Entertaining "Diva or OB8" poll happening on GS

[Urs]

Interesting! That's news to me, we have never analysed the OB-X or SEM discrete VCOs, nor did we perceive this with the CEM3340s. One percent is quite a bit... it'll sound much like noise modulation, no?

yeah it is quite interesting...I was wondering if that perhaps was why people percieve DCO to be a bit harsher/harder...perhaps even digital models also? Is this something you could verify with all your clever gear?

I could probably measure it... I feel tempted, but our studio is boiling hot. I just so manage to sit still and fix some bugs...

Maybe over the weekend, if wife lets me toddle off to the studio.

[Gamma-UT]

yeah it is quite interesting...I was wondering if that perhaps was why people percieve DCO to be a bit harsher/harder...perhaps even digital models also? Is this something you could verify with all your clever gear?

I just looked at some OB8 samples – no actual OB8 here sorry - and got no more than a single sample's worth of jitter from any of them. One did wind up a bit at the start, which might be related to voltage drop with a sharp attack at a guess, but that change was less than 1% of the target frequency. Once it was playing, there was no greater jitter than one sample.

[News From The Sky]

Hey Urs, any chance we could get an "enhanced" Dual VCO at some point?

- individual PWM for each VCO
- subosc for each VCO
- independent noise source

I've noticed that the Voyetra 8 uses the same CEM 3340 but squeezes a bit more out of it than the Jupiters. I know the chip isn't everything, though.

[Edit: I mean as a higher-CPU alternative, of course]

[AusDisciple]

Interesting! That's news to me, we have never analysed the OB-X or SEM discrete VCOs, nor did we perceive this with the CEM3340s. One percent is quite a bit... it'll sound much like noise modulation, no?

That one percent thing grabbed my attention too. That sounds like an awful lot of jitter to me. Something like that should be quite noticeable on a 'scope. I wonder if the Roland VCO's are in the same one percent ballpark? I might fire up my 'scope and compare System 100 to Juno 6 and see if there's anything noticeably different in the jitter. I've never noticed the System 100 VCO's sounding that 'fuzzy' though.

[ghettosynth]

Interesting! That's news to me, we have never analysed the OB-X or SEM discrete VCOs, nor did we perceive this with the CEM3340s. One percent is quite a bit... it'll sound much like noise modulation, no?

That one percent thing grabbed my attention too. That sounds like an awful lot of jitter to me. Something like that should be quite noticeable on a 'scope. I wonder if the Roland VCO's are in the same one percent ballpark? I might fire up my 'scope and compare System 100 to Juno 6 and see if there's anything noticeably different in the jitter. I've never noticed the System 100 VCO's sounding that 'fuzzy' though.

I wonder if there is some misinterpretation going on, along with some perhaps some exaggeration. I find it very difficult to believe that a JFET has that much variation with respect to how "fast" it switches to discharge an integrating capacitor. I can believe that many factors adding up, including variability within the comparator, will contribute to some minor variations cycle over cycle. I'm not sure that 1% means 1% of the frequency though. Perhaps, with respect to the JFET, they more reflect how quickly the capacitor discharges and so will shape the reset edge of the sawtooth, and consequently, the harmonics. Comparators, also, will have some minor variability with respect to their trigger voltage, hence timing, especially if local circuits, including other modules, impact the supply rails.

[zerocrossing]

Interesting! That's news to me, we have never analysed the OB-X or SEM discrete VCOs, nor did we perceive this with the CEM3340s. One percent is quite a bit... it'll sound much like noise modulation, no?

yeah it is quite interesting...I was wondering if that perhaps was why people percieve DCO to be a bit harsher/harder...perhaps even digital models also? Is this something you could verify with all your clever gear?

I'm curious about this too. When I picked up a MoPho the usual suspects denounced it due to it's DCOness and I wondered if they indeed had something. I ended up picking up a used Studio Electronics ATC-X and I must say that it seemed to have a "rounder" more "warm" sound, but it's a totally different synth of course. I assumed it was the VCOs, at least in part.

[Urs]

Well, DCOness would be the opposite of jitter.

[Gamma-UT]

I wonder if there is some misinterpretation going on, along with some perhaps some exaggeration. I find it very difficult to believe that a JFET has that much variation with respect to how "fast" it switches to discharge an integrating capacitor. I can believe that many factors adding up, including variability within the comparator, will contribute to some minor variations cycle over cycle. I'm not sure that 1% means 1% of the frequency though. Perhaps, with respect to the JFET, they more reflect how quickly the capacitor discharges and so will shape the reset edge of the sawtooth, and consequently, the harmonics. Comparators, also, will have some minor variability with respect to their trigger voltage, hence timing, especially if local circuits, including other modules, impact the supply rails.

This, or something along these lines, makes more intuitive sense to me. I can't believe no-one would notice 1% frequency jitter if it was there – a basic 555 timer must do way better than that. Plus it it surely would be relatively easy to build into a model compared to the more subtle feedback-led or supply-rail effects that seem more likely to be the culprits.

[SWAN808]

its could be that 1% was an exaggeration because that comment started out as a post on Youtube! This developer I know was discussing something and he resent it to me as part of our conversations. Still its interesting - because - it could explain the slight feeling of sterility to DCO and digital Osc...even if it was a very small amount of movement - these things amazingly do add up IME...

[ghettosynth]

its could be that 1% was an exaggeration because that comment started out as a post on Youtube! This developer I know was discussing something and he resent it to me as part of our conversations. Still its interesting - because - it could explain the slight feeling of sterility to DCO and digital Osc...even if it was a very small amount of movement - these things amazingly do add up IME...

I think that this comes down to the definition of "sterility," but, it seems to me that the majority of difference in long term stability which creates some movement in the voices comes from thermal drift in the expo converters. This is more noticeable in two oscillator synths because they have two expos per voice. It's also the case that synths that use linear VCOs, or linear VCOs with a shared expo converter, also have some similarity to DCOs. The Korg Poly 6, for example, has only one expo converter.

[ddeez]

Urs i guessed it right on twitter...u favorited my tweet go back and look at it!!! i've got the golden ears you need to employ me as a tester i'm telling you!!!

[Breeze]

@Urs: While this discussion has narrowed into the sound of the basic oscillators, we had briefly discussed OB-X envelopes as well in the "New DIVA Components" thread earlier this year (when your studios must of been a much more comfortable work environment... or too cold? )

Apart from the oscillators and their features, another component of a synth's sound is the envelope ADSR curves themselves. As I'd mentioned back then, there are some behaviors that are simple to achieve on an OB-X but seem to be harder to pin down using the available envelopes in DIVA; and even with the shift modifier key, it's sometimes frustratingly hard to "get there". I still think Diva could benefit from either OB-X style envelopes, or to be wider in scope, some new customizable generic envelope that allows user tailoring of the individual curves.

[AusDisciple]

... I still think Diva could benefit from either OB-X style envelopes, or to be wider in scope, some new customizable generic envelope that allows user tailoring of the individual curves.

That sounds like a great suggestion to me. Customizable EG curves would be really useful. Linear, log and expo would be a good starting point IMHO.

[Urs]

Well, yeah… if we have the time and if it doesn't make Diva too complicated.

Myself I'm getting more and more fond of the idea to turn Diva into a synth on her own right rather than an assembly of attributes to model specific vintage gear. The Uhbie filter was such a choice, as was the Digital Osc.

We'll see where it goes. There's going to be a year (at least) between now and concpetual additions to Diva.

[Breeze]

Linear, log and expo would be a good starting point IMHO.

Actually I was thinking along the lines of what several other devs have explored: having a boxed 2D space that allows reshaping of each curve for each leg of the EG in an "inifinite" number of ways. This would allow finer control over acceleration and deceleration within each moving envelope leg, especially with more than one pivot point. Come to think of it, you could replace all the current envelopes by introducing presets for the generic one...