Josep M Comajuncosas wrote:
> 
> Another idea to try could be making the delay time variable...
> I really need an allpass unit with variable delay time to easily
> implement phaser effects without deltapi units.

For a phaser, I think that the delay time remains the same - 1 unit
sample. The feedback/feedforward gain coefficients are changing
dynamically.  Run several of these allpass stages in series, mix the
output of them with the input signal, and you got yer phaser.

However, allpass units with variable delay time would be VERY useful for
other purposes.  For short delay times, they could be used to create a
deep flanging effect with less comb filter coloration.  For longer delay
times, they can be used in reverberation networks, with some very slight
low frequency modulation being used to dynamically alter the resonances
of the reverb network.  Joe Dattorro talks about using both of these in
his recent JAES articles.

I would like to see several new allpass generators:

- A single unit delay allpass stage, for constructing allpass networks.
Similar to delay1, but ideally the k-rate would not have to be set to
the a-rate for things to work.  The gain coefficient could be an i-rate,
k-rate, or a-rate variable.  Would be useful in constructing phasers,
Dome filters, and for creating nonlinearities in waveguide physical
modeling.

- A unit generator similar to alpass, but with variable delay time, as
well as output taps.

I may end up working on such unit generators for my spring project for
my computer music class; my C isn't super strong, but I will be working
on it this quarter.  Any advice for learning the internals of Csound is
always welcome. I have the source code here, Kernigan & Ritchie,
"Elements of Computer Music" by F. Richard Moore, as well as lots of
other good books on computer music, but any additional tips would be
great.

Thanks,