[PD] subnormal numbers explained

Ivica Ico Bukvic ico at vt.edu
Thu Apr 4 02:27:03 CEST 2013


If freeverb~ is compiled properly and you are not running it on a platform
whose processor does not gracefully handle denormals, you should not have to
do anything.

> -----Original Message-----
> From: pd-list-bounces at iem.at [mailto:pd-list-bounces at iem.at] On Behalf Of
> João Pais
> Sent: Wednesday, April 03, 2013 1:36 PM
> To: katja; James Dunn
> Cc: pd-list
> Subject: Re: [PD] subnormal numbers explained
> 
> I add some very very very low noise to the audio signal, it really doesn't
> affect the audio. something like
> 
> [noise~]
> |
> [lop~ 5]
> |
> [*~ 0.0001]
> 
> it's dirty, but it works (?). Do you have any arguments against it?
> 
> João
> 
> 
> > Quoth katja, on 13/03/2013 10:14:
> >> Subnormal numbers are a pain in the ass, they cause substantially
> >> increased CPU load without doing anything useful for audio DSP.
> >> Unfortunately it can happen in some Pd objects (notably [freeverb~]),
> >> and spoil the performance of a Pd patch.
> > Yes I eventually found this out and started putting [freeverb~] in a
> > subpatch with a [env~] connected to [sel 0] and then to a [switch~]
> > object to turn off dsp when the reverb tail has finished. The logic
> > ended up getting quite complicated but it did the job!
> >
> > James
> >
> > _______________________________________________
> > Pd-list at iem.at mailing list
> > UNSUBSCRIBE and account-management ->
> > http://lists.puredata.info/listinfo/pd-list
> 
> _______________________________________________
> Pd-list at iem.at mailing list
> UNSUBSCRIBE and account-management ->
> http://lists.puredata.info/listinfo/pd-list




More information about the Pd-list mailing list