Re: [alsa-devel] Help! Need to switch between normal speaker and BT headset dynamically
As my question is related to the query what pradeep is asking and mark brown also suggested to me to use PA. I have some additional questions regarding PA which i would like to clarify before starting:
1.Does PA have good resampler code if my hardware doesn't support resampling?
I have a system where we are using ALSA as the base and on top(application side) of it there is a thread running which is pumping data and this thread is started by java code.Combination of LINUX+JAVA. This thread can perform these tasks: 1.Resampling for some rates. 2.Talking to PV framework to get the decoded PCM data and pumping audio data to alsa. 3.Routing support using asound.conf file. Can i replace this thread with PA by writing some wrapper over PA?? What extra features PA is going to provide additionally if i start using PA other than what is listed above??As i know PA is kind of sound deamon which my thread is not.
Is there any good resampler code which i can use without any licensing problem and which can be easily integrated???
'Twas brillig, and ANISH KUMAR at 06/04/10 12:37 did gyre and gimble:
As my question is related to the query what pradeep is asking and mark brown also suggested to me to use PA. I have some additional questions regarding PA which i would like to clarify before starting:
1.Does PA have good resampler code if my hardware doesn't support resampling?
Yes. Infact this code will likely be needed in several cases when the h/w does support resampling. If PA has multiple streams playing at once, the h/w will likely only be able to cope with one rate at a time.
The resamplers are pluggable and can have optimized implementation for some hardware - e.g. I believe Palm wrote an ARM specific resampler to do this for their Pre platform (I could be mis-remembering tho').
I have a system where we are using ALSA as the base and on top(application side) of it there is a thread running which is pumping data and this thread is started by java code.Combination of LINUX+JAVA. This thread can perform these tasks: 1.Resampling for some rates. 2.Talking to PV framework to get the decoded PCM data and pumping audio data to alsa. 3.Routing support using asound.conf file. Can i replace this thread with PA by writing some wrapper over PA?? What extra features PA is going to provide additionally if i start using PA other than what is listed above??As i know PA is kind of sound deamon which my thread is not.
One very interesting feature of PA that is particularly relevant to mobile platforms is the use of a timer-based model to drive the hard ware. This allows for very large latencies to be used (e.g. 10-20s) allowing for very efficient processor usage - only waking up when absolutely necessary, not to service interrupts.
More info: http://0pointer.de/blog/projects/pulse-glitch-free.html
This approach can bring power savings and thus increased battery life.
Is there any good resampler code which i can use without any licensing problem and which can be easily integrated???
I pretty sure the included resamplers are OK to use, but it's all clearly documented so just take a look.
Like I say it's easy enough to include your own too if needs be.
Col
participants (2)
-
ANISH KUMAR
-
Colin Guthrie