Buchla (ish) modules!

Would love to see 'source of uncertainty' and LPG modules!

Maybe SoU would be possible to make?



  • I believe this is roughly an LPG. Note that the EnvAD affects both the filter and the amp. The filter is LP6, so only first-order (an RC circuit). The Osc and Scope are just for testing.

  • edited December 2022

    try this for starters

    an LPG is my wishlist, too.

  • edited December 2022

    (i have no idea how to replicate the stored random stuff from source of uncertainty )

    but it also offers noise and s/H ...

    I think what happens

    it takes a sample from random noise and stores the result in buffers (bbds)

    and then does some hopping over in the buffers read out ? no?

  • Thanks guys I knew that someone smart would have some drambo workarounds! Thank you! Still fiending for those vactrols though! :D

  • edited December 2022

    hm, LPG is a very musical wish burned into one module

    VCA+Filter combo mode

    if signal is louder signal is brighter if its low in volume its a little muffled ...

    (this imitates what "real" instruments do)

    on every synth in the world you have to set this up yourself ( I wonder why? its so logical to do that ...)

    with an LPG you have to do nothing

    you just plug it in turn it up or down with something - and it does its thing

    its not really meant to be a filter in the classic sense, I guess

    I think of it as "musical vca" (the somehow "slow" reaction from vactrols adds to this)

    but there is one disadvantage : usually lpgs dont close totally. (sure dont want that, a vca that doesn't shut up ...)

  • edited December 2022

    Talking 'west coast', I'd really love a function generator: control over rise and fall time, rise and fall shape, CV input to engage cycling ('looping') or one shot mode, and a bi/unipolar output option. Integrated EOR, EOF gate outputs would be the icing on the cake (end of rise, end of fall). Something drawing from the likes of Make Noise Maths or Frap Tools Falistri

    It's a bit tricky to roll your own in drambo at the moment (or it was last i tried several months ago). The current envelopes don't quite have the time range or control over stage shape.

  • edited December 2022

    close but no cigar

    u can trick around by switching from loop to one shot via morph or scene morph ...

    "end of this envelope part" is indeed interesting

  • @aleyas Sounds much like an extended Env AD Mod.

  • @aleyas I built one month ago, not perfect but it works.

  • edited December 2022


    looking at your patch: oh interesting. I should use slew. 🙏🏻

    thx for pointing this out. :)

  • edited December 2022

    In the new version there is a shift register

    so „source of uncertainty“ should be doable.

    something like this?

  • edited December 2022


    i don’t have the module I can’t look what it’s output looks like

  • edited December 2022

    I need to look closer into how exactly the flip flops are used in that module

    but it’s something like this

    im not sure if it’s jumping randomly over the Flipflops

    or if there is more logic involved ?

    i guess there is …

    more some other time

    have fun

  • I think source of uncertainty had some unique distribution / probability curves. Something like biasing random output to low/mid/high values. So it's more than just pure random. I tried to build something like it a few months ago, but gave up with the distribution curves 😅

    Would be curious what solutions there are for this!

  • edited December 2022

    What do you mean with distribution?

    more or less clicks ?

    playing with the filter does that

    If I turn tone down it gos brrrrr the more I turn it up, is more sporadic click click silence silence silence click

    i need to read up a little and what he is doing with the flip flops , i bet there is some trick with read this output read that output involved

  • edited December 2022

    The distribution stuff is just some other noise spectra than white

  • Aha

    Correlation" is basically a cross fader between the noise-synced triangle wave and the output of the sample and hold. At zero, the output changes randomly from one value to the next based on the external clock input with no correlation to the previous voltage level. At the max setting, the voltage doesn’t change at all, as it is just sampling itself. In the middle, the output is influenced by its previous states, with the Correlation position effectively defining how far the new random value can deviate from the old one—a simple, but profound way of implementing control of just how random the output is

  • edited December 2022


    that sounds like much more Chaos

    Hah 😁😎

    It’s this. (Without the crossfade)

  • edited December 2022

    🤔 so far not bad

    I need help with the math and the logic now for the stored random thing

    how to do the 2^n. Thingy ?

    I always fail at This 🤷🏻‍♂️

  • edited December 2022

    It’s some math with the outputs of the shift Register

    i suck at math, I can’t do the simplest things

    logic is fine, but math me no clue.


  • Ach who cares,

    instead of 2n I use axb

    we are after random things. 😃

  • Maybe I can help with 2^n. What does it actually do? What are the inputs and expected outputs?

  • edited December 2022

    the inputs are complex random clicks and the expected output is ähh lets say more of that but altered

    im reading something with much bigger internal blah blah ...

    English isn't my native language and I understand shit about math

    this is the math/logic part I dont understand : what does this say?

    "if n=1, then the n+1 output can produce two different values (as 1+1=2), and the 2^n also can produce those same two values (as 2^1=2) ... If n=2 then the n+1 output would choose between three values; the 2^n output, however, would choose between four values (2+1=3, but 2^2=4). With 3 selected, the n+1 is 4 and the 2^n is 8, etc."

  • Ok. That sounds pretty clear. A couple of questions:

    How high is "n" likely to be? 2, 5, 20, 100?

    How are you selecting the output from the number of available values for the "n+1" case? Is it a random draw from 0..N+1? Presumably you'll use the same method for 2^n. Do you actually need the value of 2^n to make the selection?

  • edited December 2022

    i have 8 hm different streams of random numbers that fall out of door 1,2,3,4,5,6,7,8

    i dont understand what the math says

    do I need to calculate blah plus 1 and compare it to the other numbers or something ? or compare numbers and then do if it then ?

    it must be something very simple but I dont get it

    currently I do door 3 x door 5 (this is interesting but not what this says it should do)

    damned I always went to the mens room smoking when I had math classes. I should have payed attention. 🤦🏻‍♂️

  • edited December 2022

    2^n used with a shift register usually means that each output has a different weighting: 1,2,4,8,16... so if you multiply each oitput signal with the respective number and add them all then you have a shifted binary to decimal converter.

  • I believe you're just meant to select one of the doors. So, for n=3, n+1 selects one of doors 1..4, 2^n selects one of doors 1..8.

  • @lala Could you maybe post the document you're using, or a link to it? Maybe if we can see the context, we might understand what it means.

Sign In or Register to comment.