Path: utzoo!censor!geac!yunexus!xrtll!mark From: mark@xrtll.UUCP (Mark Vange) Newsgroups: comp.sys.amiga Subject: Re: WARNING: Gadget Alert! Keywords: reverb, sound, amiga Message-ID: <351@xrtll.UUCP> Date: 11 Sep 89 04:18:26 GMT References: <3278@wasatch.utah.edu> <860@abvax.UUCP> Organization: Wcom, Toronto, Ont Lines: 31 In article <860@abvax.UUCP>, set@xanadu (Scott Townsend) writes: > > I haven't done anything with the audio device, but shouldn't it be possible > for a new/possibly massively changed audio device to do digital reverb within > the machine, instead of buying an add-on? Maybe not quite as great in sound > quality, but it could provide some reverb and maybe stereo imaging effects. > (speculation) > I'm thinking along the lines of audio.device remembering the previous N samples > sent to the D/A (somehow in sync with the DMA) and altering the samples yet > to some to get the reverb effect. Maybe 8 bits would be too much quantization > error in the reverb components. Anyone care to comment? experiment? The problem is that while you are piping the "reverbed" (I shall call it digested) sound, you must also keep playing the sound which came down the pike more recently. To do this you must either break the sound down into it's spectrum, and do a bit of addition, along with FFT there and back, or you must double your playback rate, and "mix" the new sounds with the old reverbed stuff. The former is probably not practical except in the case of a dedicated playback system because it would be quite CPU intensive. The second is still overly simplified, because the new sound would have to be fed back into the digested signal and mixed with the older stuff, although this could be accomplished by an algorithmic hack rather than the FFT approach, especially if you didn't mind a constant rate of fade. -- Mark Vange Phone Death Threats to: PAS Systems - "Plain and Simple" (416) 730-1352 mark@xrtll 8 Everingham Ct. North York "Every absurdity has a champion Ont, Canada M2M 2J5 to defend it." - Oliver Goldsmith