On Thu, Feb 06, 2003 at 06:03:47 +0100, Dave Griffiths wrote:
It happens on all IEEE machines, though on some (eg. the PS2's vector
units) you can turn it off.
If your machine is memory bound then you wont notice as much.
Cheers for the info guys, I haven't come across this before
On Thu, Feb 06, 2003 at 11:13:45 +0100, David Olofson wrote:
On Thursday 06 February 2003 16.28, Steve Harris wrote:
[...]
#define FLUSH_TO_ZERO(fv) (((*(unsigned
int*)(fv))0x7f80)==0)?0.0f:(fv) I think it came from the
music-dsp list.
There's a conditional in there, though.
Yes,
On Fri, Feb 07, 2003 at 12:56:38 +0100, David Olofson wrote:
I say we should make it as simple as strongly recommending that
plugins with algorithms that tend to generate denormals easily,
should deal with it in a suitable way. Sounds both better and easier
to me.
Agreed, plugins that can
On Friday 07 February 2003 10.55, Steve Harris wrote:
On Thu, Feb 06, 2003 at 11:13:45 +0100, David Olofson wrote:
On Thursday 06 February 2003 16.28, Steve Harris wrote:
[...]
#define FLUSH_TO_ZERO(fv) (((*(unsigned
int*)(fv))0x7f80)==0)?0.0f:(fv) I think it came from the
On Fri, Feb 07, 2003 at 02:16:08 +0100, David Olofson wrote:
On Friday 07 February 2003 10.55, Steve Harris wrote:
On Thu, Feb 06, 2003 at 11:13:45 +0100, David Olofson wrote:
On Thursday 06 February 2003 16.28, Steve Harris wrote:
[...]
#define FLUSH_TO_ZERO(fv) (((*(unsigned
On Friday 07 February 2003 21.02, Steve Harris wrote:
[...]
Well, casting to int avoids getting the denormal into the FPU.
This is probably the only safe way to deal with a denormal
without forcing the FPU to burn cycles evaluating it. I don't see
how this has anything to do with the
On Fri, Feb 07, 2003 at 09:38:46 +0100, David Olofson wrote:
Yeah - but not as cheap as generating occasional spikes, a tone or
maybe even white noise, to prevent the generation of denormals in the
first place.
Besides, if you *get* a denormal to kill, it has to come from
somewhere.
On Thu, 6 Feb 2003 14:07:06 +, Steve Harris wrote
On Thu, Feb 06, 2003 at 12:18:47 +0100, Dave Griffiths wrote:
I had a vague attempt at doing something like this (after noticing that
filters filtering silence uses up a lot of cpu). Each sample buffer object
OT: Thats probably because
Dave Griffiths wrote:
ahah, I was hoping for an explanation :) any ideas on how to combat this, what
the squashing threshold should be?
I wrote a paper on denormalized number issue, you can check
it here:
http://www.musicdsp.org/files/denormal.pdf
-- Laurent
On Thu, Feb 06, 2003 at 03:59:31 +0100, Dave Griffiths wrote:
On Thu, 6 Feb 2003 14:07:06 +, Steve Harris wrote
On Thu, Feb 06, 2003 at 12:18:47 +0100, Dave Griffiths wrote:
I had a vague attempt at doing something like this (after noticing that
filters filtering silence uses up a lot
perfect!
On Thu, 06 Feb 2003 16:09:11 +0100, Laurent de Soras [Ohm Force] wrote
Dave Griffiths wrote:
ahah, I was hoping for an explanation :) any ideas on how to combat this, what
the squashing threshold should be?
I wrote a paper on denormalized number issue, you can check
it here:
On Thu, 6 Feb 2003 15:28:00 +, Steve Harris wrote
On Thu, Feb 06, 2003 at 03:59:31 +0100, Dave Griffiths wrote:
On Thu, 6 Feb 2003 14:07:06 +, Steve Harris wrote
On Thu, Feb 06, 2003 at 12:18:47 +0100, Dave Griffiths wrote:
I had a vague attempt at doing something like this
On Thu, Feb 06, 2003 at 04:46:10 +0100, Dave Griffiths wrote:
Is this processor specific? I used to get it loads on my PII desktop, but I
haven't noticed it as much my PIII machine (might just be because it's twice
the speed).
It happens on all IEEE machines, though on some (eg. the PS2's
On Thu, 6 Feb 2003 16:15:46 +, Steve Harris wrote
On Thu, Feb 06, 2003 at 04:46:10 +0100, Dave Griffiths wrote:
Is this processor specific? I used to get it loads on my PII desktop, but I
haven't noticed it as much my PIII machine (might just be because it's twice
the speed).
It
On Thursday 06 February 2003 16.28, Steve Harris wrote:
[...]
#define FLUSH_TO_ZERO(fv) (((*(unsigned
int*)(fv))0x7f80)==0)?0.0f:(fv) I think it came from the
music-dsp list.
There's a conditional in there, though.
Another method is to add noise or some other signal (beep at Nyqvist)
David Olofson wrote:
On Thursday 06 February 2003 16.28, Steve Harris wrote:
[...]
#define FLUSH_TO_ZERO(fv) (((*(unsigned
int*)(fv))0x7f80)==0)?0.0f:(fv) I think it came from the
music-dsp list.
There's a conditional in there, though.
Another method is to add noise or some other
On Friday 07 February 2003 01.22, Simon Jenkins wrote:
[...]
(IIRC:)
If a plugin has got denormals coming out of its outputs, its pretty
certain that its got them circulating around an internal feedback
loop also, since thats where denormals generally come from.
Killing denormals at the
On Thursday 06 February 2003 16.28, Steve Harris wrote:
[...]
#define FLUSH_TO_ZERO(fv) (((*(unsigned
int*)(fv))0x7f80)==0)?0.0f:(fv) I think it came from the
music-dsp list.
There's a conditional in there, though.
Another method is to add noise or some other signal (beep at
On Friday 07 February 2003 00.57, Tim Hockin wrote:
On Thursday 06 February 2003 16.28, Steve Harris wrote:
[...]
#define FLUSH_TO_ZERO(fv) (((*(unsigned
int*)(fv))0x7f80)==0)?0.0f:(fv) I think it came from the
music-dsp list.
There's a conditional in there, though.
Maybe a -200dB sine at 1Hz and Nyquist? But then a BP filter
screws you.
White noise is pretty good...
A very simple Denormal-Zapper plugin which injects white-noise at -200dB
(or lower) can be inerted anywhere in the chain. A Very Useful Plugin.
Tim
On Friday 07 February 2003 01.38, Tim Hockin wrote:
Maybe a -200dB sine at 1Hz and Nyquist? But then a BP filter
screws you.
White noise is pretty good...
A very simple Denormal-Zapper plugin which injects white-noise at
-200dB (or lower) can be inerted anywhere in the chain. A Very
David Olofson wrote:
On Friday 07 February 2003 01.38, Tim Hockin wrote:
Maybe a -200dB sine at 1Hz and Nyquist? But then a BP filter
screws you.
White noise is pretty good...
A very simple Denormal-Zapper plugin which injects white-noise at
-200dB (or lower) can be inerted
Tim Hockin wrote:
Maybe a -200dB sine at 1Hz and Nyquist? But then a BP filter screws you.
Someone once suggested a slight DC offset, though I can't see how that would
solve things like a reverb, unless they preserve it and shift their 0.
I generally suggest to add random peaks here and
23 matches
Mail list logo