||Crazy Combined Feedback/Feedforward Patch
| br>I was reading a recent article by Sean Costello (algorithm programmer of EOS and Shimmer VST reverbs amongst other things), and came across this crazy flow chart for implementing an allpass filter:
Here's the original article:
Now, not necessarily trying to do the same thing at all (regarding reverbs and delays etc.), but just looking at that patch generally, and thinking about applying it to my own modular synth patches, has started to fry my brain.
I've used feedforward and feedback techniques in my own patches before, but never thought of combing them both in one patch. What would this do? Some kind of weird cancelling effect? Would be interesting to see how it works in a compressor patch.
I love seeing flow charts and then thinking about how I could implement them on my modular, but with a totally different goal in mind. br> br>
| br>Sean Costello's blog is really simply fascinating if you have an interest in reverb, especially early ones and various approaches taken.
So what is z-M there, a delay? I can see the b0 is a dry signal with the inverted effect -aM mixed into it. So the effect processes the dry plus it's negative feedback. The final mix is the effect mixed with b0 (dry + inverted effect).
So the comb effect that would normally happen and be audible is being canceled out leaving an all pass? Or did I interpret that wrong.
I guess what engineers have been thinking about and tweaking for decades are various tricks to be able to feed enough output back into multitapped delays that it keeps on decaying yet doesn't generate the typical byproducts of feedback, you know, in the worst case squealing building up to overload. I understand decay relates to the amplitude going back in being high enough. So the secret is processing what's going back into the system so it's not the same leading to bad feedback results but yet not so different that it sounds wrong.
The original Eno implementation of "Shimmer" is both very clever & ironic. As I understand it he had a reverb running but he split the output and sent some to a vintage digital pitch shifter to be used just as feedback back into the reverb. I gather the early pitch shifter messed up the returned feedback so he could run it back in with a good amplitude and get an extended shimmer. br> br>
| br>Hey Nick,
Some interesting points!
Yes, I believe Z-M is the delay line. But I'm more interested in trying it out with other modules. How about a filter or a phaser, for example?
I often use a frequency shifter or a pitch shifter in the feedback loop of a delay, just as you say, in order to be able to get a lot more audible feedback without runaway feedback. I think I posted a little about it here:
The Shimmer VST is well worth getting, a manual has just been released. Very nice sounding, and not too pricey! br> br>
| br>>Yes, I believe Z-M is the delay line. But I'm more interested in trying it out with other modules. How about a filter or a phaser, for example?
Might be interesting, though I get the feeling the flowchart demonstrates how to make a delay functionally like an all-pass with potentially light code involved.
A phaser is a series of allpass filters allowed to phase cancel the dry signal
Though I'm not totally sure maybe one can call the above diagram turning a flanger into a phaser?
>The Shimmer VST is well worth getting, a manual has just been released. Very nice sounding, and not too pricey!
Yes! I picked that up after giving it a demo spin. He also has a free Valhalla FreqEcho with a simpler reverb and a frequency shifter br> br>
Powered by phpBB © phpBB Group