< Back to IRCAM Forum

Creating embedded tuplets algorithmically

OpenMusic
#1

Hi,
I have a question regarding embedded tuplets. (See attachment.)

I have a sequence of proportions (1 2 2 1) which I am using to create a rhythm tree using pulsemaker. How would I go about getting OM to take each one of these proportions and convert it into a further series of proportions (eg. (1 3 1)) to create an embedded tuplet. The example on the right converts the second proportion to an embedded tuplet manually. However, I would like to do this algorithmically, where the system would make two decisions:

  1. Convert to second layer tuplet - yes or no
  2. If yes, convert to tuplet using proportions selected from a list of possible proportions.

I hope I have described my process clearly. Any pointers would be much appreciated.

Thanks,
Peter

#2

Apologies - new users can’t upload attachments…

#3

Dear Peter,

First very sorry for this silly and absurd limitation concerning the attachments.
I will assume that i nearly understood your question.
So here it is:

We will use subst-rhythm function (functions->score->trees) for this.
Starting with a one measure voice, and a loop which will be our algorithmical “manager” :slight_smile:

Screenshot_2021-01-13_16-25-25
Screenshot_2021-01-13_16-25-251366×727 39.3 KB

The loop will prepare our embedded proportions on each expressed note of our voice:

t-> insert tuplet
nil-> leave it as it is.

Here is the patch:
Patch 7.omp (16.9 KB)

Hope this helps and that next time you post you will have the privilege to post attachments :wink:

Best
K

1 Like
#4

Hi Karim,
Many thanks for your help - your example and explanation were most helpful.

You did indeed understand the core of my question. I have included your example combined with my patch to explain a slightly different application of the t/nil choice I am aiming for. [EDIT: I am still unable to include attachments]

I would like the omloop, at each point in the original rhythm, to make a choice whether to change that time-point (so t or nil using, for example, a coin toss or other random process) and create an embedded tuplet proportion as opposed to simply running through a static list of three values (t nil t) and stopping there. So the possibility of transforming a note into a tuplet is present at every time-point in the list.

Does that make sense?

Also, does it add any value to the process to use pulsemaker as I do in the example?

Thanks,
Peter

#5

Maybe this will do ?
By changing the loop computation:

Screenshot_2021-01-13_17-41-47
Screenshot_2021-01-13_17-41-47550×797 19.4 KB

Here is the patch:
embeded_tulet_algo.omp (37.7 KB)

Best
K

#6

That is exactly what I was looking for. Many thanks indeed!
It may take me the remainder of the year to take your patch apart to learn how it is put together!

#7

Hi Peter,

Here’s another way, totally random, and without the power of omloop.

Screenshot_2021-01-14_05-04-28
Screenshot_2021-01-14_05-04-28518×720 19 KB

best,
rc

1 Like
#8

Many thanks for that suggestion.

#9

Sorry to trouble you with another question but could you explain what om-random is doing here? It chooses a number between 1 and 4. How is that choice then used?
Many thanks,
Peter

#10

Dear Peter,

If i might explain:
It’s permut-random that will randomize the positions of the output of arithm-ser which is 0 to n-number of pulses minus 1. Which means it could be any arithmetical series but shiffled. then the om-random will pick n-first numbers.
In brief, to understand a lisp patch/code it is good to start looking from output to source (ie in om patch start looking from subst-rythm and upwards.)

I hope my explanation is not shuffled :-))

Best
K

#11

Thanks for the explanation!

#12

Hi, sorry for the late reply and thanks K for the explanation!
Indeed, the om-random ‘decides’ the quantity of rhythms to substitute.
best,
rc