[PD] DSP and Gem in the same instance of Pd

Roman Haefeli reduzent at gmail.com
Thu Mar 31 11:19:30 CEST 2016


On Thu, 2016-03-31 at 11:13 +0200, Roman Haefeli wrote:
> Hey all
> 
> It's a well know issue and the solution has been discussed many times:
> When you're using audio and Gem rendering, it's wise to separate those
> two into their own instance of Pd.
> 
> I'm working on a project that does both, audio processing and Gem
> rendering. The OpenGL rendering is very light-weight and does only
> require minimal resources. Also the audio part uses a fraction of the
> available CPU power. What I observe is that I get a zipping noise for
> ~15 seconds, then clean audio for another 15 seconds before the zipping
> noise returns. The zipping noise occurs only when Gem rendering is
> turned on. 
> 
> To me it looks like the problem is not related to limited resources,
> rather it seems that both, OpenGL and audio rendering, are tied to their
> respective clock and those clocks are not coupled and thus are not
> completely in sync. 
> 
> I know I could put both parts into their own instances of Pd and that's
> what I'll do if there is no resolution, but for convenience reasons I'd
> prefer to have only one patch in one instance of Pd. Now, is there a way
> to loosen the coupling to OpenGL clock (is it the same as the monitor
> clock?) so that audio is not affected? I don't care about any
> side-effects like tearing (it won't be visible anyway for the kind of
> image). 

BTW: Why does the graphics rendering|clock have precedence over the
audio rendering (at least, it seems to be like that in Pure Data/Gem)? I
guess most softwares do it the other way around, since clicks are much
more noticeable than a frame being a few milliseconds late. 

Roman

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: This is a digitally signed message part
URL: <http://lists.puredata.info/pipermail/pd-list/attachments/20160331/6769731d/attachment.sig>


More information about the Pd-list mailing list