[sdiy] ot: rotating speaker simulation or stupid approach

[jhaible]

> The technical term for the resulting pattern is 'a mess.' ;-)

All you have to do is make the increments small enough, and all
the "mess" will be outside the audible range.

> The *only* way to get this to work would be to get each speaker producing
> just a single wavefront. That would work. But for obvious reasons that's
> not going to be practical, or useful.

I don't know about wavefronts, but for your notches ("mess") being above
20khz, a delay resolution of 25 us is required. This is not so bad if you
consider that we're talking of a "delay amplitude" of only a few
milliseconds.

Further, for a Leslie emulation, your upper limit isn't 20kHz, but something
like 5 (or 8 ?) kHz, so you need even less resolution.

I think in practice you should be able to get good results with sampling 30
... 40
different positions. (To just cover the "rotating" effect, approx. 10 ...15
positions should be enough.) Just a rough calculation and gut feeling; feel
free to do the exact calculations if you like.

JH.