[jsyn] sample accuracy

Phil Burk philburk at s...
Tue Nov 20 00:18:55 EST 2001


Hello Brad,

> I'm fooling with some stuff where I need to get sample-accurate delays
> (like, 4 samples or 3 samples) and haven't seen any units that do quite
> this in the autodocs.

Hmmm. This would be good to add. How about if I added these units:

DelayOne for [Z-1]
DelayTwo for [Z-2]
DelayThree for [Z-3]
DelayFour for [Z-4]
DelayEight for [Z-8]

Note that if you have feedback in the circuit then you will get a delay of
at least 8 samples because of the block based calculations. So you have to
be careful of the order of execution when doing feedback, or when splitting
and mixing streams.

The order of execution is determined by the order that units are start()ed.

I am considering adding a mode where the calculation block size can be set
to one instead of eight for a circuit to allow precise control over delay in
a feedback circuit.

> If I use "delay" and specify time delays of 3/44100
> or 4/44100 would that work? I suppose another approach would be to
> cascade 1o_1p fiters and use coeffs of 0 or 1 appropriately, but that
> would involve some unnecessary multiplies.

Both of those would probably work but are not very efficient, as you point
out.

Phil Burk






More information about the JSyn mailing list