Rhythmic Randomness vs. Melodic Randomness

["Rainer Straschill" <moinsound@googlemail.com>]

> In the Vari speed thread (fascinating by the way) a lot is
> talked about doing things that generate random rhythms to
> play against.
> [...]
> Why does rhythmic randomness seem far more accepted than
> melodic or harmonic randomness?

Rick,

again, I'm entering that thread somewhat late, so I will in part be
repeating things already said.

It seems to me that in your text, you're mixing up three different
concepts/effects, namely:
 a) randomness (i.e. using some random/pseudo-random effect, like a
random number generator in a computer to affect pitch/rhythm in some
way)
 b) accepting non-calculated effects on one domain when consciusly
affecting the other (the domains being time and frequency)
 c) doing things in an unreflected way.

I have the impression that you're mainly talking about b) and c).

In a), I can do that in a way that affects pitch(frequency) or rhythm,
or both at the same time, by assigning a random variable to pitch,
time compression or varispeed. It seems to me that both of these
effects are used, but when used it is easier to work with them both
for musician and audience when you only randomly affect pitch and
leave the rhythm alone. Why is that? This has to do with what you're
indirectly hinting at - that the usual musician and listener is more
educated in the frequency than in the time domain, hence (and this
conclusion of mine is directly opposite to yours) he will more easily
accept frequency randomness than time randomness. In other words: the
listener (musician or audience) can accept the pitch going crazy as
long as the groove keeps going.

When we look at c), I think we all can agree that it is not good if
the musician does things in an unreflected way - no matter in which
domain.

The interesting thing (and also in the context of the varispeed
thread) is b): Let me discuss an extremely simple example to show what
I'm talking about here. Say you have a loop on a (varispeed-enabled)
looper (e.g. Mobius, Repeater, SMM w/Hazari, Digitech RDS/PDS)
consisting of a single note (e.g. an "a") which we will now assume is
"the pulse". Now the musician (who is more consciously at home in the
frequency than in the time domain, as I assumed when discussing a)
above) might decide to generate some melodic line out of this simple
loop by periodically varispeeding it (which can be sequenced easily
e.g. in Mobius by using the rate shift sequence). So he might want a
repeating sequence of "a a a a d d e g" (each pitch single notes) and
programs that. Now this will result in the pulse getting screwed up,
because before you had a very simple rhythmic pulse, now you get
something very odd, which the musician may not have consciously have
thought about when programm the sequence. It becomes a little tricky
to do the calculation here (and although I have studied and worked a
lot with serial composition techniques, I couldn't do that intuitively
- and I challenge all of you, including you, Rick, to quickly (without
doing calculations, to guess how long that new sequence is in the
pulse-quantized time domain), so I'd have to do the math, which is:
before, we had in the pulse domain (one pulse being length "1" now)
simply eight notes equals a length of 8.
Now you have
Four notes at normal playback speed: 4
Two notes a perfect fourth up (ratio 4:3): 8/3
One note a perfect fifth up (ratio 3:2): 3/2
One note a major second down (ratio 8:9): 8/9
Which sums up to 4+8/3+3/2+8/9 coming up to around 9.06. Anyone who
guessed it correctly?

And of course, I'm claiming that the majority of musicians who think
in the time domain (i.e. in rhythmic terms) would not be able to
instantly guess (meaning: when they're improvising) which pitches
they'd be hitting if they decided to go for a quintuple rhythm for
three of the pulses or something.

So , I am of the opinion that when working in a spontaneous/improvised
fashion, most musicians cannot consciously think and compose in both
time and frequency domain at the same time. But (and here, I'm again
challenging your statement, Rick) I don't believe that unreflected
rhythmic alteration/screwup is more accepted than unreflected melodic
alteration/screwup. Much more, most musicians think more firmly in the
frequency domain, and thus, the domain which gets screwed up is time
(although this may be the other way 'round for a drummer like you
working with varispeeding tuned instruments).

            Rainer

(there is a great counterexample to that "unreflected use", but I
can't right now remember what the track is called or who did it,
although I assume it was Amon Tobin. It's some kind of big-beat tune
where there's a melodic loop which for some tournaround gets
varispeeded first down (resulting in a dragging rhythm) and then up
(rushing back up) to come out perfectly on the tonic on the first beat
of the next chorus, beautifully set against a super-straight,
relatively simplistic programmed beat. Anyone knows which tune I'm
talking about?).