I can't participate in this interesting techtalk, but I think
Vangelis also experienced some double-triggering. At least, I think I
hear a double-triggered note on the track "I hear you now", Jon and
Vangelis. Have a listen at about 4:23 (time can differ a bit since I
listen to a mp3. Wow, even in mp3-quality you can hear it ;-) )
--- In
yamahacs80@yahoogroups.com, "dark_november2000" <jhaible@...>
wrote:
>
> Hi Scott and Group,
>
> another thought about the double trigger issue.
>
> If you prolong the trigger with a one-shot (what you described),
this
> should be a cure for _wrong_ notes to cause dissonance: you
ascertain
> that the trigger is long enough for the CV to reach its final
(right)
> value. But you'd still have two voices sounding instead of one
> (especially with a long release time or held damper pedal).
>
> An alternative could be to mask the first part of the trigger, i.e
> only to allow it to reach the keyboard scanner electronics if it's
at
> least of some minimal length. (As long as it takes for the CV to be
> completely sampled.)
> But this also means that ∗every∗ trigger would be delayed by that
> amount of time, which, depending on how much time we actually need,
> may affect the resposiveness of your playing. (Delays are always
bad
> for a good timing.)
>
> So here is another idea:
>
> For a trigger that's shorter than a certain time (let's say 50ms),
> the release time of the respective voice is decreased and/or the
> sustain pedal function is disabled for this voice. As everything is
> under voltage control inside a CS-80, this should be easy from an
> electronic point of view. (Don't know about the mechanical aspect,
> number of extra wires etc.)
>
> Example:
>
> You have a voice with a release time of 2 seconds.
>
> Now a bouncing key passes a trigger of 5ms to voice #1,
> and a trigger of 200ms (as long as you hold the key) to voice #2.
>
> Asuming voice #1 cannot build up its correct pitch CV in 5ms, you'd
> have a wrong note sustaining for 2 seconds without a correction.
>
> Applying my idea (if it works, that is), both voice #1 and voice #2
> would have their Release time temorarily reduced to "zero" (a
couple
> of milliseconds) - but only for a tim espan of, say, 50ms after the
> trigger.
> To voice #2 this would make no difference: as long as the key is
> held, the zeroed release parameter doesn't matter (we're not in th
> erelease phase yet!).
> To voice #1 (the wrong one), it makes _all_ the difference: as soon
> as the glitch (double trigger event) is over, the voice goes off
> with "zero" release time (just a ghost note, where pitch perception
> isn't really possible anyway).
>
> What do you think?
> Could this work?
>
> Best Regards,
>
> JH.
>
>
>
>
> > I tinkered with a corrective circuit, but am unsure if the
> correction
> > is 'better' (essentially a one-shot with a time of 50ms or so for
> each
> > trigger switch--16 in all).
> >
> > Scott
>