MUFF WIGGLER Forum Index
 FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 
WIGGLING 'LITE' IN GUEST MODE

Missing something fundamental about generative/random
MUFF WIGGLER Forum Index -> Modular Synth General Discussion  
Author Missing something fundamental about generative/random
akaye
Hello! New to modular and the site, but so far I've learned an incredible amount of knowledge from you all. Thank you.

I've put together a modest first rack so far: u0_C, Ears, Rings, uClouds Monsoon, Doepfer A-183-3 (for a cheap way of bringing in external line signals from my computer to be processed -- will eventually go expert sleepers route im sure), and an intellijel mixup for mix/outs.

So, my main goal currently is diving into generative and random/chance type patches. Though I'm missing something in my head about how the ones I like and hear so much are made.

I'm taking a clock signal from my keystep into the trigger input of O_c and using the quantermain quantized turing machine..... which is working out great so far. My problem is, the rhythm of the random notes is so completely steady, no variation. I was thinking/hoping that turing machines not only randomize pitch but also timing/rhythm?

The patches that I hear online and seem to like so much obviously have random pitch (pitch quantized, per my taste), but they also have random rhythms.... moments of silence and rest that will spontaneously burst into a flurry of notes at odd rhythms. That's what I want!

So.... am i missing something theoretically simple here??

Thank you!
the bad producer
The easiest way to do this is using a clock which has voltage control of rate, which you control with the random voltage. VC LFO, Function Generator etc.
akaye
Okay yeah, that makes sense to me, thanks. Does anyone familiar with O_C know if there's a way to do this within the module? Or somehow with the hemisphere firmware?
Abyssinianloop
Make sure you have lfsr probability set to some value other than zero so the turing machine will randomize.

Edit: Sorry, nevermind. I see that you had already achieved that and were asking about something different.

Although, you should be experiencing some notes on and some off in ever changing sequences. Sometimes a bunch of notes in a row, sometimes long spaces between notes.
mgscheue
I don't think there's a way to get O_c to randomize timing other than feeding it a clock from a randomized source. If you use the internal trigger on Rings, rather than feeding the Strum input, it'll only play a note when the note changes, so that gives you a bit of randomness. Mutable Instruments Branches is a very nice way to add gate randomness. I also have the Ladik R-110, which is handy. Both are relatively inexpensive.
BenA718
+1 to Branches. It's (or its equivalent) a must-have.
Umcorps
A simple way to thin out regular clock signals is to use a comparator. Put your clock into the comparator along with noise source and set the threshold/hysteresis to taste. Then use that to clock your shift register.

Doepfer A 167 is probably the cheapest option. If you lack space, Disting also has a good one (and a tonne of other useful stuff)

Here's two Distings doing just that. The one on the right is in comparator mode thinning a regular clock, the one on the left is in quantised shift register (Turing Machine) mode, clocked by its friend.

CLee
If you’re setting up a non rhythmic patch, one thing I like to do is DUSG as clock and patch pitch CV from whatever is generating the random pitch into the DUSG RATE CV. Set it so higher pitches have a shorter duration (faster clock). It has a natural feel because IRL higher pitches tend to decay faster.
akaye
Thanks for all the replies! Definitely helping out a lot.

CLee wrote:
If you’re setting up a non rhythmic patch, one thing I like to do is DUSG as clock and patch pitch CV from whatever is generating the random pitch into the DUSG RATE CV. Set it so higher pitches have a shorter duration (faster clock). It has a natural feel because IRL higher pitches tend to decay faster.


^^ this seems like it's exactly what I'm after. Is this something maths could be used for? Or would it have to be a true DUSG like one from Befaco or Serge?
CLee
akaye wrote:
this seems like it's exactly what I'm after. Is this something maths could be used for? Or would it have to be a true DUSG like one from Befaco or Serge?


Any voltage controlled LFO/VCO would work. The patch isn’t using anything unique to the DUSG
mt3
Umcorps wrote:
A simple way to thin out regular clock signals is to use a comparator. Put your clock into the comparator along with noise source and set the threshold/hysteresis to taste. Then use that to clock your shift register.

Doepfer A 167 is probably the cheapest option. If you lack space, Disting also has a good one (and a tonne of other useful stuff)


+1 for a solution that involves a comparator.
Om
electricanada
Burst generators might also interest you.

But keep in mind that randomness is only one small subset of generative/algorithmic musical methods.
EATyourGUITAR
clock + noise into comparator is not a good way of describing it. I think this would just give you extra pulses between the periodic pulses. it needs to be something that is not a square or pulse. ideally it would be a saw lfo but a triangle or sine lfo will do the thing. saw + noise with the same range +/- 5v on the noise and the LFO. then if you want the deluxe adjustable version you can run either the noise or the saw lfo through an adjustable offset. you can automate more or less pulses with a very very slow ramp from a maths, dusg, 281 something like that. you run the noise + 5 minute ramp into a dc mixer. run the dc mixer + saw lfo into the comparator. there are modules that have this stuff all premade into one module.
Umcorps
EATyourGUITAR wrote:
clock + noise into comparator is not a good way of describing it. I think this would just give you extra pulses between the periodic pulses.


Watch the video





The clock is your maximum regular voltage level. The noise is a fluctuating voltage. The comparator level defines the voltage at which the noise is considered to be "equal" to the clock.

So, if your clock is 5v when high and the comparator level is set to 3v you only get a logical TRUE when the noise level is >3v AND the clock is high.

So you can't increase the clock frequency this way, but you can absolutely thin it.
EATyourGUITAR
His clock could be 0v to 11v. I see what you are suggesting. It could work for some people and not for others. If your clock is +/-5v and your noise is +/-4.5v peak then you will just get the clock at the output of the comparator. We should be clear when giving instructions to people who are trying to reproduce the results we are getting. There are many ways to patch a modular. This is not the only way. This could get up to some super massive patch if someone wanted to.
okiikahuna
The comparator method is my favorite. One advantage is that the rhythm rather than being totally random can be in time with a basic underlying pulse. Like good musicians usually play. Also, you can mix the random voltage with the another voltage to give you voltage controlled rhythmic density. /using an envelope with a very slow attack, the rhythmic dansity can gradually increase through can gradually increase over time , and So it can go somewhere rather than just meander, if that's what you want. Nothing wrong with meandering, of course, but you can have control over long term structure if you want.
adnauseam
akaye wrote:
So, my main goal currently is diving into generative and random/chance type patches. Though I'm missing something in my head about how the ones I like and hear so much are made.

I'm taking a clock signal from my keystep into the trigger input of O_c and using the quantermain quantized turing machine..... which is working out great so far. My problem is, the rhythm of the random notes is so completely steady, no variation. I was thinking/hoping that turing machines not only randomize pitch but also timing/rhythm?

The patches that I hear online and seem to like so much obviously have random pitch (pitch quantized, per my taste), but they also have random rhythms.... moments of silence and rest that will spontaneously burst into a flurry of notes at odd rhythms. That's what I want!

So.... am i missing something theoretically simple here??

Thank you!


The Turing machine modules you see online may be using the Pulses expander. The turing machine, on the inside, is actually generating a binary number - a series of 1's and 0's that are turned into an analog value (pitch or whatever) at the output. The pulses expander gives you access to each bit as it's shifted around. XOXOXXOX etc. One application is that you could use these many outputs as they flow through to create a series of "random" gate signals for envelopes or percussive envelopes.
PietroC
Part 1
With Maths Using the Envelope Follower ( Single Voice Patch )

- You can send one signal for regular tempo to Trig In
- Another signal that is tempo divided by 3 or 6 to the cycle and make it that cycle makes a burst when cycling Carefully Adjusting Rise / Fall and the EXP / LIN Shape to get a Triplet or Quintuplet

Also Add a Random Stepped Cv to Rise and/or Fall for more variation ( This Works Amazing with Noise also to get a Snare type Sounds )

Part 2
To Make this even More Random
Use Channel 1 EOR To Delay the trig and to trigger Channel 4 ( Patching like the first part with Cycling )

Only Thing Any Tempo Alteration will make everything Misaligned
3hands
I’m going to give this a go today! Thanks!
Parnelli
PietroC wrote:
Part 1
With Maths Using the Envelope Follower ( Single Voice Patch )

- You can send one signal for regular tempo to Trig In
- Another signal that is tempo divided by 3 or 6 to the cycle and make it that cycle makes a burst when cycling Carefully Adjusting Rise / Fall and the EXP / LIN Shape to get a Triplet or Quintuplet

Also Add a Random Stepped Cv to Rise and/or Fall for more variation ( This Works Amazing with Noise also to get a Snare type Sounds )

Part 2
To Make this even More Random
Use Channel 1 EOR To Delay the trig and to trigger Channel 4 ( Patching like the first part with Cycling )

Only Thing Any Tempo Alteration will make everything Misaligned


I've used the Part 2 version to place a snare drum at an odd place in time before; if you fiddle about enough with the rise and fall of the two channels you can put it just about anywhere you want.

And yes, you cannot change anything once it's set up or it will move on you.
MUFF WIGGLER Forum Index -> Modular Synth General Discussion  
Page 1 of 1
Powered by phpBB © phpBB Group