[sdiy] Random DSP thoughts

[Scott Gravenhorst]

madhun2001 at yahoo.com wrote:
>Hey all,
>
>A few ideas of what the heck to do with one of these new-fangled DSP thingies:
>
>Clap Machine. No, not some kind of eeevil experment, but an 
>attempt to make a clap synthesizer that mimics real clapping. A 
>gizmo that takes either samples of claps or just generates claps, 
>and can do rhythmic or random clapping. Can even play with slight 
>delays between the channels to do some spacial stuff. Yeah, I 
>know, it's been done. But not by me. Yet. You see, I hate canned 
>clap machines. Hate, hate, hate. I think we can do better. 

Clap machines.  Hmm.  What's bad about sampled claps is there's probably only one in any
commercial machine - that's crappy and we all know it.  Amplitude modulation and pitch
modulation might help - a little, but this method doesn't really fool the brain into
thinking the claps are real, even if the sample is a very good one.  Perhaps hundreds of
samples that are randomly selected would help, along with amplitude and pitch modulation. 
But getting good clean samples is not an easy task.  And such a machine would (I think)
have to be rather memory rich to pull that off.  I would not know how to approach a
physical model of a clap...

Different people on the various synth forums and synth websites have often said that trying
to replicate natural sounds is quite difficult due to all of the nonobvious nuances within
the sound.  To synthesize even some of them begins to require more and more resources and
some of them are difficult to discover and model properly.  Thus, the answer here is
usually: if you want a hand clap that sounds natural and realistic - use a pair of real hands.

>VoSim. This is the vowel simulator out of the Chamberlin book. 
>Creates a decaying train of sin^2 pulses with 3 parameters. Might 
>make a fun module. 

I'd like to learn more about what this is.

>Drums. A simulation of an actual drum head with the ability to 
>hit anywhere on the surface. Simulates the not-so-harmonics of a 
>drum, the rim, rhe shell, and the stick. Yeah, I know, it's been 
>done. Into the ground it's been done. So what? 

Go back to what I wrote about hand claps, the same can be said about drums.  Resource
intensive stuff to include as many nuances as can be discovered and modeled properly.  I'm
going to suggest that a reasonably sized FPGA is probably the way to go.

>Inharmonic Oscillator. Venture into pseudo-additive synthesis 
>with an oscillator whose harmonics are variable in frequency. Be 
>able to have the harmonics shift over the note duration. 

Why psuedo?  Use a reasonably large FPGA and you can make a couple of hundred very accurate
sine waves all at once with EGs for each to control them.  This sounds like a lot of fun.

>Just a few things rattling around in my head. 

Hey, can we make a physical model of that rattling?  (c:


-- ScottG
________________________________________________________________________
-- Scott Gravenhorst
-- FPGA MIDI Synthesizer Information: home1.gte.net/res0658s/FPGA_synth/
-- FatMan: home1.gte.net/res0658s/fatman/
-- NonFatMan: home1.gte.net/res0658s/electronics/
-- When the going gets tough, the tough use the command line.