More about 3310 and envelopes

[Haible Juergen]

	>A while ago JH reported that one version of the Oberheim OB-8 used
the
	>attack and asymptote to control the amplitude of the ADSR being
sent to
	>the VCA/VCF for modulation depth. At the Synth do around Paul's at
the
	>weekend, we tried to find this on the schematics to no avail.
Although
	>there was a distinct lack of the 3080 used to control ADSR envelope
	>depth found in the older versions of the same synth, OB-Sx, Xa etc.
	>Perhaps they thought they would hide their little secret.
	>
	>I have tried this on a more traditional ADSR, and it works well,
but for
	>one important thing. I wanted to use this method of amplitude
control to
	>adjust the modulation for different velocities. If you set the ADSR
to
	>have a long release, and press one note again and again at
different
	>velocities, you get a major problem. If the cap is still charged up
to a
	>higher level than the new incoming attack peak, all hell breaks
lose.
	>The decay phase never gets initiated and the envelope itself
doesn't
	>sound natural. Does the OB suffer from this, or does the velocity
not
	>control the modulation depth direct. If the depth is only changed
at the
	>patch level, then this wouldn't be a problem.
	>
	>Any ideas?

The OB-8 has no velocity, so there is no such problem.

But you're raising a most interesting question: How *should* the envelope
of a retriggered voice behave when the new velocity value is lower than the
"remaining" amplitude from the previous note ?

Not easy ! 

Consider the *same* note is struck again for that voice: Should the
amplitude
jump back abruptly to the new value, or should the maximum of both values
be valid, or should the sustaining ("releasing") note dominate ? Or should
the tiny amount of the low velocity note be *added* to the sustaining
previous note ?
When a vibrating Guitar String is plucked again, I guess it will stop
vibrating
as long as the finger touches the string. (?)
When the reed of an electric piano is hit with the damper open, some
of the newly applied energy might add to the sustaining note's energy
I suppose. So there is no general solution.

Now think of the new note assigned to that voice being different to the
previous one. Say, the voice is sustaining at C with a momentary level of
70,
and a D# with an initial level of 30 (lower velocity) is bound to replace
that C.
I'd say the big question is not *how* that note should be replaced (in terms
of level), but *if* it should be replaced under these conditions at all !
Replacing a loud note with a soft note abruptly is a striking event
regardless
how you do it.

JH.