Maybe I miss the real question of the topic but I have played around with creating a FIR filter: 1. generate white noise of a desired length 2. window it with an exponentially decaying envelope 3. apply some gain, e.g. 0.5 4. add a Dirac pulse at the first sample The result is sprectrally not flat but 5. compute the excessphase of the sum = allpass = spectrally flat and use it
- Uli 2016-08-02 2:31 GMT+02:00 gm <g...@voxangelica.net>: > > > Am 01.08.2016 um 22:55 schrieb Evan Balster: > >> The most essentially flat signal is a delta function or impulse, which is >> also phase-aligned. Apply any all-pass filter or series thereof to the >> impulse, and the fourier transform over infinite time will remain flat. I >> recommend investigating Schroeder filters. >> > > I already played with them as well as FDNs. > Though Shroeder allpass filters in series (or reverbs in general) are not > strictly flat it's better than random. > > And it's a trade off to have an impact like onset. > You get that "like gaussian smear" smear, unless you set your diffusion > coefficient high > which also makes the responses longer. And the onset ist a little bit > unnatural. > (I know you can "nest" them and change that a little bit) > > Either way this way it comes down to reverb design... which quite a trap > to waste time with... > it's never finished in a way, at least for me > > And related to reverbs, the question: > - how do I create a spectrally flat short decaying noise-like and > impact-like sequence > becomes interesting again, I think. > > But maybe there is nothing that's better than Shroeder allpass? > I started to use random sequences for early reflections but found that > these colors the sound too much, so I basically came to the same question. > Though for reverb a more sparse noise would be better... > > And for allpass delays the question remains > - how to design optimal length ratios? > That's why I made the slightly nonsense remark about RNGs and reverbs the > other day. > > So far I just use my ears, assumptions and numerology. > > > > _______________________________________________ > dupswapdrop: music-dsp mailing list > music-dsp@music.columbia.edu > https://lists.columbia.edu/mailman/listinfo/music-dsp > >
_______________________________________________ dupswapdrop: music-dsp mailing list music-dsp@music.columbia.edu https://lists.columbia.edu/mailman/listinfo/music-dsp
