[LAD] [ANN] Radium V1.9.1

Kjetil Matheussen k.s.matheussen at notam02.no
Sun Nov 11 13:04:39 UTC 2012


On 11.11.2012 13:46, Fons Adriaensen wrote:
> On Sun, Nov 11, 2012 at 12:24:08PM +0100, Florian Paul Schmidt wrote:
>
>> Find /proc/cpuinfo below. Sadly I don't know the reason for this
>> crash..

I don't know either. -mtune=generic -msse should produce legal code
for your processor...


>> Maybe Fons knows something as it's in the Zita-Code?

Pretty sure the problem is that sse math crashes. It's just a 
coincidence
that zita reverb was the first code to use sse math. Radium would crash
later for you anyway, even if zita reverb wasn't used.

>
> That code was generated by Faust. IIRC it is considerably less
> efficient than the original C++ code, and it adds nothing new.
> So I wonder why you don't use the original, in which case I
> could provide support.
>

Using the Faust version made it possible to port the zita reverb in 
only a
few minutes, since I have a faust architecture file for Radium. In 
addition,
I don't need the equalizer controls and the dry/wet controls (since 
those
things are available on all sound objects in Radium), and removing
that code from your version (C++) is more work than removing the same
code from Faust code.




More information about the Linux-audio-dev mailing list