TB Sequencer Replication by Software (was another @#!! TB thr ead)

[List, Christopher]

On Wednesday, April 29, 1998 2:14 PM, Sean Costello
[SMTP:costello at costello.seanet.com] wrote:
> At 10:30 AM 4/29/98 +0100, you wrote:
> Now, this seems like behavior that can be duplicated in software.  I
haven't
> worked with MIDI that much, but it seems that if portamento can be
triggered
> at will, then a sequencer can be programmed to reproduce this
behavior.
> Maybe someone can come out with a shareware 303 sequencer (like the
> "Seq-303" sequencer I have heard about - isn't the author of that on
this
> list? I forgot who wrote it) that can automatically produce this
behavior.
> Or maybe I should get off my butt and do it myself.  I hope to take
some C++
> programming this summer, so maybe I can come up with some workable
thingys.
> 
> Sean Costello

Heh. I long ago thought this through and thought it would be a cool
addition to my Virtual-909 software (the latest revisions are pretty
cool, but not ready for public consumption, and progress has been slow
due in large part to my debilitating addiction to Age of Empires). 

The tricky part to doing this (and the only reason I didn't add it - as
it does seem simple) is that the software needs to know how the
instrument that's playing the track interprets "pitch bend"
instructions, as different instruments can be programmed to pitchbend
different amounts for the same MIDI pitchbend message. Also, because of
this, your range of sliding is limited, especially if the instrument can
only pitch bend +/- 2 semitones! Note that "smart" software could get 4
semitone slides out of this. It would send a note-on in the middle of
the of the two slid notes, at the same time as a 100% pitchbend - for
the first note, then sliding down to -100% pitch bend for the second
note...

Another minor difficulty is that most of us who've written
"electro-centric" sequencing tools have done so with very granular
timebases - because that's what's used in a lot of minimal dance music
(V909 is currently 48ppq, though there's no reason it can't be 96 or 120
or whatever - it's just a constant). This clocking makes smooth slides
trickier, but definitely not impossible (the slide doesn't even have to
be considered part of the "clocked" sequence, and could run under it's
own timer).

I seem to remember noticing that a 303 slide lasts for the life of a
note - not just for the first part of the note - meaning that the slide
is stretched for longer notes, but I forget. Maybe it was just my
imagination - or the old discharging cap half-life thing made it seem
that way. Either way is just as easy in software - you know what note
you're going to and how long it'll take to get there...

- CList