>From: Eric Scheirer <eds@media.mit.edu>
>
>> A FIR filter could be made with
>> y = x(-1)*a + x(0)*b + x(1)*c;
>> but I don't know if that is already possible. I'm not able to verify
>> it from SAOL specs now.
>
>Perhaps I don't understand what you are suggesting.
The suggestion was that notation. The above is an example how it can
also be used.
>like you suggest that this syntax could allow the implementation
>of pre-causal filters! Surely, we can't implement the above
>FIR filter except by adding a constant delay throughout,
Yep but I have a question: if I have 2001 tap FIR interpolation filter
and I do
y = FIR(x) + x;
then is FIR(x) misaligned with respect to x by 1000 samples?
And to get it corrected I should write
y = FIR(x) + delay(x,1000);
or such?
What I would like to see is that FIR is not delayed but it is pre-causal
if possible. If we have an audiofile in memory, then it is easy to
implement pre-causal filter. But if we have a complex flow structure,
is it possible to remove the delays? At least in the above example
it could be removed by implicitly adding the delay to x.
Say, if I have a piano sample and I mark the attack place such that
it can be snatched at the sequencer's bars. If I then make an arbitrary
effects with SAOL, the attack place should be still at the same place.
Yours,
Juhana
This archive was generated by hypermail 2b29 : Wed May 10 2000 - 12:14:10 EDT