Convolution reverb?

Algorithm development and general DSP issues

Moderator: frank

Post Reply
livingston
Posts: 131
Joined: Sun Nov 15, 2009 3:37 pm
Location: New Orleans, LA US

Convolution reverb?

Post by livingston »

I'm just curious if this is possible in the FV-1. I'm guessing it's too processor-intensive. You would need some external memory to hold the impulses, then when you want to use one you'd need to record the file into the delay memory, then some sort of multiply (but obviously not just a mulx). It's beyond me at the moment, but is it possible?
frank
Posts: 1244
Joined: Wed Oct 19, 2005 12:26 pm
Contact:

Post by frank »

Well, you could probably pull off a short convolution, no where near enough points to be useful. But someone may have a trick I haven't thought of, so if you figure it out would really like to see the code.
Frank Thomson
Experimental Noize
B.C.
Posts: 38
Joined: Sat Mar 06, 2010 5:28 pm

Post by B.C. »

You have peaked my interest(no pun intended). I might play around with this...
livingston
Posts: 131
Joined: Sun Nov 15, 2009 3:37 pm
Location: New Orleans, LA US

Post by livingston »

How do you plan to approach it? I couldn't think of any way to even conceptualize the code for this.

I used to have a lot of fun on computer recording systems, convolving a piano with a cymbal, for example. The convolution concept can be extended to a lot of interesting things beyond reverb. For something like this, you don't need a really long impulse, so it might be more attainable on the FV-1 than a hall reverb convolution, for example.
B.C.
Posts: 38
Joined: Sat Mar 06, 2010 5:28 pm

Post by B.C. »

Well, I plan on exploiting the idea of convolving a signal with itself. Should be a pretty interesting result sound wise.
livingston
Posts: 131
Joined: Sun Nov 15, 2009 3:37 pm
Location: New Orleans, LA US

Post by livingston »

Hmm, so how would that differ from multiplying a signal with itself? Is there some sort of time component?
B.C.
Posts: 38
Joined: Sat Mar 06, 2010 5:28 pm

Post by B.C. »

Right, well Convolution is multiplication in the frequency domain, but in the time domain it has a time dependence. I believe I won't have to take as many samples (or atleast find a way around it) by doing it this way.
Post Reply