Stacking Static Parts and 100% Feedback in live looping

[Rainer Straschill <moinsound@googlemail.com>]

Rick,

I'm curious: did I write this to you in private correspondance (also
about the low-cost looper), or to this list, or did I only THINK about
writing that for me, currently variable feedback (of the EDP/delay
type) and unrounded multiply are the most important loop-modification
options, and for that reason, the DD20 (in delay mode) would be my
preferred low-cost looper? ;). However, I digress...

Of  course, as you say, different approaches aren't worse or better
than others, only some lend themselves better to specific musical
goals than others.

What interested me in your post was, as you said, "but I keep finding
more and more things to do just by using starkly naked loops that
don't morph over time but that are,  through arrangement,
recontextualized because of the parts that end up on top of them."

I never consciously worked in that way, although I have found that
some of the stuff you do (and also, more by chance), uses that effect
(and I hope I correctly understand what you mean here): for example
laying down a theme or short motif which seems to start on the beat,
but by overdubbing a drum groove (or other unambigous reference), it
suddenly appears offbeat.

Of course, a lot of these things can be used in a much more flexible
fashion by using a multitrack looper like Mobius or the Looperlative.
Per already mentioned that he uses fades on individual tracks instead
of varying feedback. Using the layers in the example before on tracks
with different lengths will recontextualize them on each pass. This,
however, both depends on looper implementations which are still, more
complex than the "average looper", at least in the hardware domain.

...and with that, we're at the point where it gets interesting for me.
When working with multiple loops, EDP-style feedback and time-variant
processing, things will happen that while being strictly
deterministic, are of a nonlinear fashion and thus chaotic.
Loops of different lengths make a great basis for ever-changing
rhythms (I tried something on a very dilletantic level here
http://vimeo.com/4565667). Now set those loops with different lengths
to different feedback values - and you got a changing volume
relationship, which makes for some beautiful fadeouts to play against.

The beautiful thing is that this works completely without user
interaction from the point the feedback is set - thus freeing my to
walk around and play while the loops change and slowly vanish (or
slowly get replaced by something I choose to add).

But I'd like to ask you as an intercessor of 100% feedback: how do you
"remove density" from an existing loop arrangement? Ways I can think
of:

1. mute/turn down individual tracks (requires multiple tracks)
2. reduce feedback (requires adjustable feedback)
3. "start over" with a new loop altogether (tricky with some user
interfaces as with the SMM w/Hazari, as Andy Butler pointed out)
4. unrounded multiply/divide to cut to a part of the loop with low
density (requires unrounded multiply/divide)
5. undo (requires undo functionality)

Of course, using a Looperlative, you're in the nice situations that
all of those are available to you. Which ones do you use - or don't
you use any of these and have your arrangements always climax
Ravel-style (or do you use another approach I didn't think of)?

Best,

             Rainer