Repro-1 (out now)

[pdxindy]

I didn't have a Pro-1, but I used it's great grandkid, the Prophet 6 as a measure of what a similar filter in an analog synth would sound like... And I thought that 4 was a fairly clear winner. Of course, the Pro-1 isn't the Prophet 6, but I thought that would be a better way to judge than to leave me to my own devices. I do remember thinking before I did the comparison that #5 was one of my favorites for some types of sounds. More muted high end and a bit smoother. 4 gets a bit screechy when you start cranking osc mod and resonance in the Prophet 6, just like filter #4 does. Actually, the best synths that I've played with that do that well are Moogs or Moog filter clones. Something about that ladder filter remains kind of musical at extremes in a nice way.

In software, currently the System 100 plug in does a superb job of that kind of mayhem. Reaktor 6/Monark too.

I prefer 3 over 4... at some extreme settings, 4 has some artifacts where 3 doesn't...

[PAK]

Well, self-oscillation on 2 and 5 dies when cutoff is swept beyond 20-30KHz, which is unlike not some but every Pro One out there. So, I guess that none of RePro filter algos is working properly then, right?

Depends what your definition of "properly" is. I don't think fmr is alone in what he picked up on. Also, the specific emulated synths output is the 2nd post on page 16 of this thread. Can RePro recreate all the filter's characteristics with any of them?

I'm not sure if I'll have the time, but if I have, I will try to create some simple filter sweep patches with high resonance in DIVA with the Minimoog, the Júpiter and the MS-20 filters, and then some closer patch in Re-Pro, and bounce all the results.

Even for the OSC B modulation Diva can go towards those sounds. You just need to modulate the Filter FM and main filter mod slot with an LFO multiplied by itself and set to max rate. Then moving the filter control will produce similar sounds.

The best structure for it would likely be a Dual VCO into feedback then Cascade 12dB or Bite Rev 1. The feedback used more for gain, since the results volume may lack otherwise. You might want to set only Osc 1 active, but move the mix control all the way to osc 2 to hear.

It's quite interesting what you get when you push Diva like this, and how the quality settings don't really make any difference to the outcomes! I guess throwing CPU at things won't help if what it seeks to solve doesn't explain the actual differences you hear. Rather, it's simply a waste of resources.

Anyway, I think you're right to trust your ears on this, and comparing to the audio example on page 16 will maybe tell people much more than walls of text might too.

[PAK]

Having said all that, I like Diva's filters much better. There is no contest.

Perhaps that's not surprising if you keep in mind Diva's filter code was also (if not primarily ??) the work of Clemens Heppner. Given Diva's reception, perhaps Urs feels under some pressure to replicate Diva's success in equivalent products. However a key ingredient for that success, Clemens, is no longer there.

I see an awful lot of assumptions being made based on Diva. Some people clearly don't "get" what it means when a key coder leaves. Consider Synth Squad without Andrew Simper.

Of course, it doesn't mean "the end of the line". But it does mean what you get, moving forward, will obviously be very different. Stagnated, worse, or better you can debate. But it will be different.

I'm always mindful not to under-estimate how difficult it is to push boundaries. For all the talk so many people do, the hard bit is the implementation, and we still waited the best part of a decade for native plugins to arguably surpass the work of Klaus Piehl (Minimax) on the Creamware platform. A decade!

Anyway, hopefully another audio dev might get Clemens Heppner back into creating audio plugins some day..

[EvilDragon]

I wouldn't suggest that Urs' coding knowledge is necessarily lesser than that of Clemens'... If code is well-documented, there are no problems even if the main coder leaves.

[hakey]

we still waited the best part of a decade for native plugins to arguably surpass the work of Klaus Piehl (Minimax) on the Creamware platform

iirc, Urs said something about Diva's Moog filter model including a detail (a stage of waveshaping?) not found in the Minimax.

[PAK]

I wouldn't suggest that Urs' coding knowledge is necessarily lesser than that of Clemens'...

That's the wrong way to look at it. By that measure the example would also be saying the same about Angus, who is clearly an extremely talented coder.

Yet he still brought Andrew in. Why? Likely because he recognized the value in having people with different skill sets involved in a project and, if you can check your ego in, frequently working together will create something greater than if either had worked alone.

What the words above are trying to say is that the actual people who create the products matter. That's why I remember Klaus Piehl's name. Would it be saying the rest of the Creamware guys weren't crazy good, at what they did, if you said he'd be missed if they were creating a synth? Of course not.

And it sort of goes without saying that Diva would not be what it was without Urs involved too. And, btw, whoever did the oscillator code on RePro has done a very nice job, so it's not like it's saying other people can't have talent!

If code is well-documented, there are no problems even if the main coder leaves.

To an extent that is true. But not entirely. That coder would have had his own "style", his own methods and ways of doing things, his own ideas. Like I said, it doesn't mean to say something will either necessarily be better or worse, when someone is lost. But, moving forward, the one thing you can count on is that it will be different.

[EvilDragon]

Not if they have coding guidelines, coder's style is then out. And considering u-he has their own plugin framework, I'm pretty sure they have guidelines for coding as well, so that code can be easily readable by any of them.

In any case, I don't think what you imply here is going to have an impact you think it might but for sure it would be interesting to read what Urs has to say on that matter

[hakey]

Surely it's the mathematical modelling, not the coding (style, or framework, or whatever), that makes a difference so far as what we hear?

It seems pretty clear that Urs understands the modelling side. Read his blog.

[PAK]

Not if they have coding guidelines, coder's style is then out.

Style includes their problem solving abilities. Or is there only ever one way to solve a complex problem?

Granted, part of the issue with emulative stuff is also understanding and defining the problems in the first place. That's also why two people can attempt to solve the same problem, yet produce results that are pretty different. That has nothing to do with coding guidelines. It's about the personality of the coder deciding what matters, what they think is happening, and their creative approach to solving problems.

If you're using previously published answers then the importance of the individual is obviously reduced. (AKA why do all my plugins sound the same type posts)

Yes, component variation and calibration can also create differences between code. Just like code simplification does. It's left to end user judgement as to which it might be, since product makers will rarely talk about the ways in which their products fail

But, amongst all that, there is a lot of room for individuality to be expressed in problem solving, and for it to make differences people may hear.

In any case, I don't think what you imply here is going to have an impact you think it might but for sure it would be interesting to read what Urs has to say on that matter

The break down of a successful working relationship is unlikely to be a topic anyone would want to discuss much. Urs own words have made it clear he's extremely confident he can take things beyond Diva.

[aaron aardvark]

hakey,
Are you saying you followed the same procedure on RePro One as Urs' video? That is something I did not try.

[hakey]

hakey,Are you saying you followed the same procedure on RePro One as Urs' video?

Yes- the first test with increasing resonance. Filter 4 is closer to this behaviour than filter 3, which changes pitch noticeably.

Filter 3 may be the more expensive algo, but it doesn't accurately replicate every aspect of u-he's Pro One. Presumably the intention is to capture generic Pro One behaviour rather than that of a specific instance (though I seem to recall that the latter was hinted at somewhere along the way).

[fmr]

hakey,Are you saying you followed the same procedure on RePro One as Urs' video?

Yes- the first test with increasing resonance. Filter 4 is closer to this behaviour than filter 3, which changes pitch noticeably.

Filter 3 may be the more expensive algo, but it doesn't accurately replicate every aspect of u-he's Pro One. Presumably the intention is to capture generic Pro One behaviour rather than that of a specific instance (though I seem to recall that the latter was hinted at somewhere along the way).

Actually, I think that in the current state, Urs intention as just to check which model was "perceived" as the most analogue. They are still working on it, AFAIK, so, probably there still things to be perfected. And since all filters are actually the "same" (except for the calculation method, then the model is the same, which means that, after choosing one (or two?) calculation methods to stick with, the model has to be worked on the aspects that are less faithfull or accurate.

[PAK]

It seems pretty clear that Urs understands the modelling side. Read his blog.

I think it’s great that Urs has opened up his thoughts in the way he has, and hopefully it will benefit some.

Most devs also would not post things, like on Page 16, or the video you linked, since it then removes part of a snowflake defence about why differences exist. They exist partly because this stuff is hard to do.

And most “criticisms” can probably be applied to every other dev in some way, since nobody has reached perfection in their models yet, especially when it comes to a whole synth.

I certainly hope he keeps that in mind as he reads remarks which are saying things like “I don’t like this, or I don’t like that”

[hakey]

Urs intention as just to check which model was "perceived" as the most analogue.

That Urs wanted to know how other Pro One's behaved in that test makes me think something else was going on. A mismatch between the model's behaviour and u-he's Pro One would be one explanation.

[PAK]

BTW Part of the perception, with OSC B mod, might be down to just how much a single step of the on-screen control can change the sound, never mind a standard MIDI CC knob. Before I uninstall I'll try to get around to checking the perception created with higher resolution automation assigned