Hi Hans, Is this using the audio driver that interfaces with the JRE calls which fill the hardware buffers via Java? If so, one thing to double-check is whether those buffers are actually being filled only when they are empty, not just constantly filled even when they don't need to be. Is there somewhere I can look at the code?
Chris. On Thu, 13 May 2010 14:40:12 -0400, Hans-Christoph Steiner <[email protected]> wrote: > Sounds like that's the problem. [metro 1000] to [realtime] is giving me > 150-200ms. This is with -nosound. Any ideas where to look? > > .hc > > Miller Puckette wrote: >> Hi Hans - >> >> if it even misbehaves with -nosound most likely the OS isn't waking >> Pd up reliably. A test would be to try the "realtime" object to see if >> Pd is getting the time correctly from Android -- relevant for the >> -nosound >> case. If so I can't see why Pd wouldn't run "on time" unless it simply >> bogs the CPU down (uses 600% of CPU time). >> >> cheers >> Miller >> >> On Sun, May 09, 2010 at 01:21:39PM -0400, Hans-Christoph Steiner wrote: >>> So I have Pd running on Android, and am currently testing the timing. I >>> made a simple [metro 1000]--[print] and I seem to be getting 3-6 bangs >>> per second. The audio API stuff isn't entirely worked out, and I seem >>> to recall some relation between the audio I/O and timing. >>> >>> Any ideas why the timing might be so off? I've tried with both -noadc >>> and -nosound as well, and got more or less the same result. >>> >>> .hc >>> >>> _______________________________________________ >>> Pd-dev mailing list >>> [email protected] >>> http://lists.puredata.info/listinfo/pd-dev > > > _______________________________________________ > Pd-dev mailing list > [email protected] > http://lists.puredata.info/listinfo/pd-dev ------------------- http://mccormick.cx _______________________________________________ Pd-dev mailing list [email protected] http://lists.puredata.info/listinfo/pd-dev
