ouch, thank you… made a very silly mistake yesterday… renamed the project root folder, then forgot to re-add all the files to git.
all the sources are now here: https://github.com/jpenca/elektron-ctl/tree/master/elektron-ctl
1 Like
the password cracking sounds interesting, and sounds like it defo makes sense to use the olde chains for that 
haven’t really done a lot with the Markov stuff yet… so far, used it only for rhythm & pitch. and in the cheesiest way:
J.S.Bach cello suites as source material.
but it can handle any objects which have an “isEqual:” method as input… e.g. NSString, NSNumber, or your own custom classes… downside is, it’s a bit slow for long sequences. but potentially useful for fun stuff.
that would be a seriously cool project! type in a string and a p-locked talking pattern is created?
Thanks for all this work void. Huge possibilities.
I wonder if anyone could port it over to being a python library? Then it could be non OS specific - and perhaps there are more people able to use python?
That would allow it be used on a small raspberry pi for instance which would make a nice sidekick for the A4 with a few buttons and knobs.
Or were you envisaging this for writing iphone/ipad tools only? (more attractive I guess for live use if you don’t want to use a laptop)
wha?!
How?!
i’ve been writing this in ObjC because it’s the lang I’m most fluent with…
that pretty much locks it down to OSX/iOS… if I had done it in some other language though it wouldn’t be where it is today, by a long shot… and yeah, iPad is a good target… solid MIDI, powerful touch interface, good GPU, easy app distribution, and it’s popular studio gear…
a port to something which is crossplatform-friendly would be really cool though…
wha?!
How?![/quote]
Like the Monomachine, which uses synthesis to make vowel and consonant sounds alternately. For example, noise to make consonants and formant-like filters for vowels.
Stringing the consonants and vowels together gives an approximation of speech, but it takes a lot of word-specific amplitude- and pitch-changes to make it sound at all realistic.
srsly, that would be rad even if it doesn’t sound realistic…
i guess a first step could be to make a thing which let’s you easily sequence vowels…
cool. thanks!
i’ m definitely going to try work on some stuff to add on/give back. i’m pretty slow and messy/ bad commenter. i’ll try to nice it up. i actually kinda like spaghetti code because then you connect weird stuff. 
anyway, this is pretty huge. if you open source the way you talk to the hardware, (you’ve close enough for me) open sourced / modularized the hardware. i mean, i still keep an os9 machine w/ sound diver around for the stuff it can do programmatically that you wouldn’t want to do via the front panel.
1 Like
yeah the formant/vowel stuff is pretty easy to do even just by ear. i had a bunch of preset sounds made, but i needed to do a factory reset and didn’t care too much about backing them up because 1.1 was right around the corner.
and yeah software that translated the vowels into the right machine settings on each track was what i was going to try 1st.
with distortion and noise plus modulation you can get some of the other sounds. the LFOs can modulate at 2hz per second, so it’s not hi-rez, but it’s not bad!
i was thinking try and whip up a super crappy PCM(/PWM) by varying the amplitude of clicks, like waveform drawing, if that makes sense. i’ve gotten some nice sounds on the machine itself by combining that with changing the LFO depth, but again without doing it programmatically, i’m not crazy enough to try to make anything accurate! a lot of the appeal of using a program is that you can just try out settings quickly until you find what sounds right and save it.
if you look into how they did stuff on old game consoles it explains pretty well. (better than i can) 
then i was going to try and figure out how to translate that into locked parameter settings, which would give you a nice starting place to screw around with it.
hmm, now that i think about it, for some stuff you might be able to get better fidelity by splitting the signal into 4 chunks and assigning part of it to each voice, (inverse multiplexing?) but i’ve only ever worked with a single voice at a time.
i have a feeling that somehow the performance knobs could play into this in an interesting way, but i’m not sure yet…maybe morph between male and female formants…
what’s so cool about having the sysex is that instead of making feature requests we can prototype functionality that would be possible in machine itself. thanks again. 
1 Like
yeah I tried messing with this a few hours ago, manually… locked steps across 2 tracks, with the 2nd track using the 1st as neighbor, and got interesting utterance-like results… the A4 is defo flexible enough for speak&spell kinda things…
would be a ton of work to do this programatically though… will try make some convenience stuff which lets you work with the filters in a frequency domain… iirc the filter cutoff value 0-127 is basically 1:1 to MIDI notes…
////
did you check out the repo? can you build the demo project?
1 Like
void - in your probing of the A4’s midi protocol did you happen to come across a command to clear the FX buffers? E.g. - what happens when you triple click [stop] while a delay or reverb is still resounding? triggering this via midi would be super cool to me. I hope Elektron eventually makes that a trig option on the FX track.
Also, thanks for posting this repo. lots of cool stuff here.
cheers!
hmm nope, haven’t found anything like that.
this quick & dirty hack kind of does the trick: https://gist.github.com/jpenca/9605113
setting FX params such as delay feedback to 0, waiting a little, then restoring the original value…
can be clicky though…
would probably be better to just read the kit, fade the parameters down, wait, then fade them back to the original values, using CCs…
1 Like
dirty indeed - i tried something similar with a max patch and didn’t get what i was looking for. Locking ARPs might get me the stuttering effect i’m thinking of though. I hope buffer resets come with overbridge. the FX track could get so glitchy with that as a trig option.
anyway, off to dig through your A4Arp class!
thanks again!
Hi Void,
Just wanted to bump this and say thank you so much for your generosity. I’ll definitely look at this more when I go to do some more Analog Four SysEx tweaking.
Cheers!
1 Like
Hi, I just found this. Seems wonderful. Also, it seems to be deleted from github. 
Any clues?
I’m interested, too. Please let there be a fork somewhere!
Found this! https://github.com/ggozad/elektron-ctl
Also, made my own fork in case ggozad’s one is removed: https://github.com/zsszatmari/elektron-ctl .
I am not sure it compiles today and don’t have the time now to test, but it’s good to know that it’s there. In case someone really wants to resurrect, I’d gladly maintain. Pull requests are welcome ^^
2 Likes