A lot of this thread has gone towards discussing randomization
What fascinates me more than pure randomization is the constraint of
Bear with me on this:
When even a good drummer plays a two handed hihat rhythm there are very small
timbral variations that occur because the sticks are slightly wieghted differently, the pressure of
each stroke varies just slightly (no matter how many hours we have tried to make it sound as uniform as possible); the part of the cymbal changes just slightly, the pressure of the left foot that keeps the two hi hat cymbals together varies minutely.
All of this cause a slightly percolative feel to the rhythm no matter how uniformly the drummer tries to play.............and yet.........the fact that the drummer tries very hard to play and be heard as consistent
seems to have something to do with the musical result (think of a creative professional drummer playing
Louie Louie as opposed to a beginning beginner playing the same thing).
Both can be interesting but the lion share of listeners probably would prefer the former to the latter.
When we loop (unless we are manipulating the way, say, an Andre La Fosse manipulates his EDP)
we freeze a performance in time so that ever deviation from the intended norm (of rhythmic perception) repeats EXACTLY.
What makes that differ from even the drummer who created it is this exact replication of every nuance of the pattern.
There's no denying it: this can be as boring to listen to as listening to a perfectly quantized drum machine pattern looping over and over.
Assuming the drummer is trying to play as perfectly replicateable as possible, there are tiny inconsistencies built into a live performance.
In a way, this could be thought of as a constraint of randomness. It is, of course, not truly random but it most certainly can be mapped as a random (with certain limited constraints) deviation from the norm (or
the intended perfect performance might be a better way of saying this).
This is where Boid algorhythms come in (if I can be bold enough to even talk about them because I certainly don't have the mathematics skills to even understand how they are generated...........please google Boid algorhythms to see what the experts are saying...........I'm not one by a long shot).
Birds flocking will stay a relative distance from every other bird in the flock. The distances will vary
within a certain tolerance (they probably will never hit each other, nor will they get more than a certain distance away from each other because they are a flock , for god's sake!
So, the tolerance of how far away and how close can be a changeable but nonetheless mappable phenomae.
Now consider when the flock changes direction suddenly:
Simple observation will tell you that the distances (or tightness of the flocking) will widen slightly as the birds change direction in both their furthest distance from each other and their closest distance............it will still stay within a certain constraint however because they are flocking for god's sake.
Now that the flock has resumed flying in a relatively straight line (and that itself has some tolerances and yet you can map with a straight line where they will end up weeks later), their relative distances
'tighten up' and go back to their original status quo.
Why not apply these kinds of algorhythms to filter resonance, cutoff, lfo's....................programmable
contrained random deviations from each parameters beginning setting.
My feeling is that the results would feel more 'organic' (and, yes, Larry Cooperman, this is a terrible and wishy washy term if it weren't for the fact that everyone on this list has a strong feeling for what is meant when it is used).
I've noticed when programming potentially sterile drum patterns in Fruity Loops Pro that if I find
a change in a parameter (volume, panning, resonance, cutoff frequency, shift--timing) that produces
an audible difference that I can back that change off until is barely perceptable.
With every single hi hat pattern I can go in and make these really small random changes to every single note and the result is a more realistic (i.e., sounds like a real drummer drumming) rhythm.
Interestingly enough, one can change ONLY THE HI HAT PATTERNS in a piece and listeners can sometimes be faked into thinking that you used a real drummer to either program the piece (via a midi pad) or that it is actually real drums.
Seemingly, only a small (but distinct) percentage of things can be randomized to create a more
'realistic' drum pattern.
Check out the controls in the super cool free VST plugin SUPATRIGGAH. Each major
parameter in this very simple granular plugin has a control for how random it can get and what the frequency of this randomization application is.
I just imagine a granular real time plugin which combines these constrainable randomization algorhythms.
Or better yet, a hardware stomp box that would do the same thing.
It's up to a far better man or woman than me to actually make the damn thing, but I think that would be really cool.