This change is a complement to the main fix in kernel driver for the same issue (partner change #1250).
It removes clicks sometimes heard after the end of the tones while audio flinger is sending 0s to the audio output stream.
The problem was that the sleep time between two writes was more than the duration of one audio output stream buffer which could cause some underrun.
Also fixed a recent regression in ToneGenerator that made that the end of previous tone was repeated at the beginning of current one under certain timing circumstances when the maximum tone duration was specified.
This currently assumes 44k stereo (won't crash on other formats, but won't give the correct results either), and links statically with libspeex to get FFT data, increasing the size of libmedia by about 45kb.
When the AudioTrack callback notification size is relatively high (Which is the case on Sholes and over A2DP), it is likely that the end of tone is reached during the first callback. In this case, the AudioTrack is stopped before exiting the callback which causes 2 problems:
- 1: If the AudioFlinger thread is scheduled before we exit the ToneGenerator callback, the track can be stopped and reset before the data is actually marked as present in the buffer by the AudioTrack callback => no audio will be processed by AudioFlinger.
- 2: In this case, the data write index in the AudioTrack buffer is incremented after the track was reset by the AudioFlinger which leaves unplayed data in the buffer. This data will be played the next time the AudioTrack is started if not flushed in between.
The fix consists in adding an intermediate state to ToneGenerator state machine so that we exit the callback function when the stop condition is reached and stop the AudioTrack the next time we execute the callback.
This api allows to instantiate a renderer by specifying the hosting java Surface object. This hides the implementation details of (java-)Surface, (native-)Surface and friends.
Add a parameter to ToneGenerator.startTone() allowing the caller to specify the tone duration. This is used by the phone application to have a precise control on the DTMF tone duration which was not possible with the use of delayed messaged.
Also modified AudioFlinger output threads so that 0s are written to the audio output stream when no more tracks are ready to mix instead of just sleeping. This avoids an issue where the end of a previous DTMF tone could stay in audio hardware buffers and be played just before the beginning of the next DTMF tone.
* changes:
When encountering a .nomedia file, notify the MediaScannerClient, so that it can erase the data column for entries that are in the folder containing the .nomedia file. This prevents us from deleting (via a delete trigger) files when somebody adds a .nomedia file after the fact.
so that it can erase the data column for entries that are in the
folder containing the .nomedia file. This prevents us from
deleting (via a delete trigger) files when somebody adds a
.nomedia file after the fact.
commit c45bfbb97ccd05982008df47181f9c73abaf0497
Author: Andreas Huber <andih@google.com>
Date: Tue Sep 1 15:58:12 2009 -0700
This quirk should not be enabled by default in order to make the bug reproducible by the vendor.
commit 21d72e80e795fcae53d9c3bcc8ba6312b081e420
Author: Andreas Huber <andih@google.com>
Date: Tue Sep 1 15:55:45 2009 -0700
Undoing the hack to temporarily give up the lock to facilitate reading from the buffer source.
This simply causes too many issues, there need to be independent threads providing input buffers and dequeuing output buffers.
commit 84d507def8999c146ce124cc8edfe106c9ca70c2
Author: Andreas Huber <andih@google.com>
Date: Tue Sep 1 15:16:23 2009 -0700
The AAC components appear to output stereo data even if the input data is mono...
commit 3fa5f3ce910f34da0cedb8bcce1bd593db7eb740
Author: Andreas Huber <andih@google.com>
Date: Thu Aug 27 14:48:54 2009 -0700
Getting rid of more code now obsolete with the departure of OMXDecoder.
commit e0d923ec0c7bb60bec8eb97739036a4c70c0feef
Author: Andreas Huber <andih@google.com>
Date: Thu Aug 27 14:33:06 2009 -0700
Remove obsoleted OMXDecoder class. OMXDecoder is dead, long live OMXCodec.
commit 5bb012f0065f7ffaaeb4f569d71f0e3a8d6b19c3
Author: Andreas Huber <andih@google.com>
Date: Fri Aug 14 10:40:08 2009 -0700
An attempt at fixing export using the qcom encoders. More quirks.
commit 0690e76bfa48118a68287ccf1bbfa82febaa620c
Author: Andreas Huber <andih@google.com>
Date: Fri Aug 14 09:08:28 2009 -0700
Callbacks are now dispatched from a separate thread in OMX.
commit c6571a039526df29b6343f9a1971dbc019088c61
Author: Andreas Huber <andih@google.com>
Date: Thu Aug 13 15:42:25 2009 -0700
Massive API changes throughout stagefright, smart pointers everywhere.
commit 900612af6a0555664d9ba195112cd859491265f4
Author: Andreas Huber <andih@google.com>
Date: Thu Aug 13 13:33:12 2009 -0700
OMXCodecs now properly shutdown.
commit 96732f05e1b0603dcd1b11f16a23512592eeb4f5
Author: Andreas Huber <andih@google.com>
Date: Thu Aug 13 12:04:04 2009 -0700
More work on JPEG decoding using the hardware OMX component.
commit 63839a073ac393e3a130434ba467969053b694ad
Author: Andreas Huber <andih@google.com>
Date: Wed Aug 12 13:13:31 2009 -0700
An attempt to drive the JPEG decoder OMX node.
commit 3ac2fe5ab2926eda81b2123610b2434c645294ff
Author: Andreas Huber <andih@google.com>
Date: Tue Aug 11 16:38:21 2009 -0700
Renamed StateMachine to OMXCodec and put it in its proper place.
commit 247da75a96bf8881956413023dd49a84d5b4f5b2
Author: Andreas Huber <andih@google.com>
Date: Tue Aug 11 16:06:19 2009 -0700
Statemachine is now a full-fledged MediaSource.
commit 045244f6771fa0b9b329495c953afda900a84b71
Author: Andreas Huber <andih@google.com>
Date: Fri Aug 7 09:16:54 2009 -0700
Properly setup the input format when exporting to AMR audio.
commit 271b984cb32c5cd9e46e3f90ae121f334e4b8da9
Author: Andreas Huber <andih@google.com>
Date: Thu Aug 6 09:59:38 2009 -0700
Added some code to test audio encoding to the OMX harness.
commit 79af4748e4af33bd66d3fbac606e332a69741cf4
Author: Andreas Huber <andih@google.com>
Date: Wed Aug 5 14:36:22 2009 -0700
Merge the old OMXDecoder and the new, shiny, StateMachine code.
commit 91cf5dd77a8762bc10a0b2ffce35e3bbeb262231
Author: Andreas Huber <andih@google.com>
Date: Tue Aug 4 17:41:43 2009 -0700
A new harness to test OMX node compliance (and quirks).
Use integers instead of void* as input/output handles at IAudioFlinger and IAudioPolicyService interfaces.
AudioFlinger maintains an always increasing count of opened inputs or outputs as unique ID.
commit 374ea382ee3a9e3ce17e4c6357fc40d02e362810
Author: Andreas Huber <andih@google.com>
Date: Tue Jul 28 09:54:13 2009 -0700
PV's OMX implementation now uses (spec-compliant) microseconds instead of milliseconds in buffer timestamps.
commit 8d02f8ab5d7b022ad4ad34db2a9bdeea6ce2acfe
Author: Andreas Huber <andih@google.com>
Date: Mon Jul 27 14:24:26 2009 -0700
Support for using an overlay for video playback on TI hardware.
Appears to be currently constrained to CbYCrY 16-bit colorspace.
commit d17f321cb4b15c1fea378f33a7ef5998f23dd0fc
Author: Andreas Huber <andih@google.com>
Date: Mon Jul 27 09:45:38 2009 -0700
Added '--audio-only' commandline option to stagefright tool.
commit d8beef6be5c668c46451446d87e622933371cd75
Author: Andreas Huber <andih@google.com>
Date: Fri Jul 24 13:35:00 2009 -0700
Generalize the various workarounds for OMX nodes with their own unique interpretation of the spec.
commit c7dfd53eeadf8ed5a39bf2b19b014dcd62f3324d
Author: Andreas Huber <andih@google.com>
Date: Thu Jul 23 16:06:36 2009 -0700
Fixed erroneous TI AAC decoder behaviour wrt shutdown.
The AAC decoder appears to not return out buffers on an Executing->Idle transition, implemented a workaround that does a flush on all ports followed by the Executing->Idle. Oh, and flush with OMX_ALL doesn't properly work either. Fun.
commit d6377282e75060881336578f166f9c7feacf3f8f
Author: Andreas Huber <andih@google.com>
Date: Thu Jul 23 14:06:50 2009 -0700
Apparently the "channels" parameter in AudioTrack's constructor no longer counts channels but is a bitmask of enabled destinations, update the code accordingly.
commit ff698c79e851a2e57d362e9c3a09828af4048087
Author: Andreas Huber <andih@google.com>
Date: Wed Jul 22 16:54:44 2009 -0700
Fix MPEG4 decoding using TI's hardware decoder that I broke earlier...
commit 2ef78bb87cd856eb7f0b3d7dd68782a8650c12bf
Author: Andreas Huber <andih@google.com>
Date: Wed Jul 22 15:43:18 2009 -0700
Now that the qcom decoder properly advertises its own custom colorspace, update dependent code to reflect this fact.
commit bbaec96910727080fd7c8a6907c04facb9f5220f
Author: Andreas Huber <andih@google.com>
Date: Wed Jul 22 14:32:03 2009 -0700
Finishing up previous, incomplete change.
commit 76f14a1ae816b6f434771f8d12bdad81196f351e
Author: Andreas Huber <andih@google.com>
Date: Wed Jul 22 14:25:17 2009 -0700
The TI video decoder now properly decoder AVC/H.264 content.
commit e106130d8c100d5c94603e43864a7a93cca10252
Author: Andreas Huber <andih@google.com>
Date: Wed Jul 22 08:56:04 2009 -0700
Experimental support for the TI H.264 decoder, various modifications to OMXDecoder, non-functional still.
commit 241c3062dec3447db1a1ee74558cb4b9098fc404
Author: Andreas Huber <andih@google.com>
Date: Tue Jul 21 12:13:09 2009 -0700
Enable TI hardware OMX decoders (except for AVC/H.264 which still has issues).
This particular set of OMX nodes does not appear to properly return our buffers when sending the "disable" command on a port. Rather it reqires manually flushing that port and _then_ disabling it instead.
commit 1c34506a46e32ce25f2a86f3b4250dcfc037356a
Author: Andreas Huber <andih@google.com>
Date: Tue Jul 21 08:51:35 2009 -0700
Make it simpler to switch between the stagefright player and PVPlayer.
commit 249c6de05671d403f8dd51f095d49bf190430c9c
Author: Andreas Huber <andih@google.com>
Date: Mon Jul 20 14:38:15 2009 -0700
Prepare to use soon-to-be-available hardware decoders in the OMX decoder.
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()
Metadata.java:
Fixed typo 8k != 8092. The comment was correct though.
In Metadata.h, the new Metadata class is declared in the ns android::media
to limit the chances of conflict with other packages.
The MetadataType in MediaPlayerInterface is gone and moved to Metadata as
an inner typedef.
Similarly the SortedVector<MetadataType> instance have been replace by a
new type Metadata::Filter.
All the keys declared in the java counterpart are also in Metadata.h.
Metadata.cpp:
Contains the implementation of the native metadata packing.
There an associated change in the opencore package that should go in
at the same time as this one.
The method passes a list of metadata ids to be retrieved and
a parcel where the metadata records should be appended.
If the list of ids is empty, all the metadata should be returned.
