I guess the bufferCount is the sizeof integer which is 4. The second term is something audioFlinger and audioTrack is introducing. Needs to be figured out. The latency introduced is not acceptable for real time applications.
On Tue, Nov 30, 2010 at 10:42 PM, RockNCode <[email protected]> wrote: > I Sambhav. > > I have this same question regarding the buffering. What I can say is > that I think the reason that the second term is bigger than the first > is because the mFrameCount can be 4 times the size of the driver's > output buffer (or a little bit more). mFrameCount is calculated in > MediaPlayerService.cpp in status_t > MediaPlayerService::AudioOutput::open() this way: > > frameCount = (sampleRate*afFrameCount*bufferCount)/afSampleRate; > > Where bufferCount is a constant on the media player side (with a value > of 4). I'm also clueless what the buffercount value means or how was > it calculated. So the only variable in the equation seems to be the > clip's sampling rate (as afSampleRate, afFrameCount and afLatency seem > to be constants obtained from the driver). > > I hope we can get this doubts resolved. > > - Regards. > > On Nov 30, 1:54 am, Sambhav <[email protected]> wrote: > > Hi, > > > > Audio Track has a API to query the latency *AudioTrack::latency()* > > > > This latency is calculated with the equation > > > > mLatency = afLatency + (1000*mFrameCount) / sampleRate; > > > > where mFrameCount and sampleRate are parameters given in AudioTrack > > initilzation. > > *afLatency* is the latency obtained from the alsa driver. > > The second part of the equation is significantly a big number. > > > > Does anyone know how AudioFlinger/AudioTrack is adding to the latency ? > > How buffering is done at the framework level before giving it to the > drivers > > ? > > > > Regards, > > Sambhav > > -- > unsubscribe: > [email protected]<android-porting%[email protected]> > website: http://groups.google.com/group/android-porting > -- unsubscribe: [email protected] website: http://groups.google.com/group/android-porting
