Analogue-sequencer

Hi Colin

Can you tell me what actually happens when I apply a % of randomisation to a
value in an aux? I understand that it affects the likelihood of the
programmed value being produced but does it in any way affect the likelihood
of any other value being produced? Or put another way, is the randomly
created value any likelier to be closer to the orginal programmed value or
is there an absolutely equal
likelihood of any value within the available range being produced? The
mathmatical and musical effects of these two possibilties strikes me as
being very different

I'm getting very interested in some of Buchla's ideas and in seeing how I
can using the P3 as a random pattern creation source -  in a way that
randomness can be controlled, shaped and modulated. What I would really like
to have more random modes available at both the track and aux levels.  It is
neat having instant
randomness and the random pattern playback modes are great, and it is great
to be able to control the amount of random variation for each step in the
Auxes. But I would like more control over the how the randomness is produced
and applied.I did make some vague requests in this direction last year but I
am clearer
about what I would like to be able to do: some of these ideas would be more
appropraite for the track level and some would make sense in the Auxes.

I'd like, in the Auxes for example, to be able to use a controllable
"smoothing" algorhythm so that the random number generation can be softened
and biased towards the programmed value. e.g. set at 0% it would just
produce the same value over and over, set at
100% it would be truly random, but set at 50% it would show a 50% bias
towards values closer to programmed step value - as if you could change the
strenghth of gravitational force of the orginal value. The controls for the
mode could be available in some form in the auxes, as both sources and
destinations for further modulation. I think some quite subtle evolving
effects could be created in this way which would be difficult to create
otherwise.

Similar but different (I think) would be a Gaussian distribution which would
bias numbers towards the centre of the available range, I think voltage
controlled random generators tended to have this chracteristic? It would be
nice to have it on the P3 too. This should also have a % control and also a
polarity control, biasing numbers towards
to the top or to the bottom of the available range.

In track mode I could also imagine a random mode with similar controls that
compared the new random value to the last value created. There could also be
some musicially interesting way of biasing whether or not a random number is
likely to be above or below the previous value, but I haven't quite
formulated that yet - I am sure some mathematician will have a name for it.
I think both these ideas
would be good for creating "melodic", more linear streams of values, an
interesting alternative to the more typically "random" sets of values
created when each new value is a completely new roll of the dice.

A way of temporarily overriding programmed values with random modes (and the
existing sculpt mode) would also be an interesting feature for live
playing: when a button is pressed the programmed values of the selected
track are temporarily replaced in real time, when it is released the
preprogrammed values are reverted to.

Another thing I would appreciate is basic templates for LFO shapes and a
range of different arpeggios and interval patterns, so for example if I want
an aux to function like a clocked LFO I could just dial up whatever wave
shape I wanted to use rather than painstakingly programming the values step
by step. Using these I
could, for example, dial up a template of perfectly divided values for all
16 steps. I could then, say, use a tempo synced sine wave pattern to
modulate the distribution biases
within the randomisation being applied to all the values in a pattern.

Phew... my brain hurts.

Don't want much, do I :)

Richard