MIDI looping

[buzzard@world.std.com (Sean T Barrett)]

Wow.  I just found this list, and it's amazing to discover
not only so many people involved in looping, but even several
effects processors dedicated to it.

Just a year ago, I mentioned to a friend, "I wish there
was some kind of digital delay where you could record a loop,
then set it aside and record another one, and then switch back
to the first, and then overdub, but why would anyone bother
manufacturing something nobody else would want?"

Foolish me.

Ok, anyway.  I have three big questions, but I'll give
them time to breathe, so let's start with the first only.

MIDI looping.

I've checked all of the '97 archives (although there's
no June, so maybe you've all talked it to death last month),
and the only mention of MIDI looping I saw was discussion
of the Cyclone (which sounds a bit more like sequencing/
arpeggiation tech, although I can see how they become similar),
and one person mentioned he was writing custom software.

So what's the deal?  Does nobody on this list do MIDI looping?
Is there simply no good technology to carry it out?  Are
people talking about it on some synth mailing list?  Or is
it just that audio looping is so much cooler because guitars
(or accordians or trombones or voice) can do much cooler things,
and MIDI isn't expressive enough?  Or vice versa, that synths
are already powerful enough instruments that they don't need
the crutch of delay technology before they become interesting
solo performance instruments.  (Hey, I'm a guitarist myself,
I'm just getting the theories on the table, not advocating
them.)  Or is the list full of MIDI loopers who are just
keeping quiet?

I first experimented with MIDI looping in '87 or so (to
answer that age question, I was 20 at the time).  I took
my friend's Atari ST, wrote a BASIC program to do MIDI
looping, mapped program changes from the input to output
channel routing (so from one synth you could loop multiple
different timbres), and my keyboardist friend used it to
create backing textures for my pointless guitar solos.

Then, because I was a guitarist, I forgot all about it.
Then, as I said, I got the idea for this cool looping
technology, but I figured it didn't exist.  So then I
was looking for other ways to expand my instrumentality,
so I'm getting a guitar synth.  And then I figured, hey,
MIDI looping should be a lot simpler than digital looping,
maybe I could do that.  A search on the web, and here I
am--nobody anywhere seems to be talking about MIDI looping.
(I'm not doing it myself--but I want to be.)

Oh, duh, a quick definition in case anyone can't guess
(or if MIDI loop is a common term for something else):
a MIDI looper is like an audio looper.  You play in a
sound source througha a MIDI in, and on the MIDI-out
it plays the notes of the loop.  Basically it just passes
through the notes you play, and then plays them again after
a delay, etc.

Here are some of the obvious issues I've thought of for
MIDI looping:
  con: no effects in the feedback path
   but: most people don't use their loopers that way anyway
  pro: actually, you can pitch shift and bounce notes between several
       instruments during feedback
  pro: easy to have multiple different speed loops (in terms of internal
       implementation--user interface still a nightmare); even do odd 
things
       like every pitch gets its own loop length
  con: requires MIDI input
  con: MIDI guitar with pitch bends requires one MIDI channel per note,
       which will get eaten up really quickly when you layer a loop
  con: another MIDI delay in your signal path
   but: you can use your performance synth (e.g. guitar synth or keyboard)
       to provide the initial tone, and then the extra MIDI delay can
       be compensated for by reducing the first iteration's delay time
   but but: now you need another sounds source with very similar sounds
       to your initial sounds
  pro: requires much less RAM; infinite UNDO is plausible
  con: probably harder to create the software for
  pro: probably requires much less CPU crunching power
  pro: you can "record" your performance into a sequencer,
       storing the notes you played & MIDI patch changes or
       such that changed the looper's performance--then just
       play the sequence out into the MIDI looper to repeat it.
       Now you can edit your performance.
  con: drops out notes if the layers get too thick
   but: get more sound sources to avoid this (and possibly multiple
       MIDI outs on the looper to get more channels with distinct
       pitch bends).  Also, audio looper must distort or clamp
       or compress if the total audio volume gets too thick (different
       but similar sort of problem)
  
Well, I could go on and on (well, I guess I already have), but
I'm interested to hear some comments before I go too far over
the top with it.

(I'm pretty sure there are some existing MIDI "multiplexers" with
some kind of MIDI delay features, but it seems the dedicated digital
audio loopers have obvious performance features for doing all sorts
of things a pure "delay" won't have, and I doubt such multiplexers
implement a delay which deals correctly with pitch-bent notes, thus
making not too useful.  But I'd be just as happy to be proven wrong,
as the Echoplex and JamMan have already done in the digitial audio
dimension.)

Sean Barrett
computer game programming: 
http://www.geocities.com/SiliconValley/Vista/7788