Architect: Does anyone have a * macro/script

[cturner]

If there’s a problem, then running the cable from the output of [==] should do the trick.

[AZZIN]

If there’s a problem, then running the cable from the output of [==] should do the trick.

Yes, I think this would "disambiguate" the order of execution. But still I'd like a word from Colin on this specific topic..

[ThomasHelzle]

But then you would have the wrong value, wouldn't you?

Cheers,

Tom

[cturner]

But then you would have the wrong value, wouldn't you?

Yes, you’re absolutely correct. I shouldn’t answer these kinds of questions from my phone when I’m out on the street. :-(

[mannymang]

Generative Midi Machine UPDATE:

Added Alberto's a-chaos algorithms:
a-clifford
a-dejong
a-duffing
a-henon
a-hopalong
a-logistic
a-lorenz

You can now choose for each track which algorithms for the notes and velocity. Feedback is welcome!

Special thanks to AlbertoZ for graciously providing the a-chaos algos!
viewtopic.php?p=7299827#p7299827

Special thanks to Charles Turner for graciously providing the logistic algo!
viewtopic.php?p=7284039#p7284039

Enjoy!

[AZZIN]

Wow, looks great!

[mannymang]

Thanks Alberto!

[mannymang]

Generative MIDI Machine UPDATE:

Fixed: Track on/off.
Fixed: Scales drop-down was reporting incorrectly.
Added: Randomize scale & tonic button.
Added: Note duration. Finally got around to doing note duration rather than relying on the Hold module. Hold on/off button is still there to provide some random staccato. With the addition of note duration,
everything is smoother and less choppy. There are some extraneous notes (it seems) getting through sometimes. Will need to track that down.




Enjoy!

[colin@loomer]

That's pretty incredible. Are there any debugging features I could implement that would help you track down these extra notes?

[mannymang]

Hi Colin, I'm pretty sure it's my multiple instances of metronomes where these extra signals are coming from. I need to go through it all with a fine toothed comb. I'll report back what I find. Thank you!

[mannymang]

That's pretty incredible. Are there any debugging features I could implement that would help you track down these extra notes?

I think I figured it out, it's either 2 or more choices:

a: The extra notes are coming from the aether where no amount of debugging will help.

b: When a duration of a note or notes are longer than the arrival of the succeeding note(s), the succeeding just need to be patient and wait their turn before making their debut thus surprising everyone with the unexpected appearance wherein the performance brings down the house.

c:

Happy accident, that can remain a feature, until later...

[cturner]

Here's another installment of "something no one requested": "sieves" in Architect, used to generate attack time points for percussion, but obviously other things as well.

The "sieves" reference is to the Sieves of Eratothenes, which is a method of deriving prime numbers. They are put into the service of music, if not for the first time, certainly most extensively by Iannis Xenakis.

EDIT: If you wanted a general introduction to Xenakian sieves, I'd recommend a paper by my friend Dimitris Exarchos:

https://research.gold.ac.uk/15753/1/11. ... -Jones.pdf

There are a few generalizations of his work in code, but I believe they mostly deal with the generation of harmonic intervals, less so rhythm. I'd postponed fiddling with this for a long while, largely for lack of an inspiring platform on which to explore it. On offer are sketches really, but I imagine that bits could be adapted for service in many ways.

Here's video of Steve Schick performing Psappha, the first forty notes of which are roughly generated in Architect presets:

https://www.youtube.com/watch?v=Yge7GNl5p_k
You can download my files here:

http://vze26m98.net/loomer/psappha-1902 ... 190209.zip

There's a PDF in the ZIP archive that contains my notes, working backwards from a final product:



Sieves construct moduli that can be shifted along the set of integers, and as sets, they can be combined in basic ways: union, intersection and complementation. The above shows the finished forty-note line and the combination of moduli Xenakis used to produce it. The first combination: (8|0 u 8|1 u 8|7) V (5|1 u 5|3) is read as follows:

The union of moduli of 8, shifted by 0, 1 and 7 is shown in "dot form" on the top, the union of moduli of 5, shifted by 1 and 3 shown on the bottom. The intersection of these two 40 point lines are circled in blue.

The PDF works backward into this and into Ellen Rennie Flint's wonderful doctoral dissertation on Psappha. My notes as a whole are probably still rather cryptic, but I'd be happy to provide a copy of Flint's dissertation if you PM me.

I started work in Architect using visual objects, but as you can see from the next image, it really becomes quite large for the simplicity of work involved:



So I stopped after the first combination and moved to Lua. Given that the task is really one of pre-calculation, it's eminently suited for the job. The whole task collapsed into one visual object.

What's valuable about the first sketch is the sieve macro:



which in a simpler combination of moduli and the addition of knobs, could make a useful performance tool.

If you look at the page of moduli above, you'll see that Xenakis made things needlessly complex, such as the 8|3 intersection with the entire groups of 5-moduli. That's really just the 8|3 group; no need to do the intersection. I think the explanation is in his hand-calculation: with revision, it was always expensive to re-calculate. So the simple clean-up we are now able to do via computer just wasn't a part of his working method.

I hope this is of interest. Happy to talk further about it. Enjoy!

[AZZIN]

Veeeery interesting. I Have to dig into it. Thanks for sharing..

[cturner]

Thanks for sharing.

You’re welcome AZZIN! It’s my pleasure.

[cturner]

I haven't heard much talk recently about automation in Architect, but here's a first bit of a larger thing I'll post eventually: the processing required to convert a 14bit value into it's respective 7bit MSB/LSB CC message streams:

DL: http://vze26m98.net/loomer/cc_14bit-190 ... 90211.frag



Not elaborately tested, but seems to work a treat on the u-he synths that support 14bit CCs.