euclidean rhythm as list

sonic shrubs's icon

Is there any implementation out there which creates euclidean rhythm patterns as precomputed lists ?

Miles Davies's icon

http://www.11olsen.de/code/max-objects/11euclidean

ZeroValue's icon

The list created doesn't follow the godfried one.

Roman Thilenius's icon


most "euclidian" max4live devices out there calculate it wrong.

11OLSEN's icon

Hi ZOLTAN, it's the godfried ones, but they maybe rotated / shifted.

10110110 rotated -2 = 10101101  
 

ZeroValue's icon

Hey @11Olsen
I sent you a message via your website.
Yes this what I was wondering, the result is correct but not it the good "position" need to be rotated
I'm trying to understand why :)

ZeroValue's icon

one other strange issue is when setting 0 to hits and steps still getting a 1.

11OLSEN's icon
ZeroValue's icon

Thank you for sharing this !

Mario Fernando Cardoso's icon

If you add [zl.rev] and [zl.rot 1] at the end of all those "wrong-calculated" euclidean generators, you'll end up with the correct Godfried patterns.
At least with the ones I've been trying.

Mafo.

11OLSEN's icon

This includes a JS and a pure Max version of a true björklund algorithm which is what has been used to generate the sequences from Toissant's paper. It also includes some examples which calculate incorrect sequences and others have rotated output (compared to björklund), at that time i extracted the eucl. generator part from different M4L devices and forum posts and compared the output.

Roman Thilenius's icon

i was always wondering how "important" it is for a musical context do create the patterns "right".


this refers to the question whether beat 1 should also be true or not, but even more relevant is probably that you do not really need to have to know which is the correct starting point when you added a "rotation" option to your generator anyway.*

the only thing one needs to have "right" is the pattern itself.

*)
or doesn´t "rotation" break the paradigm already? well, or maybe the repetition does it already... :) "evenly distributed" does not really match to an infinite number of steps...

oh, and i should add that i rotate through all steps, not only through the onsets. :)

bertrandfraysse's icon

It is a bit important, if you always want an event to mark the beginning of a cycle. (or some offbeats and not the 1)

I made a rotation variation that rotates the parttern, but always starting with an onset, which makes some nice "logical" rotations and as a more steady output.

11OLSEN's icon

It's not about musical context. The realization was just that when you describe these sequences with 2 numbers like Toissant does in his paper e.g. E(5,8), you also need to specify the algorithm that has been used (björklund algorithm is described at the beginning of the paper) or you may get rotated output if another one was used. The second realization was that the björklund algorithm seems to be the most efficient one as Max and JS version.

LDMdesign's icon

I coded it myself into a JSUI object I made for a m4l device, and because i'm stubborn I made sure that it does it 'right' according to that paper. Can share if u want