On Tue, Jan 15, 2002 at 11:16:57 +0200, [EMAIL PROTECTED] wrote: > See http://www.johnson-amp.com/pdf/CabinetImaging.pdf for info on > speaker simulation.
Thanks, looks like they are using acoustic models or raytracing to model cabinets, which was my first idea, but its very time consuming. > Best way (but not effective) to simulate speaker would be to sample the > impulse response and use convolution (like AcousticMirror does). I guess > Johnson and Line6 use other regular IIR and FIR filters. I thin there > are some filter design apps available for the job. This is what I'm trying at the moment. My major concern is latency, I don't think I can get it below 256 or 512 samples (5 or 11ms @ 44k1) with reasonable CPU use. Also it would only work well at the capture frequency, I don't know how bad the effects of resampling impulses from 44k to 48k are offhand. I suspect that I would get better results with a swept sine, rather than an impulse, but deconvoloving impulses is trivial ;) and I'm just playing about at the moment. - Steve
