The long DIVA thread

Official support for: u-he.com
Post Reply New Topic
RELATED
PRODUCTS
Diva

Post

Urs wrote:My current workplace is 15sqm and tumbles under the space occupied by the machines I accumulated
PM me, i think i can help you clear some space!!


:hihi:

PS the two Avid jokes made me smile before breakfast!! (well done you two)


:lol:

on a Real On Topic note!

i think Diva is a excellent idea

U-he already has the sound but lack the ATCU interface (Any Twat Can Use)

Diva would would fill that hole!

Subz

Post

djsubject wrote:U-he already has the sound but lack the ATCU interface (Any Twat Can Use)

Diva would would fill that hole!
Well, call her a sweet-spot-synth ;)

Post

Hmmm... regarding a/b tests... last night I had the idea of writing a little software that takes an audio file of a single note played on a vintage synth as an input. The other input would be a hand-matched preset for the Diva engine.

The software would then analyse oscillator phase and frequency of the audio sample.

Then it would start an iterative process rendering the software synth at the right frequency and phase, and measuring how good the phase inverted signals cancel each other out. With each iteration it would adjust parameters slightly and compare to previous results. This would be a genetic algorithm, with "silence" being the selection criteria (there may be more though)

The best result would show the exact difference between the original and the simulation.

Wonder if that could work and if it could help to convince the naysayers.

Post

Urs wrote:
djsubject wrote:U-he already has the sound but lack the ATCU interface (Any Twat Can Use)

Diva would would fill that hole!
Well, call her a sweet-spot-synth ;)
Oh, so many jokes I could make that would potentially get me banned..

Post

Urs wrote:Hmmm... regarding a/b tests... last night I had the idea of writing a little software that takes an audio file of a single note played on a vintage synth as an input. The other input would be a hand-matched preset for the Diva engine.

The software would then analyse oscillator phase and frequency of the audio sample.

Then it would start an iterative process rendering the software synth at the right frequency and phase, and measuring how good the phase inverted signals cancel each other out. With each iteration it would adjust parameters slightly and compare to previous results. This would be a genetic algorithm, with "silence" being the selection criteria (there may be more though)

The best result would show the exact difference between the original and the simulation.

Wonder if that could work and if it could help to convince the naysayers.
This sounds like a great "cloning" tool :D
Would be surely nice, to have a softsynth, which can produce many facettes of old analog synths, specially if it has such superb quality like ace.

If there would be later many of such "cloned" synths in there, would Diva on the first hear the false name i think, also if Diva is only a shortcut for 4 words, the most users will think diva - opera/singer.
And under the point of the "cloning" why not something like...

....Fantomas :lol:
Image

Post

Urs wrote:...With each iteration it would adjust parameters slightly and compare to previous results. This would be a genetic algorithm, with "silence" being the selection criteria...
Good idea, I had thought about doing something similiar myself at one point, having doine some ai at uni, but don't really have the time/sKilZ/wherewithal etc

I'm actually a Juno 6 user, not sure how much of i difference there is between the 6/60, a lot less than between either and the 106 though I beleive

I'd be glad to hear some a/b's though :)

Post

todd_r wrote:I'm actually a Juno 6 user, not sure how much of i difference there is between the 6/60, a lot less than between either and the 106 though I beleive
Yeah... the voice architecture is the same of course, but I wonder... are the faders still digitised first and then connected to the circuit via sample & holds?

- me pulling up schematic -

Oh. The Juno 6 has... hmmm... quite a few differences there... nothing serious at first sight, but the difference in parameter control might also mean different control ranges.

Post

Urs wrote:...the Juno is very in-your-face during a short attack with filters open..
Those are the sounds I really love from juno
Urs wrote:... quite a few differences there... nothing serious at first sight, but the difference in parameter control might also mean different control ranges...
So the only difference lie in the controls themsleves, the osc, filters, env are the same?


Ooh! ooh! ooh! Do you think you could replicate the Juno's unison sound, it's very a unusual/thick/phasey/muddy, not a nice sound as such, but definatly unique

Post

Well yes, the Juno 6 and 60 seem to share they same (or similar enough) components, but aside from the switches the Juno 60 has 16 sample & hold units. These are set to the parameter values stored in ram until a fader is moved. So the faders are not directly connected to the voice circuits. They are connected to the cpu which samples them with an adc and then converts the values back to analogue.

There's only 1 adc and 1 dac, so they use demultiplexers to iterate through them with a timer. On the output side then the analogue values are stored in the sample & hold units.
todd_r wrote:Ooh! ooh! ooh! Do you think you could replicate the Juno's unison sound, it's very a unusual/thick/phasey/muddy, not a nice sound as such, but definatly unique
Really dumb question... how do I switch the Juno 60 to unison?

:oops:

Post

Turn off, put arepeggio mode in up position, turn Juno on with transpose button held in. I beleive you can then go between normal/unison using the arepgio switch :)

Post

Hi Urs, I don't know if this would be helpful for you or not, but this guy implemented a similar genetic parameter-matching algorithim with the SuperCollider programming language. I think he made the code available to the public, but I haven't looked at it yet and if I did, I'm sure it would be over my head! Here's the link, with a video demonstration:

http://www.earslap.com/news/naturetoolk ... ercollider

Post

todd_r wrote:Turn off, put arepeggio mode in up position, turn Juno on with transpose button held in. I beleive you can then go between normal/unison using the arepgio switch :)
Amazing! But how do I detune the voices?

Is there a similar trick on the alpha Juno 2?

Post

Urs wrote:Amazing! But how do I detune the voices?
I dont think there is any way unfortunatly, not the most usefull/versatile of unisons i know :lol: apparently it's some kind of diagnostic mode, I'm not quite sure what it would help you diagnose though. Mine sounds different everytime, like all the Osc's phase diiferently each time you turn it on, can get some nice bass sounds
Urs wrote:Is there a similar trick on the alpha Juno 2
Not that I know of

Post

Really interesting, thanks!

So maybe if one lets it do enough rounds it doesn't even need an initial state...

I wonder though if it needs a different set of selection criteria though...

I might just try this for fun over the weekend, considering that the DHL guy will again only drop a card in my letter box because the Jupiter 6 I'm getting delivered is too heavy for him ;)

;) Urs

Post

Urs wrote: Amazing! But how do I detune the voices?
It's analog...just wait a few minutes, they'll detune themselves. :wink: :)

Post Reply

Return to “u-he”