[music-dsp] Is beating the same thing as flanging?
scoofy at inf.elte.hu
Sun Nov 14 13:05:57 EST 2010
Richard Dobson wrote:
> True, but it depends. I have checked the DAFX book (Zolzer) about this
> (surprising how so many general dsp books hardly discuss allpass filters
> at all!), and there is a clear distinction between first and second
> order allpass behaviour. For first order, the group delay (delay wrt
> frequency) is indeed a curve(flat at the extremes) ranging between 3
> samples and fractionally below 1 (the fractional delay being a very
> important application of allpasses). For second order, the range is much
> greater - at Fc the group delay can be surprisingly large - the Zolzer
> figure indicates around 28 samples - down to almost zero elsewhere.
I don't understand something. How does Zoelzer get a figure of '28
samples' at Fc, when a 2nd order allpass filter changes the phase of a
sinusoidal frequency component by 180 deg at the cutoff freq? That means
a half cycle delay, but the length of a half sinusoid cycle (in samples)
is different at every frequency. So, for every different Fc, he should
get a different number of delay (in samples).
Or did I miss something here?
More information about the music-dsp