As far as causality is concerned it's the *group* delay that should be
non-negative. Phase delay assumes sinusoidal inputs over (-inf, inf) for
which whether a system is causal makes no difference.
-Original Message-
From: music-dsp-boun...@music.columbia.edu
From: music-dsp-boun...@music.columbia.edu
[mailto:music-dsp-boun...@music.columbia.edu] On Behalf Of robert
bristow-johnson
well, even group delay is negative with the peaking filters, for
*some* frequencies.
Yes, but only if the filter has high (negative?) dispersion at that
- when considering finite duration there is the uncertainty principle, so
you always deal with a pack of frequencies rather than one frequency, which
makes latency dependent on the content of that pack.
- however, using FT[th(t)]=j(FT[h(t)])', one can show that the centroid of
time w.r.t. the
AD2136x doesn't seem to have FP division instructions (?) therefore I guess
one has to engage some homemade routine to do that. Some chips have a
built-in lookup table of low-precision reciprocals to help cooking faster
division routines. A usual digital arithmetic textbook will have the
