The long DIVA thread

[Gonga]

MIDI Learn suggestion: Allow a single CC to be assigned to multiple controls.

+1

[PAK]

I thought the main difference was in the behavior of the envelopes. Or do you think its just the envelopes in his Diva patch are not properly adjusted?

It is. In those examples Divas envelopes are behaving a little bit more linear. Some are going to continually make a big deal out of those differences unless Urs gives people a way to modify the shapes of the envelopes. In fairness they do have a point IMO - it makes Diva sound a bit "more precise" in that example, where the Jupiters filter sounds a tiny bit more "gloopy" (and arguably more musical) and this is largely down to the shape of the filters attack envelope. The cutoff is more closed in the Jupiter 8 example too though, so it should indeed be possible to tweak Diva closer. But I really wish Urs would just bite the bullet on this stuff and add a second separate trimmer tab for envelopes at some point

Re Voice Trimmers - It would also be nice to have enough assignable sources so that cutoff and resonance can be adjusted independently. In voice variation 99.99% of the time the only things that matter much (IMO) are pitch, pan, cutoff and resonance. As long as those basics are covered then that mainly leaves the... envelopes.

Envelope adjustments wouldn't only help shut up people who don't have much else left to pick at, it'd be musically useful and would also leave a framework where possible future synth modules, which have more variations between voices, would already have everything they need in place to enable that sort of behaviour to work very well. Envelope adjustments (by which I mean controls to actually shape an exponential curve, and not merely a linear/expo switch) would actually be my number one request for new features in Diva. The second would be a more advanced key follow along the lines of (dare I say it ) SEM V.

Both of these things would be very helpful in nailing down differences where code alone cannot. After Urs and co are out of "bug squashing" mode I really hope they'll think about this. Measurements and code will only take you so far. To get "the last few percent of the way" some of it actually involves handing some control over to the users so they may mess up some of those measurements slightly Provided a good job has been done on the overall modelling (and it has in the case of Diva), the very next place people are likely to notice differences are in envelope shapes - that thread being just one example of that.

[Urs]

In those examples Divas envelopes are behaving a little bit more linear. Some are going to continually make a big deal out of those differences unless Urs gives people a way to modify the shapes of the envelopes. In fairness they do have a point IMO - it makes Diva sound a bit "more precise" in that example, where the Jupiters filter sounds a tiny bit more "gloopy" (and arguably more musical) and this is largely down to the shape of the filters attack envelope. The cutoff is more closed in the Jupiter 8 example too though, so it should indeed be possible to tweak Diva closer. But I really wish Urs would just bite the bullet on this stuff and add a second separate trimmer tab for envelopes at some point

Well, maybe he's using a Rev 1 Jupiter and ours is a Rev 2

The envelopes of our Rev 2 are indistinguishable from our Juno 60, and Diva's "analogue ADSR" matches these perfectly.

There will always be people who will make examples where the analogue will sound better than Diva. There will also always be people who make examples the other way round.

We could at any time produce a dozen or so examples where Diva sounds better, more alive, more "3D", more silky, more anything than any of the analogues her parts are modeled after. But that would be just completely useless.

Diva speaks for herself. If you need audio examples to prove that she's either right or wrong, then Diva is not for you

[Nielzie]

Diva speaks for herself. If you need audio examples to prove that she's either right or wrong, then Diva is not for you

My Diva thinks she's always right and never wrong

I could record some audio examples of that (but they won't sound very pretty)

[hakey]

We could at any time produce a dozen or so examples where Diva sounds better, more alive, more "3D", more silky, more anything than any of the analogues her parts are modeled after. But that would be just completely useless.

I agree. But wouldn't comparisons showing how close DIVA gets to your Mini, Juno, JP8 etc (ie not better or worse than) have value?

[Fritze]

My Diva thinks she's always right and never wrong

I could record some audio examples of that (but they won't sound very pretty)

I guess she will use some very analog envelopes for that...

[Urs]

We could at any time produce a dozen or so examples where Diva sounds better, more alive, more "3D", more silky, more anything than any of the analogues her parts are modeled after. But that would be just completely useless.

I agree. But wouldn't comparisons showing how close DIVA gets to your Mini, Juno, JP8 etc (ie not better or worse than) have value?

Well, we've done that quite often in the past.

The gist of an analogue emulation is that no sound is identical twice. Thus we won't have a chance to get two identical audio files. And thus there'll always be naysayers who believe to hear a fundamental difference that only magically trained ears can perceive even though "they are quite obvious". It's a stupid game. Not worth wasting time for.

[Fritze]

There will always be people who will make examples where the analogue will sound better than Diva. There will also always be people who make examples the other way round.

We could at any time produce a dozen or so examples where Diva sounds better, more alive, more "3D", more silky, more anything than any of the analogues her parts are modeled after. But that would be just completely useless.

Diva speaks for herself. If you need audio examples to prove that she's either right or wrong, then Diva is not for you

This!

[Urs]

(I'm completely nerved by Apple's complete lack of practical documentation of multicore support for time critical applications... currently pulling my hair out... shouldn't post in that mood)

[metrosonic]

(I'm completely nerved by Apple's complete lack of practical documentation of multicore support for time critical applications... currently pulling my hair out... shouldn't post in that mood)

Fear not Urs. I find it really productive to occasionally throw my toys out of the pram in pubic especially when there are so many 'experts' stirring the pot with their own skew on what you should be trying to achieve. I don't often comment on this stuff but I think you're one of the most responsive and least defensive of the developers out there and I'm pretty certain most of your loyal customers feel the same way.

Keep up you sterling efforts and I'm sure Diva is going to live up to her name!

JM
------------
http://soundcloud.com/leftside-wobble

[hakey]

It's a stupid game. Not worth wasting time for.

You've had a recent change of heart about the idea? iirc, there was talk, first of blind tests, then some kind of objective comparisons? Anyway, fair enough, that's your prerogative.

[lagavulin16]

(I'm completely nerved by Apple's complete lack of practical documentation of multicore support for time critical applications... currently pulling my hair out... shouldn't post in that mood)

http://www.kvraudio.com/forum/viewtopic.php?t=334284

Regardless, your input is welcome.

[Bronto Scorpio]

In those examples Divas envelopes are behaving a little bit more linear. Some are going to continually make a big deal out of those differences unless Urs gives people a way to modify the shapes of the envelopes. In fairness they do have a point IMO - it makes Diva sound a bit "more precise" in that example, where the Jupiters filter sounds a tiny bit more "gloopy" (and arguably more musical) and this is largely down to the shape of the filters attack envelope. The cutoff is more closed in the Jupiter 8 example too though, so it should indeed be possible to tweak Diva closer. But I really wish Urs would just bite the bullet on this stuff and add a second separate trimmer tab for envelopes at some point

Well, maybe he's using a Rev 1 Jupiter and ours is a Rev 2

The envelopes of our Rev 2 are indistinguishable from our Juno 60, and Diva's "analogue ADSR" matches these perfectly.

There will always be people who will make examples where the analogue will sound better than Diva. There will also always be people who make examples the other way round.

We could at any time produce a dozen or so examples where Diva sounds better, more alive, more "3D", more silky, more anything than any of the analogues her parts are modeled after. But that would be just completely useless.

Diva speaks for herself. If you need audio examples to prove that she's either right or wrong, then Diva is not for you

We could at any time produce a dozen or so examples where Diva sounds better, more alive, more "3D", more silky, more anything than any of the analogues her parts are modeled after. But that would be just completely useless.

I agree. But wouldn't comparisons showing how close DIVA gets to your Mini, Juno, JP8 etc (ie not better or worse than) have value?

Well, we've done that quite often in the past.

The gist of an analogue emulation is that no sound is identical twice. Thus we won't have a chance to get two identical audio files. And thus there'll always be naysayers who believe to hear a fundamental difference that only magically trained ears can perceive even though "they are quite obvious". It's a stupid game. Not worth wasting time for.

+1

It's a stupid game and it's impossible to win. The guys who think they hear a difference will always hear a difference. Ever tried to convince one of those audiophiles that his 3000$ USB cable doesn't sound better than my 3$ cable? Won't be possible.

Cheers
Dennis

[lagavulin16]

I do think it is possible to win - but I don't think the hardware or software is there yet. Go compare soft synths from 10 or 15 years ago (FM/digital ones aside) to ones today. Same with compressors and EQs trying to emulate Fairchilds or Pultecs... we get closer every year but we aren't there yet.

[Urs]

It's a stupid game. Not worth wasting time for.

You've had a recent change of heart about the idea? iirc, there was talk, first of blind tests, then some kind of objective comparisons? Anyway, fair enough, that's your prerogative.

We had some discussions and I changed my mind. I won't say never, but in the current situation I don't see any benefit in it - we've lost maybe 4 sales to die hard naysayers, and no A/B test will bring them back in.

Truth is, Diva sounds different - IMHO in a surprisingly nice way. We would need to eq the analogues to get them there.

Hence I'm currently more interested in getting her finished before we go on holidays.