-Tungsten branch.
Change-Id: I95372d913a0761d90168edb4016f5ece0ea74502
commit bc7c46aa629f9883e959ef23de8da297f9eb508b
Author: Jason Simmons <jsimmons@google.com>
Date: Mon Jun 20 13:59:17 2011 -0700
Create a separate class for timed AudioTracks
commit 43be3231034ff8537fdd84422a7954780038671f
Author: John Grossman <johngro@google.com>
Date: Mon Jun 27 18:59:12 2011 -0700
Move libaah_rtp over from the vendor directory.
Also move factor PipeEvent out into utils.
Change-Id: Id3877c66efe22d771cf3ef4877107e431b828e37
commit 17526eb3148c9c3d4365b6d5b47e8dc13bca71b6
Author: John Grossman <johngro@google.com>
Date: Mon Jun 27 17:06:49 2011 -0700
Name changes for the TRTP Players s/tungsten/aah/g
Change-Id: I55e9ad13003f6aa6a36955b54426a7efbe31ac51
commit 423fc1bfc0fda799c421a650c83c4b9293b1a08c
Author: Jason Simmons <jsimmons@google.com>
Date: Mon Jun 20 17:56:09 2011 -0700
More timed AudioFlinger changes requested by code review:
* change trimTimedBufferQueue to trimTimedBufferQueue_l
* create one timed audio buffer heap per client process instead of one per track
* grow the silence buffer on demand
* some error handling fixes in timed getNextBuffer
* calculate the next output PTS in all mixer and track hooks
Change-Id: Ifc51a08b55029b7c48902ab2f22933ad7bafe1ad
commit a148e2674b1d3cb73289b82b85c333f0a66824a9
Author: John Grossman <johngro@google.com>
Date: Mon Jun 20 17:02:24 2011 -0700
Move the A@H time service into frameworks/base
Change-Id: I5c570cde70e8931e205516cb33517585804ce841
commit dfa438fa49bdaeeb2ec5fd0d17b30d881608b6b1
Author: John Grossman <johngro@google.com>
Date: Mon Jun 20 11:55:36 2011 -0700
Fix the build after Mike's code moving.
Change-Id: Ia883643ded252168bcc5a70584ab6ce97bb05266
commit 04489474ec8e73efe1bf52918831f41659033162
Author: John Grossman <johngro@google.com>
Date: Fri Jun 17 14:19:50 2011 -0700
Refactor the local/common clock services.
This change is one of a set of 5 changes made to different repositories. Look
for this comment in all of them.
Refactor the local/common clock services in tungsten to match android best
practice. Notable changes include
+ The kernel no longer knows anything about common time. Common time has been
moved completely up into user land. This has an impact on the accuracy of the
timesync debugging code, and the netfilter assisted approach to network based
timesync is going to have to be modified.
+ The timesync driver used by A@H is now just local time driver.
+ The kernel no longer needs access to the linear transform math code, and it
has been removed.
+ A new HAL has been introduced to expose the concept of local time to the
system.
+ A non-slewable stub implementation of the local time HAL based on
CLOCK_MONOTONIC has been added.
+ The TungstenTime library has been eliminated. Its functionality has been
distributed among the common time binder service, the local time hal and the
linear transform utility code.
+ All clients of the old TungstenTime library have been changed to be clients of
the binder service, the hal and the utility code.
+ The reset_tt utilities have been removed, they no longer have a purpose in the
system.
+ more progress has been made in eliminating the word "tungsten" from the code.
Things left to do include
+ Finish getting rid of tungsten from the time service.
+ Move the time service into the framework; AudioFlinger's new timed mode
depends on it and the service cannot continue to live in vendor tungsten.
Change-Id: I999b6cfb4a9d267818a86d747c35eecfc6693101
commit d48194545eed1116a84d81e2fb53315d2b0701a7
Author: Jason Simmons <jsimmons@google.com>
Date: Thu Jun 16 14:22:46 2011 -0700
Change the interface of the AudioMixer and AudioBufferProvider to accept a presentation timestamp
Change-Id: Ice2df5628d45a7f77100e7008103b35b3d3160a4
commit 02561419db82b01ffb28df38000716c612988427
Author: John Grossman <johngro@google.com>
Date: Tue May 10 14:00:21 2011 -0700
Put in a hack for controling master volume in the policy manager.
Fix initial master volume reporting.
Change-Id: Ia6caf2bbc6083c5f99fab852baa40fff10fc5fc7
commit 549cdc3ba115dc654cdade261fb055c72c6cdb79
Author: John Grossman <johngro@google.com>
Date: Wed May 4 11:46:17 2011 -0700
Make certain the logic for computing the output stream mixing point is hardened
against underflow and overflow when input and output sample rates don't match.
Change-Id: I5ebea07c9938107b435bec7413418622767e4e16
commit 8043d8ed63f51e76d452d22be7d453d4a7794530
Author: Jason Simmons <jsimmons@google.com>
Date: Wed Apr 27 18:06:27 2011 -0700
Add the patch for timed audio support to the mono resampler
Change-Id: I526f34ae9d1e8e3b0ed2fb05af3d024d5c5fe711
commit 2be89486ef23f0b0b0cc2dc25a4c0ee691043f00
Author: John Grossman <johngro@google.com>
Date: Wed Apr 27 10:38:57 2011 -0700
Extend the AudioHWInterface to allow it to specify the initial master volume used by AudioFlinger.
Change-Id: I8823330801c927494cf7ca31a6b8f9264fbfbb26
commit ff89a4d5e37e6a05a2b03f79ab4e97833dd66393
Author: John Grossman <johngro@google.com>
Date: Wed Apr 27 09:07:14 2011 -0700
Fix an issue with inconsistent volume reporting.
Changed masterVolume() to return the same value as the last call
to setMasterVolume when the HW layer is implementing master
volume control. The masterVolume/setMasterVolume API seems to be
an idea which was abandonded a long time ago; as of today the
system only ever sets it to 1.0 at startup and then never changes
it. Until we can figure out how the concept of external
amplifier gain control fits into the Android audio framework,
Tungsten is exposing this API via a hack-tastic invoke back door
in the TungstenRXPlayer and needs the getter/setter results to be
consistent.
Change-Id: I2ac730fa8fc9ee28c88f1a8e6f2e493eb5b65544
commit 086511b2d19cceb976747ac23e12b73fc7c28bea
Author: Jason Simmons <jsimmons@google.com>
Date: Mon Apr 25 16:07:19 2011 -0700
Add handling of timed audio tracks in the generic resampling mixer
Change-Id: Ic3be1d21b1117f1b233808be543c28a0dcec4792
Change-Id: I6ec5d2bca9b8ebc0acd395a7dd92e1a48fcdfa9b
Signed-off-by: Mike J. Chen <mjchen@google.com>
Signed-off-by: John Grossman <johngro@google.com>
Signed-off-by: Jason Simmons <jsimmons@google.com>
Fixed problem in AudioTrack::restoreTrack_l() causing a permanent
failure if the IAudioTrack interface to AudioFlinger could not be
restored at the first attempt.
Change-Id: I039d4fe2dca8d3baf71f1a6c51119f27a67b6611
Do not force wake up the AudioTrack thread every 10ms if no timed
events (loop, markers..) have to be processed.
This will help reduce power consumption.
Change-Id: Icb425b13800690008dd07c27ffac84739e3dbba3
The problem occurs when activating or deactivating A2DP connection
while SoudPool has a channel active. This can happen quite frequently now
that the UI sound effects are enabled by default.
If PCM data is remaining in the AudioTrack buffer when it is restroyed and
re-created on the new AudioFlinger output thread, this data is flushed.
As a consequence, no underrun or request for new data callback is sent to
SoundPool and the sound channel remains active for ever as the end of the
sample is never detected.
Change-Id: I13e0c11e4ce3f83bff7f58d347ca814b6a86712b
When the A2DP headset is connected, there is a possible
race condition when the audio tracks are moved from
the mixer thread attached to the speaker output to the thread
attached to A2DP output.
As the request to clear the stream type to output mapping cache in
the client process is asynchronous, it is possible that the flag
indicating to the client audio track to re-create the IAudioTrack
on the new thread is processed before the cache is invalidated.
In this case, the track will be attached to the old thread and
music will continue playing over the device speaker instead of being
redirected to A2DP headset.
Change-Id: Ib2ce1eb5320eaff83287b93779061bf4e7a330df
AudioTrack::stop() is not synchronous, so a stop() followed
by flush(), which is synchronous, will not always report
a playhead position of 0 after being called.
This CL adds a flag to mark a track as flushed, and report the
correct playhead position in this state.
Bug 5217011 has been created to address the real issue in the
future, where flush could be made synchronous, to properly
address bug 4364249.
Change-Id: Icf989d41a6bcd5985bb87764c287f3edb7e26d12
Don't remove effects until the session they are in goes away or all
AudioEffects have been explicitly released. This allows the control
panel process to die without stopping the effects.
Change-Id: I4496e5df080230ca1af149dec95c1309ab8ea888
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
The first fix (commit 913af0b4) is problematic because it makes threads
in mediaserver process block on the cblk mutex. This is not permitted
as it can cause audio to skip or worse have a malicious application
prevent all audio playback by keeping the mutex locked.
The fix consists in using atomic operations when modifying the control
block flags.
Also fixed audio_track_cblk_t::framesReady() so that it doesn't block
when called from AudioFlinger (only applies when a loop is active).
Change-Id: Ibf0abb562ced3e9f64118afdd5036854bb959428
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
Make sure that all read/modify/write operations on the AudioTrack
and AudioRecord control block flags field are protected by the
control block's mutex.
Also fix potential infinite loop in AudioTrack::write() if the
written size is not a multiple of frame size.
Change-Id: Ib3d557eb45dcc3abeb32c9aa56058e2873afee27
This change fixes the stability problems experienced when using
a bluetooth headset supporting both A2DP and SCO. Problems occur
when starting the video chat at which time the A2DP output is being
stopped to start SCO. At that time, active AudioTracks are invalidated
by AudioFlinger so that a new AudioTrack binder interface can be
recreated by the client process on the new mixer thread with correct parameters.
The problem was that the process to restore the binder interface was not
protected against concurrent requests which caused 2 binder interfaces
to be created sometimes. This could lead to permanent client deadlock
if one of the client threads was waiting for a condition of the first
created binder interface while the second one was created (as the AudioFlinger
would only signal conditions on the last one created).
This concurrent request situation is more likely to happen when a client
uses the JAVA AudioTrack as the JNI implementation uses simultaneously the
native AudioTrack callback and write push mechanisms. By doing so, the code
that checks if the binder interface should be restored (in obtainBuffer()) is
much more likely to be called concurrently from two different threads.
The fix consists in protecting the critical binder interface restore phase
with a flag in the AudioTrack control block. The first thread acting upon the binder
interface restore request will raise the flag and the second thread will just wait for
a condition to be signaled when the restore process is complete.
Also protected all accesses to the AudioTrack control block by a mutex to prevent
access while the track is being destroyed and restored. If a mutex cannot be held
(e.g because we call a callback function), acquire a strong reference on the IAudioTrack
to prevent its destruction while the cblk is being accessed.
Modified AudioTrack JNI to use GetByteArrayElements() instead of
GetPrimitiveArrayCritical() when writing audio buffers. Entering a critical section would
cause the JNI to abort if a mediaserver crash occurs during a write due to the AudioSystem
callback being called during the critical section when media server process restarts.
Anyway with current JNI implementation, either versions do not copy data most of the times
and the criticial version does not guaranty no data copy.
The same modifications have been made to AudioRecord.
Change-Id: Idc5aa711a04c3eee180cdd03f44fe17f3c4dcb52
Use a Mutex wherever atomic operations were used in AudioTrack,
AudioRecord, AudioFlinger and AudioEffect classes.
Change-Id: I6f55b2cabdcd93d64ef19446735b8f33720f8dbc
This issue showed that when an AudioTrack underruns during a too long period
of time and is therefore disabled by audioflinger mixer, it takes an additional
delay of up to 3 seconds to recover.
This fix adds a simple mechanism to recover immediately when the client application
is ready to write data again in the AudioTrack buffer
Also throttle warnings on record overflows
Change-Id: I8b2c71578dd134b9e60a15ee4d91b70f3799cb3d
Added methods to AudioTrack and MediaPlayer java classes to enable use of
auxiliary audio effects. The effect can be attached and detached by specifying its
ID and the send level controlled.
Change-Id: Ie74ff54a453096a742688476f612ce355543b6f3
Audio sessions are used to associate audio effects to particular instances (or groups) of MediaPlayers or AudioTracks.
Change-Id: Ib94eec43241cfcb416590f435ddce7ab39a07640
First drop of audio framework modifications for audio effects support.
- AudioTrack/AudioRecord:
Added support for auxiliary effects in AudioTrack
Added support for audio sessions
Fixed left right channel inversion in setVolume()
- IAudioFlinger:
Added interface methods for effect enumeraiton and instantiation
Added support for audio sessions.
- IAudioTrack:
Added method to attach auxiliary effect.
- AudioFlinger
Created new classes to control effect engines in effect library and manage effect connections to tracks or
output mix:
EffectModule: wrapper object controlling the effect engine implementation in the effect library. There
is one EffectModule per instance of an effect in a given audio session
EffectChain: group of effects associated to one audio session. There is one EffectChain per audio session.
EffectChain for session 0 is for output mix effects, other chains are attached to audio tracks
with same session ID. Each chain contains a variable number of EffectModules
EffectHandle: implements the IEffect interface. There is one EffectHandle object for each application
controlling (or using) an effect module. THe EffectModule maintians a list of EffectHandles.
Added support for effect modules and effect chains creation in PlaybackThread.
modified mixer thread loop to allow track volume control by effect modules and call effect processing.
-AudioMixer
Each track now specifies its output buffer used by mixer for accumulation
Modified mixer process functions to process tracks by groups of tracks with same buffer
Modified track process functions to support accumulation to auxiliary channel
Change-Id: I26d5f7c9e070a89bdd383e1a659f8b7ca150379c
The problem is due to a too big difference between the buffer size used at the hardware interface and at the A2DP interface.
When no resampling occurs we don't notice problems but the timing is very tight. As soon as resampling is activated, the AudioTrack underruns.
This is because the AudioTrack buffers are not resized when moving the AudioTrack from hardware to A2DP output.
The AudioTrack buffers are calculated based on a hardware output buffer size of 3072 bytes. Which is much less than the A2DP output buffer size (10240).
The solution consists in creating new tracks with new buffers in AudioFlinger when the A2DP output is opened
instead of just transfering active tracks from hardware output mixer thread to the new A2DP output mixer thread.
To avoid synchronization issues between mixer threads and client processes, this is done by invalidating tracks
by setting a flag in their control block and having AudioTrack release the handle on this track (IAudioTrack)
and create a new IAudioTrack when this flag is detected next time obtainBuffer() or start() is executed.
AudioFlinger modifications:
- invalidate the tracks when setStreamOutput() is called
- make sure that notifications of output opening/closing and change of stream type to output mapping are sent synchronously to client process.
This is necessary so that AudioSystem has the new stream to output mapping when the AudioTrack detects the invalidate flag in the client process.
Previously their were sent when the corresponding thread loop was executed.
AudioTrack modifications:
- move frame count calculation and verification from set() to createTrack() so that is is updated every time a new IAudioTrack is created.
- detect track invalidate flag in obtainBuffer() and start() and create a new IAudioTrack.
AudioTrackShared modifications
- group all flags (out, flowControlFlag, forceReady...) into a single bit filed to save space.
Change-Id: I9ac26b6192230627d35084e1449640caaf7d56ee
Condition must be initialized with SHARED for the old behavior, where
they can be used accross processes.
Updated the two places android that require SHARED conditions.
PRIVATE conditions (and mutexes) use more efficient syscalls.
Change-Id: I9a281a4b88206e92ac559c66554e886b9c62db3a
At some point the implementation became complicated because of
SurfaceFlinger's special needs, since we are now relying on gralloc
we can go back to much simpler MemoryDealer.
Removed HeapInterface and AllocatorInterface, since those don't need
to be paramterized anymore. Merged SimpleMemory and Allocation.
Made SimplisticAllocator non virtual.
Removed MemoryDealer flags (READ_ONLY, PAGE_ALIGNED)
Removed a lot of unneeded code.
This is a second attempt to fix the audio routed to earpiece syndrom.
The root cause identified this time is the crash of an application having an active AudioTrack playing on the VOICE_CALL stream type.
When this happens, the AudioTrack destructor is not called and the audio policy manager is not notified of the track stop.
Results a situation where the VOICE_CALL stream is considered as always in use by audio policy manager which makes that audio is routed to earpiece.
The fix consists in moving the track start/stop/close notification to audio policiy manager from AudioTrack to AudioFlinger Track objet.
The net result is that in the case of a client application crash, the AudioFlinger TrackHandle object (which implements the remote side of the IAudioTrack binder interface) destructor is called which in turn destroys the Track object and we can notify the audio policy manager of the track stop and removal.
The same modification is made for AudioRecord although no bug related to record has been reported yet.
Also fixed a potential problem if record stop is called while the record thread is exiting.
This change goes with a kernel driver change that reduces the audio buffer size from 4800 bytes (~27ms) to 3072 bytes (~17ms).
- The AudioFlinger modifcations in change 0bca68cfff161abbc992fec82dc7c88079dd1a36 have been removed: the short sleep period was counter productive when the AudioTrack is using the call back thread as it causes to many preemptions.
- AudioFlinger mixer thread now detects long standby exit time and in this case anticipates start by writing 0s as soon as a track is enabled even if not ready for mixing.
- AudioTrack::start() is modified to start call back thread before starting the IAudioTrack so that thread startup time is masked by IAudioTrack start and mixer thread wakeup time.
Create a new IAudioTrack interface to AudioFlinger when start() fails due to a broken pipe error.
Do the same if start fails due to the same error after time out in obtainBuffer().
Do not indicate that the AudioTrack is started to AudioPolicyManager if IAudioTrack start fails.
This avoids that an AudioTrack keeps a dead IAudioTrack after a media server crash.
Same modifications for AudioRecord.
Add a flag to ToneGenerator indicating that the callback thread can call Java. Without it, when the media server crashes and restarts, the AudioSystem error callback will crash in JNI if the IAudiotrack is created from AudioTrack callback thread.
Initial commit for review.
Integrated comments after patch set 1 review.
Fixed lockup in AudioFlinger::ThreadBase::exit()
Fixed lockup when playing tone with AudioPlocyService startTone()
Merge commit '88e209dcf8c2ebddda5c272f46d1bd5478bc639c'
* commit '88e209dcf8c2ebddda5c272f46d1bd5478bc639c':
Fix issue 1743700: AudioTrack: setPlaybackRate can not set the playback rate to twice of the ouputSR
Store sample rate on 32 bits instead of 16 bits in audio_track_cblk_t.
Removed sampleRate() methods from AudioTrack and AudioRecord: replaced by getSampleRate().
AudioTrack::setSampleRate() no returns a status.
AudioTrack, AudioRecord:
- remove useless mAudioFlinger member of AudioTrack and AudioRecord.
- signal cblk.cv condition in stop() method to speed up stop completion.
- extend wait condition timeout in obtainBuffer() when waitCount is -1 to avoid waking up callback thread unnecessarily
AudioFlinger:
- remove some warnings in AudioFlinger.cpp.
- remove function AudioFlinger::MixerThread::removetrack_l() as its content is never executed.
- remove useless call to setMasterVolume in AudioFlinger::handleForcedSpeakerRoute().
- Offset VOICE_CALL stream volume to reflect actual volume that is never 0 in hardware (this fix has been made in the open source): 0.01 + v * 0.99.
AudioSystem.java:
- correct typo in comment
IAudioflinger, IAudioFlingerClient:
- make AudioFlinger binder interfaces used for callbacks ONEWAY.
AudioHardwareInterface:
- correct routeStrings[] table in AudioHardwareInteface.cpp
