More about 3310 and envelopes

[Magnus Danielson]

From: Haible Juergen <Juergen.Haible at nbgm.siemens.de>
Subject: RE: More about 3310 and envelopes
Date: Wed, 15 Nov 2000 17:21:45 +0100

> 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 ! 

Certainly. Tony had some trouble in clarifying for everyone what the
problem really was about.

> 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.

Right. I agree. There are many solution to this dilema but the worst
solution is to avoid looking at it since then you migth totally miss
the train and loose the note or something.

> 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.

Should one has a "fade-to-new-note-time-constant" possibly? Basically
acting as a decay when you do a retrigger before the decay/release is
below the velocity of the new note on that voice.

Should one possibly just jump over the attack and go straight for the
decay phase?

These should be simple enougth to implement on most ADSR curcuits.

BTW. I have looked through all my OB-8 material and no mentioning of
the asymtotic voltage pin.

Cheers,
Magnus