Re: [Faudiostream-users] [Faudiostream-devel] SR now a float

2016-06-16 Thread Yann Orlarey
Hi Julius, You can. Actually the fSamplingFreq field is still an int and still accessible. In the definition of SR it is converted to float or double according to the floating point size option of the Faust compiler. Is it enough ? Cheers, Yann Yann Orlarey Directeur scientifique www.grame.fr

Re: [Faudiostream-users] [Faudiostream-devel] SR now a float

2016-06-14 Thread Julius Smith
Hi guys, Making SR a float is great, but can we make it double? A 32-bit float can specify a quartz clock frequency to only about 1% accuracy. - Julius On Tue, Jun 14, 2016 at 11:40 AM Albert Graef wrote: > Hi Yann, > > > On Tue, Jun 14, 2016 at 7:40 PM, Yann Orlarey

Re: [Faudiostream-users] [Faudiostream-devel] SR now a float

2016-06-14 Thread Albert Graef
Hi Yann, On Tue, Jun 14, 2016 at 7:40 PM, Yann Orlarey wrote: > But you know that I would not make you such a bad thing, right ? > Really? BTW, talking about high blood pressure, I'm still waiting for that LAC17 proposal to show up on lac-team some time... Touché. ;-) >

Re: [Faudiostream-users] [Faudiostream-devel] SR now a float

2016-06-14 Thread Yann Orlarey
Hi Albert, Sorry for the heart attack ! But you know that I would not make you such a bad thing, right ? Concerning a test suite, you can use the one in tools/faust2appls/test. It tests all the faust2xxx scripts. Run first ./testfailure it tests that all scripts correctly report Faust and C++

Re: [Faudiostream-users] [Faudiostream-devel] SR now a float

2016-06-14 Thread Albert Graef
I'm relieved. This almost gave me a cardiac arrest, I'm still working on getting the blood pressure down again. ;-) Stéphane, thanks a bunch for clarifying, and sorry for the false alarm, but Yann's post didn't make this awfully clear. Yann, I hope you enjoyed that prank! :) On Tue, Jun 14, 2016

Re: [Faudiostream-users] [Faudiostream-devel] SR now a float

2016-06-14 Thread Stéphane Letz
Albert; look at the commit : SR definition has been changed in math.lib, it does not change any public interface and so should not break anything… Stéphane > Le 14 juin 2016 à 17:21, Albert Graef a écrit : > > Yann, maybe it would be possible to discuss such changes

Re: [Faudiostream-users] [Faudiostream-devel] SR now a float

2016-06-14 Thread Albert Graef
Yann, maybe it would be possible to discuss such changes *before* they actually happen? ;-) This opens a can of worms for me. Probably means that the Pure-Faust bitcode interface is broken now. And this might affect other architectures as well. We really need a Faust test suite which covers at