need sample code for IR or convolution reverb
-
- KVRian
- 922 posts since 26 Mar, 2003 from Guildford, England
Convolution itself is pretty simple, you just need a nested loop and a few multiply adds to convolve your IR array wiyth the incoming stream.
There is a dsp guide at: http://www.dspguide.com/
this covers basic convolution, and gives an implementation of 'brute force' covolution.
It can get more complex though, with the likes of FFT convolution, or other methods, which are neccessary when using long impluse responses.
Here is some stuff at musicdsp.org regarding convolution.
http://www.musicdsp.org/archive.php?classid=3#66
There is a dsp guide at: http://www.dspguide.com/
this covers basic convolution, and gives an implementation of 'brute force' covolution.
It can get more complex though, with the likes of FFT convolution, or other methods, which are neccessary when using long impluse responses.
Here is some stuff at musicdsp.org regarding convolution.
http://www.musicdsp.org/archive.php?classid=3#66
-
- KVRer
- Topic Starter
- 17 posts since 21 Jul, 2003
some good examples...thanks a lot.texture wrote:Convolution itself is pretty simple, you just need a nested loop and a few multiply adds to convolve your IR array wiyth the incoming stream.
There is a dsp guide at: http://www.dspguide.com/
this covers basic convolution, and gives an implementation of 'brute force' covolution.
It can get more complex though, with the likes of FFT convolution, or other methods, which are neccessary when using long impluse responses.
Here is some stuff at musicdsp.org regarding convolution.
http://www.musicdsp.org/archive.php?classid=3#66
anyone have example code for applying IR maps to create reverb?
-
- KVRAF
- 8388 posts since 11 Apr, 2003 from back on the hillside again - but now with a garden!
You could look at these also...
http://membres.lycos.fr/amycoders/tutor ... sfftm.html
http://www.ludd.luth.se/~torger/brutefir.html
http://ccrma.stanford.edu/~jos/filters/
HTH
DSP
http://membres.lycos.fr/amycoders/tutor ... sfftm.html
http://www.ludd.luth.se/~torger/brutefir.html
http://ccrma.stanford.edu/~jos/filters/
HTH
DSP
-
- KVRer
- Topic Starter
- 17 posts since 21 Jul, 2003
great links. thanks.duncanparsons wrote:You could look at these also...
http://membres.lycos.fr/amycoders/tutor ... sfftm.html
http://www.ludd.luth.se/~torger/brutefir.html
http://ccrma.stanford.edu/~jos/filters/
HTH
DSP
-
- KVRian
- 992 posts since 16 Dec, 2002
The musicdsp.org examples are almost exactly what I'm looking for - with one exception. I'm trying to figure out how to implement a simple convolution module in Reaktor, which ought to be doable (if incredibly slow) given the math primitives that are available. But - as far as I know there's no way to implement recursion in Reaktor, so ... does anyone have an example of a *linear* algorithm that can convolve two signals? I don't need the math background, just a simple brute force alg that I can try to translate into Reaktor-speak ... ?texture wrote:Here is some stuff at musicdsp.org regarding convolution.
http://www.musicdsp.org/archive.php?classid=3#66
-
- KVRian
- 922 posts since 26 Mar, 2003 from Guildford, England
I *think* this would work:
N would need to be equal to or greater than the number of samples in your IR.
It would be shockingly slow.
The Z-1 blocks are unit delays.
N would need to be equal to or greater than the number of samples in your IR.
It would be shockingly slow.
The Z-1 blocks are unit delays.
-
- KVRian
- 992 posts since 16 Dec, 2002
LOL - I love that term 'shockingly slow' ... hehehe ...
Thanks for the help. I'll let y'all know if I come up with something vaguely usable ...
Thanks for the help. I'll let y'all know if I come up with something vaguely usable ...