The long DIVA thread

[fmr]

Hey,

Diva is still supposed to become a simple, standard synth. As research/brainstorming went on, she'll have the following structure:

2 LFOs just like in TyrellN6
Oscillator Section: Different models ranging from 1 to 3 oscillators
Mixer: Two or three models
Filter Section: A selection of models, mostly HP/LP combinations.
VCA: One fixed section but maybe some choice in character
Env Section: Two envelopes, 3 or 4 models to choose from, one maybe with 4 rate/level controls
FX Section: Probably just a simple Chorus (but a good one)
Urs

Please consider adding a delay besides the chorus. Delay is a highly useful effect in synth programming, IMO, and although the analog synths you mentioned as your "paradigms" do not have it shouldn't prevent you from add one.
And what about the Xpander filter model? Did you made any advance on it?

[Urs]

Hi Urs,
if possible, could you please integrate in Diva different quality modes
for modules that neeed much power?
And a "high rendering mode" that works only when a song is rendered in a DAW?
So many people with an older computer could use DIVA.

The cpu load will mainly depend on the filter model and the accuracy of modeling.

At the moment it looks like for each filter type there'll be a version with the quality of ACE/Tyrell/Zebra (which is already very, very good). These solve the differential equations with an approximation. But for some (maybe all) filters there'll be an accurate model which uses iterative methods to eleminate the legendary "one sample delay" that is allegedly necessary for digital computation (of course it isn't, but the implications are mind boggling).

The idea here is that for offline rendering there's a choice of replacing the approximation with the accurate model. Or have the offline render the same model as used in realtime. If however an accurate model is chosen for realtime, it'll be monophonic (if it plays at all, that is...).

There could however be a draft mode with lesser oversampling. Maybe.

Urs

[Urs]

Please consider adding a delay besides the chorus. Delay is a highly useful effect in synth programming, IMO, and although the analog synths you mentioned as your "paradigms" do not have it shouldn't prevent you from add one.
And what about the Xpander filter model? Did you made any advance on it?

Delay is an option, yes. As is a Phaser. We'll see.

We currently don't have a CEM 3372, but I'm toying with the idea to get one - without the dongle around it. The Matrix 12 I recently had was in such a bad shape, I don't ever want to see such a device again. I admittedly do hold a grudge against that machine. I also don't think I want to emulate any of those digital envelopes and modulations Oberheim used in their late synths.

That said, what's wrong with the XMF? - I thought it was pretty accurate as it is... just more flexible and more *balls*.

[fmr]

Please consider adding a delay besides the chorus. Delay is a highly useful effect in synth programming, IMO, and although the analog synths you mentioned as your "paradigms" do not have it shouldn't prevent you from add one.
And what about the Xpander filter model? Did you made any advance on it?

Delay is an option, yes. As is a Phaser. We'll see.

We currently don't have a CEM 3372, but I'm toying with the idea to get one - without the dongle around it. The Matrix 12 I recently had was in such a bad shape, I don't ever want to see such a device again. I admittedly do hold a grudge against that machine. I also don't think I want to emulate any of those digital envelopes and modulations Oberheim used in their late synths.

That said, what's wrong with the XMF? - I thought it was pretty accurate as it is... just more flexible and more *balls*.

You mean, the Zebra XMF filter? Nothing, AFAIK. It's a great filter, as every other module in Zebra. However, it's not the same thing, IMO.
Anyway, I guess that the greatness in Matrix-12/Xpander relies more in the whole sum of the parts rather than any of them in particular, and I am aware you are not in the same mood as I regarding these, so ...
I'm sure DIVA will be a great synth, nonetheless.

[eXode]

I think Doepfer sells a CEM 3372 based filter (A-106-6 XP VCF) I don't know if it's based on the actual circuit though.

Please consider adding a delay besides the chorus. Delay is a highly useful effect in synth programming, IMO, and although the analog synths you mentioned as your "paradigms" do not have it shouldn't prevent you from add one.
And what about the Xpander filter model? Did you made any advance on it?

Delay is an option, yes. As is a Phaser. We'll see.

We currently don't have a CEM 3372, but I'm toying with the idea to get one - without the dongle around it. The Matrix 12 I recently had was in such a bad shape, I don't ever want to see such a device again. I admittedly do hold a grudge against that machine. I also don't think I want to emulate any of those digital envelopes and modulations Oberheim used in their late synths.

That said, what's wrong with the XMF? - I thought it was pretty accurate as it is... just more flexible and more *balls*.

[News From The Sky]

The Matrix 12 I recently had was in such a bad shape, I don't ever want to see such a device again. I admittedly do hold a grudge against that machine. I also don't think I want to emulate any of those digital envelopes and modulations Oberheim used in their late synths.

Please don't let the poor quality of '80s digital modulators deter you from incorporating the underlying concepts of those modulators. Done right (ie with higher speed and resolution, and variable slopes) they can be very powerful. There are lots of analog-inspired softsynths with just a couple ADSRs and basic LFOs. It'd be nice to see more tracking generators, lag generators, multistage loopable envelopes, complex LFOS and so forth, that had the speed and smoothness of analog but the flexibility and power of digital sources.

(Judging by Zebra I don't think I need to tell you this, though.)

BTW have you looked at the modulation processors (lag, diode, quantizer, abs value, etc) available in the last generations of E-mu synths and samplers? Might be some interesting ideas there.

That said, what's wrong with the XMF? - I thought it was pretty accurate as it is... just more flexible and more *balls*.

Is the XMF based in part on the Xpander filter? Innnnnnnteresting...

[Gonga]

If you could get the overdriven sound of the Memorymoog filter even close, without too much digital harshness, I would be a happy man.

The Memorymoog is one of the "holy grails" that still hasn't got an emulation. I know at least a few people who've practically gotten on hands and knees to beg Arturia for an emulation. Myself, I wouldn't mind nailing that Miami Vice lead sound...and therefore I'd say a smooth polyphonic sync sweep is at least just as important as the filter, but that's just me.

[eXode]

http://www.memorymoon.com/memorymoon.htm

It's not a 'pure' emulation but you can see where it's coming from. :)

The Memorymoog is one of the "holy grails" that still hasn't got an emulation. I know at least a few people who've practically gotten on hands and knees to beg Arturia for an emulation. Myself, I wouldn't mind nailing that Miami Vice lead sound...and therefore I'd say a smooth polyphonic sync sweep is at least just as important as the filter, but that's just me.

[Gonga]

Wow, interesting. I know at least a few people who might be interested. Thanks for the heads up.

[Urs]

The thing about the Memorymoog is, when Hans Zimmer sold me his, I had to promise to never make an emulation. Hans found that the Memory has a certain "screechyness" that the Minimoog does not have. Having all this gear side by side, it's easy to spot the subtle differences, and I totally agree. A polyphonic Minimoog will sound much stronger, yet smoother than a Memorymoog. It'll be all fine, and it (the Mini within Diva) will have more modulation options than the Memory.

The only thing that will not be present is PWM and Sync of course. Which can be achieved quite sufficiently by some other oscillator models.

Again some words about the Matrix/Xpander. The concept of those is simply too complex for Diva. If anything then we should rather look at Zebra and add the missing bits (lag generators, analogue fm). I'm pretty sure that this would cover *all* grounds of the hardware, but with faster modulations and a way more intuitive user interface.

I don't think anyone would want the user interface of any such synth replicated for mouse interaction. I found the Matrix interface clumsy and unintuitive by todays standards. A software version would have to look completely different in order to be usable. We can create a Zebra skin which exhibits just that feature set for instance, with specifically scaled parameters and all, and be done with it. Maybe we do that one day

Cheers,

Urs

[eXode]

@ urs: Another thing, do you have any intention of emulating an ARP 2600 or Odyssey in DIVA? How about Oberheim SEM? ARP and SEM (2/4 voice) was synths I couldn't spot in the gearporn images on the new site!

I understand that you probably have to draw the line somewhere but still interesting to know! :)

[Urs]

@ urs: Another thing, do you have any intention of emulating an ARP 2600 or Odyssey in DIVA? How about Oberheim SEM? ARP and SEM (2/4 voice) was synths I couldn't spot in the gearporn images on the new site!

I understand that you probably have to draw the line somewhere but still interesting to know!

Send me an ARP 2600 or some Odyssees and I'll see what I can do :-p - seriously, I know almost nothing about those machines. The good ones are abviously made so that one can not reverse the circuit (they're glued in a block or something). So I'm actually not too keen on spending too much money on these just yet. The primary intention of Diva is rather Rolandish anyway, and I think we got a rather representative collection, including the original diode ladder design.

The TyrellN6 filter is loosely based on the SEM Circuit, but we don't yet have an original too verify some aspects. It had to be self-oscillating, so I had to betrickst it.

Some of the gearporn is still missing... there's an OBXa that's almost causing a lawsuit ( owner lent it to someone who doesn't want to give it back), there's an Opera6 in the mail (SSM filter yay) and we're looking into some Doepfer modules.

And as said before, we're simply building some of the circuits ourselves with some of the vintage OTAs, CEMs and other parts we've stocked. We're anxiously awaiting the time after the current updates so we can finally do this!!!

U

[rockin1]

Again some words about the Matrix/Xpander. The concept of those is simply too complex for Diva. If anything then we should rather look at Zebra and add the missing bits (lag generators, analogue fm). I'm pretty sure that this would cover *all* grounds of the hardware, but with faster modulations and a way more intuitive user interface.

I don't think anyone would want the user interface of any such synth replicated for mouse interaction. I found the Matrix interface clumsy and unintuitive by todays standards. A software version would have to look completely different in order to be usable. We can create a Zebra skin which exhibits just that feature set for instance, with specifically scaled parameters and all, and be done with it. Maybe we do that one day

This sounds like a great plan Urs! The sound is what is important. I don't always care for the ancient UI designs, but sometimes the sounds are incredible. That is certainly the case with the Matrix/Xpander stuff...

[eXode]

@ Urs: If I could (If I had one or the other) I would! On another note I think amsynths.co.uk are in the future releasing a Eurorack filter module based on the Odyssey MK1 (white face) that featured a SEM-like 12dB LP filter, also STG Soundlabs has a "Post Lawsuit Filter" module in Eurorack format that is based on the design ARP used after being accused to mimic Moog's ladder design too closely.

There's a load of documents by amsynths on the 2-pole odyssey filter as well as various other filter designs here.

Don't know if that helps.

Send me an ARP 2600 or some Odyssees and I'll see what I can do :-p - seriously, I know almost nothing about those machines. The good ones are abviously made so that one can not reverse the circuit (they're glued in a block or something). So I'm actually not too keen on spending too much money on these just yet. The primary intention of Diva is rather Rolandish anyway, and I think we got a rather representative collection, including the original diode ladder design.

[fmr]

Again some words about the Matrix/Xpander. The concept of those is simply too complex for Diva. If anything then we should rather look at Zebra and add the missing bits (lag generators, analogue fm). I'm pretty sure that this would cover *all* grounds of the hardware, but with faster modulations and a way more intuitive user interface.

I don't think anyone would want the user interface of any such synth replicated for mouse interaction. I found the Matrix interface clumsy and unintuitive by todays standards. A software version would have to look completely different in order to be usable. We can create a Zebra skin which exhibits just that feature set for instance, with specifically scaled parameters and all, and be done with it. Maybe we do that one day

This would make very happy, I can assure you. That's not the picture that I am keen of, it's the synth architecture. If you can mimmick it entirely, filling the missing pieces, I'd say: Go for it.