From the original SynthMaker forum (2004–2014), with its members’ names and dates.
tor ·
@ 44100 SR
Is the only cure for frequency warping close to nyquist in iir's oversampling?
I guess it is the same question. But.. analogish bell curves close to nyquist in iir's also only possible with OS?
I see that when using the OS toolkit the warping is gone. But the OS toolkit also have a LP filter at the end at SR*1???? Why is not the output of the down sampler warped? 😕
Is there a place where I can read to understyand this with out actually do massive and advanced maths. I'm trying to read 'Digital Signal Processing' by Paulo Diniz. Eduardo Silva, Sergio Netto. But I do quicly get lost caused by lack of math and language knowledge. This is heavy reading......... and as far as i can tell there is notting about coeffisient maths in this book. If there is a solution to my qustion it must be about coeff's.
Slowly learning each day 😉
MichaelBenjamin ·
i would guess read just the parts you understand and ignore/overfly the rest.
if the oversampler works for you i would not be so much concerned about the lp filtering, for practical uses, its irrelevant.
tor ·
The reason I ask this is because I decided to use double precision floats in the biquad equation. And this is as far as I can see not possible to do at 2*SR in SM. The warping is only a problem at SR 44100 and the benfit of double precision is as I can understand better than OS. Exept for the warping near nyqusit at 44100. In sampling rates from 48000 and above the warping is less or if at all audible. Orfanidis has caught my attesion.. but I have not yet managed to wrap my head about the info I found so far.
Wrote something here? To have a post removed or your name hidden, write to [email protected].