[sdiy] Ideal polyphonic digital reverb

[cheater cheater]

let's enumerate some of the possibilities:


voice 1: vco1 -> (a1) -> vcf1 -> (b1) -> vca1 -> (c1)
voice 2: vco2 -> (a2) -> vcf2 -> (b2) -> vca2 -> (c2)
...
voice n: vcon -> (an) -> vcfn -> (bn) -> vcan -> (cn)

at a1, you want to have one of a few things:

(i) a convolution engine that takes the output of vco1, applies an IR,
and puts it into the input of vcf1
(ii) same as (i) but the input is all vcos (1..n) and the output is vcf1
(iii) same as (i) but the output is all vcfs (1..n)
(iv) same as (i) but the input is the last few notes you played in
decreasing order: 1/2 * the vco from the most current note + 1/4 * the
vco from the previous note + 1/8 * the vco from the note previous to
that etc
(v) same as (iv) but you do that with the outputs
(vi) same as (i) but there's also input from b1 or c1

then you want the same topology at b and c

They're all obviously going to have different results and different
applications.

And obviously you can think about all sorts of dsp in here, not just
convolution, and when it comes to convolution, then not just reverb.

On Mon, Mar 22, 2021 at 7:03 PM Mike Bryant <mbryant at futurehorizons.com> wrote:
>
> Thought I'd change the topic name as we're definitely moving away from cheap.
>
> So given a polyphonic synth with unlimited voices and unlimited budget, what should the structure of an ideal reverb system look like ?
>
> Once there's a consensus I'll try coding it into my digital synth.
>
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at synth-diy.org
> http://synth-diy.org/mailman/listinfo/synth-diy
> Selling or trading? Use marketplace at synth-diy.org