## Musical Idea 21: Psychoshuffle

02.23.2014 § 1 Comment

Let’s add a layer of craziness! Shuffle in its typical form is simply turning a knob to introduce an increasing amount of difference between alternating notes. Per Nørgård has done something funky where binary divisive rhythm goes by the golden section each time. All that I’m calling for here is adding a random +/- to your atomic units of time.

The subtlety would be stipulating that the total length of your music is non-variable. In this case, the easiest way to accomplish this would be to say each random plus must be countered by an equal and opposite minus, and vice versa; since the point of this is vaguely destablizing your entire music, you wouldn’t want to make them neighbors, but randomly place these opposites among the atom units consisting your music. But you could achieve this other ways as well, perhaps even insisting that no two intervals have the same absolute value difference from the original.

The next subtlety is that this effect isn’t necessarily imposed in general across an entire music, but can be chosen to be imposed only within a specific window. For example, you could take a blast of sixty-four sixteenth notes (four bars worth) and say “each of these sixteenth notes can be up to a thirty-secondth note shorter or longer, but by the time sixty-four of them have elapsed, the total duration must add up to what it would have before.” In this case, you can rest assured that you have a clean set of exactly four bars worth of material to integrate into the rest of your music, however, you cannot be sure that at each bar a strike will be aligned. If you wanted to stipulate that, then you would need four of these windows, each saying “…by the time sixteen of them have elapsed, the total duration must add up…”

What’s beautiful about this is how it can lead to this effect of things drifting in and out of coordination with each other — if you have multiple entities with aligned windows but the randomization happening differently within them, things might get really off in the very middle, but then you’ll feel it starting to pull itself back together as it approaches the point where everything will finally hit sameltimeously again.

You can also play with adjusting the level of randomness through the window, such as saying stuff has more of a chance of going to extremes toward the beginning of the window, but less toward the end, so there will be a longer stretch of it approaching that realignment (but you can do any sort of graph to indicate the level of randomness, or the chance of longs or shorts at any time throughout).

The next point to touch on is that you may want to respect the integrity of your atomic unit of time by never allowing one of your +’s to be more than twice the length of your shortest minus. In other words, if you want to preserve the original rhythms unambiguously, you should not allow a +/- greater than or equal to ⅓ of the atomic unit’s length. The above example wouldn’t meet this requirement since a thirty-secondth is ½ of the atomic sixteenth, and thus it would be possible to introduce a situation where it was unclear whether a resultant interval was actually longer than another in the original music or if it was just made that way via psychoshuffle.

But this generally not quite far enough. A related point has to do not with the atomic unit of time but the rhythmic intervals that you actually use in the music. That is, your atomic unit of time might be a thirty-secondth note, but you might not actually use thirty-secondth notes in your music: however, since you do use both sixteenths and dotted sixteenths, the thirty-secondth is their greatest common factor (in short your atomic unit must be the GCF of every original rhythmic interval). If you also wish to prevent any ambiguity on the level of the actual rhythmic intervals you use built from that atomic unit, then you will have to consider the smallest ratio between rhythmic intervals used in your music. In divisive rhythm, start with 2:1 = 2, then if you permitted dotted notes, 3:2 = 1.5, and if you permit triplets 4:3 = 1.333. Then if you permit a whole tied to a quarter then you’ll start having stuff like 5:4 = 1.25. Anyway, if you look at this number, you have to make sure your +/- isn’t too big such that it would ambiguate it. Note that there is a chance here because the extremes have to clump together for it to happen. Here’s the formula: +/- 1/xth of your atomic unit -> (x+1):(x-1); if that’s bigger, it can happen (for instance if you had ¼ be your +/-, your atomic unit could become as big as 5/4 and as little as 3/4ths as long, so if you had a rhythmic interval built out of 5 of them and another built out of 4, and the 5 guy lucked out and got all 3/4th lengths, and the 4 guy lucked out and got all 5/4th lengths, then you’d have your originally longer interval come out only 15 long and your originally shorter interval come out 20, longer now; if you wanted to guarantee the unambiguous preservation the original interval classes, you’d need your +/- to be less than 1/9th the atomic unit’s length).

The next thing to point out here is that these principles can be applied vertically as well, that is, to tuning systems. You could spice up your tuning with a little beating by applying this random alteration to multiple entities in a section, but still requiring that, say, their octaves all perfectly line up, for example. Or for a single entity, each pitch could get different randomizations for the same durational window, and each strike could get different pitch randomizations within the same pitch window, resulting in a perfectly rectangular frame over a reaaalllly busted up grid.

[…] but it involves non-live stretching of intervals applied in a way kind of more like the later Psychoshuffle chapter insofar as you set a window of […]