AW: EGs

Haible Juergen Juergen.Haible at nbgm.siemens.de
Mon Jun 23 12:03:38 CEST 1997


	> i'm planning to build a machine which will produce sounds
continuously
	> evolving and morphing one into another (pardon my english).
so, rather
	> than have a usual saw/pulse/tri alternative, i'd like to make
something
	> which would continuously morph between these. so far, the only
idea that
	> comes to my mind is to feed an EG with a pulse. in theory,
i'll have six
	> parameters (pitch, pw, adsr) to play with, but will this work
in real
	> world? will i have to make my own egs (oh, no!) or it's
possible with
	> existing chips?

Oh yes, you can make your own EGs ! (;->)

First, you could use a static crossfading like my "Interpolating
Scanner",
which works very nicely at audio rate. (Piecewise linear - PWL
-approximation
of waveforms)

But I really like your idea as well, and what you need is *not* a full
VCADSR
circuit, but only a very simple "1-stage envelope" which is reprogrammed
with a new asymptode CV ("sustain") and time CV ("decay") for each step.
This "envelope" would not be more than a 1-pole VCF, i.e. 1 OTA plus
a few discrete devices. The two CVs would be updated with a simple 
shift register (sequencer). You'd want to have the time CV input
exponential,
with a keyboard tracking function included.
The whole thing would not be as predictable as the PWL solution, but
this
may be a feature. You can also chosse between linear and exponential 
slopes by either using a straight VCF or a slew limiter circuit for the
EG part.

JH.





More information about the Synth-diy mailing list