Add support for deep audio buffers
Allow AudioSink to use deep audio buffering when the
source is audio only and its duration is more than
a certain threshold.
This helps improve battery life but implies higher
audio latency.
Change-Id: Ie79915b61c370292f05aabda9779356570e03cbb
diff --git a/media/libmediaplayerservice/MediaPlayerService.h b/media/libmediaplayerservice/MediaPlayerService.h
index 2a8cfd2..95b1b05 100644
--- a/media/libmediaplayerservice/MediaPlayerService.h
+++ b/media/libmediaplayerservice/MediaPlayerService.h
@@ -91,7 +91,8 @@
virtual status_t open(
uint32_t sampleRate, int channelCount, audio_channel_mask_t channelMask,
audio_format_t format, int bufferCount,
- AudioCallback cb, void *cookie);
+ AudioCallback cb, void *cookie,
+ audio_output_flags_t flags = AUDIO_OUTPUT_FLAG_NONE);
virtual void start();
virtual ssize_t write(const void* buffer, size_t size);
@@ -135,6 +136,7 @@
int mAuxEffectId;
static bool mIsOnEmulator;
static int mMinBufferCount; // 12 for emulator; otherwise 4
+ audio_output_flags_t mFlags;
// CallbackData is what is passed to the AudioTrack as the "user" data.
// We need to be able to target this to a different Output on the fly,
@@ -190,7 +192,8 @@
virtual status_t open(
uint32_t sampleRate, int channelCount, audio_channel_mask_t channelMask,
audio_format_t format, int bufferCount = 1,
- AudioCallback cb = NULL, void *cookie = NULL);
+ AudioCallback cb = NULL, void *cookie = NULL,
+ audio_output_flags_t flags = AUDIO_OUTPUT_FLAG_NONE);
virtual void start();
virtual ssize_t write(const void* buffer, size_t size);