[PD-dev] does pd use dual buffer approach ?

Miller Puckette mpuckett at imusic1.ucsd.edu
Sun Aug 19 02:28:26 CEST 2007


Depending on the OS, you can get at least 100 milliseconds of buffering and
perhaps much more... so it should be fine.

cheers
Miller

On Sat, Aug 18, 2007 at 01:50:46PM -0700, Sergei Steshenko wrote:
> Hello All,
> 
> I have a very basic question regarding pd - does it use dual buffer approach ?
> 
> That is, suppose I'm going to use a FIR (FFT, for example) with long (several
> seconds) processing time (latency), and there is enough CPU performance to do
> processing in real time.
> 
> Still, if I want to hear uninterrupted sound, the sound card, having, say,
> a buffer of 4096 samples has to be fed new data much more often - with 44.1KHz
> sampling frequency it has to be fed about 11 times per second.
> 
> So, a possible solution is to have to buffers - while one of them is waiting for the
> new data from the FIR, the other piece by piece feeds data to soundcard, and when the
> feeding buffer is empty, the buffers are swapped.
> 
> So, how does pd deal with long latencies like in the case described above -
> does it use dual buffer approach in order to have uninterrupted sound ?
> 
> Thanks,
>   Sergei.
> 
> Applications From Scratch: http://appsfromscratch.berlios.de/
> 
> 
>        
> ____________________________________________________________________________________
> Choose the right car based on your needs.  Check out Yahoo! Autos new Car Finder tool.
> http://autos.yahoo.com/carfinder/
> 
> _______________________________________________
> PD-dev mailing list
> PD-dev at iem.at
> http://lists.puredata.info/listinfo/pd-dev




More information about the Pd-dev mailing list