State of that drum synth?

[tapiodmitriyevich]

Hi guys,

when will that drum synth be done? Or is it discontinued? Or did your research find no way into a product? I'm all open for a plugin with realistic drum synthesis and no samples.

Thx

[Funkybot's Evil Twin]

I'd still love to see just a basic, old school, analog drum synth done at U-he level quality. Seems like an obvious gap in the marketplace and U-he product line to me. Add some nice effects, a fancy sequencer, and boom, next thing you know, it's been a solid 18 months of development. I guess what I'm saying is: I'd like that after Zebra 3 and Diva 2.

I'm still not sure I get what the more realistic drum product even is. How you do realistic drum synthesis without modeling rooms, and microphones, and bunch of other stuff which seems crazy prohibitive from a CPU perspective? And at that point, wouldn't samples just be the way you go? I'm thinking Pianotech for drums, so maybe I'm just confused about what it is. An analog style drum synth I understand.

[sascha]

A while back, Urs posted this:
http://www.kvraudio.com/forum/viewtopic ... 7#p6861383

It's still all alive, but ongoing, and we're thinking about going two directions. I can explain a bit further:

What we had at that stage is probably very suitable for including it into a synth-based product. One can think of a drum instrument as a 'sound source' (aka oscillator) within the scope of a polyphonic synth, or it even be part of its post processing, since it's all about vibrating objects and sympathetic resonances. We think there should be a nice amalgam of techniques and concepts that, at least how we think it, hasn't been worked on that much, or with that level of attention. What we have at this point suits a synth like Zebra perfectly, and we're within the process of implementing a couple of cool ideas as we speak.

On the other hand, that PM project led us to a number of conclusions. The most important one is probably, that, in order to achieve a concinving degree of realism (if you want it all the way 'acoustic' and equal playing feel and immersion), one needs to look much further and explore principles and mechanisms way beyond what is currently done in realtime audio DSP. Problem with drums is that 'realistic' modeling of an instrument (or worse: a set of drum instruments as a whole drumkit entity) is about dealing with a complex nonlinear system. As apposed to simple delay-based physical-modeling techniques, you just can't specify ins & outs, junctions, ports etc., it's just that every tiny piece is a nonlinear thing, and everything has memory, is dependent on internal and external states. Similar to analogue-filter design: you can't solve equations analytically, you have to do it the numerical way, and iterate over time to find solutions.

In a nutshell: we've come to the point where we need to clear the table, and explore different paths, and check out mechanisms similar to those of more traditional industry sectors. Imagine you're building a car, and want to know how it deforms during a crash. Or, you're designing an aircraft wing, and you need to know how the airstream flows along the surface and how much bending, vibration and buildup is tolerable within certain range and material setup. Alas, methods like finite elements, which divide a nonlinear problem up into tiny segments and solve each numerically, take up a hefty chunk of numbercrunching.
We're currently thinking about what aspects are essential, what needs to be addressed at what level of detail, and what can probably be addressed linearly, or de-coupled in a black-box approach. Like we did with filter design, it's science & art in one: you have to find ways to steer around the impossible; to find the right balance of realism vs. cpu load, to explore what makes simulation differ from reality, and where a 'feel' of realism starts, and at what point it all falls apart once you make too many shortcuts in favour of less CPU cycles.
That is a complete journey into the unknown, as opposed to trivially do a thing like 'let's take some off-the-shelf algorithms and make them sound like drums'. This definitely hasn't been done before, at least not at the level we think it should be done, and we're nowhere near a fast solution. It needs a novel approach.
Since the resources we'd have to invest is way more than we could typically handle, we're evaluating public funding programs and writing up piles of documents on that matter right now. We want to make it right, but can't just do it on the side, otherwise it would take us forever, apart from the many risks involved.

[tapiodmitriyevich]

Good luck! I also had "Pianoteq for drums" in mind. I definitely think with a good product and marketing, you could have a successful product since it has not been done before (at least not relevant), as far as I am aware of. And U-He already has that important thing called reputation when it comes to sound creation.

[synth_punk]

I would love a simple, synthetic U-he drum synth with pedal effects, but that's just me. I do not mind using Reaktor and Microtonic for this still and Chromaphone for modelling until you find the magic your looking for.

[xx JPRacer xx]

A while back, Urs posted this:
http://www.kvraudio.com/forum/viewtopic ... 7#p6861383

It's still all alive, but ongoing, and we're thinking about going two directions. I can explain a bit further:
...
...
...

[lunardigs]

... a complete journey into the unknown, as opposed to trivially do a thing like 'let's take some off-the-shelf algorithms and make them sound like drums'. ... It needs a novel approach.
Since the resources we'd have to invest is way more than we could typically handle, we're evaluating public funding programs and writing up piles of documents on that matter right now. We want to make it right, but can't just do it on the side, otherwise it would take us forever, apart from the many risks involved.

No doubt.
Btw, I'm a huge fan of this. I'll definitely pitch in if you crowd-fund.

[UmFuFu]

Wow this sounds amazing.

[nevis]

I'll definitely pitch in if you crowd-fund.

Same here.

[Beamboom]

I'll definitely pitch in if you crowd-fund.

Same here.

Same here too. I'm not a fan of crowdfunding in general (too many broken promises) but this project here I'm all in on. Heck I sit here throwing money at my monitor at this very moment, my right hand instinctively reached for my wallet!

A sexy drum machine, U-He quality level, is more or less the last brick in my studio puzzle. Just to have some virtual cymbals to even remotely behaving like real cymbals is a dream I've had ever since I tried my first drum machine back in the late eighties.

[sascha]

We're not talking crowdfunding, we're evaluating public (EU- or local-based) funding for small- & medium-sized enterprises. There are programs that help companies to start research and development on something that includes high technical risks (= something that otherwise would most probably not get addressed). Which is the key aspect, and very different to crowdfunding. It's not that we just need money in the first place, and then - bam! - comes a product which was pretty much layed out beforehand (as a premise to open peoples' wallets). It's more that we can't foresee delicate risks involved along the way, and the overall feasibility. It could also mean that we invest 2 years and only come up with the conclusion that it's not doable the way we thought.

[sascha]

(oops, double-post hiccup)

[nilhartman]

Sascha, would it be possible to share some audio samples of what can be done with what you already have? What I've heard from the low quality youtube vids of what you showed at Superbooth 2017, I'd pay big bucks to have that in my arsenal. Even if buggy / not even an alpha version etc...

A drum/percussion oriented, single voice synth by U-he would already be fantastic. I don't know, something semi-modular à la Diva: you pick an engine (analogue / physically modeled, membrane / cymbal etc...) as a starting point and go from there. I avoid samples as much as I can, and I can't find any definitive drum synth.

As I type, I'm listening to a kick drum patch I've just made with Bazille, which sounds great (sort of 808ish)... but I'd love something a tad more predictable and to the point. And sampling it somehow defeats the points of RT synthesis, no?

[sascha]

I currently don't have the time to make any audio demos, but... have you seen this one?

(It's German, but maybe the YT auto translation partly works, at least the got a direct audio feed from the DAW)

I'm pretty certain that the possibilities within Z3 would be endless, way more than we could show with that concept back then.

[rod_zero]

It is very interesting to know u.he are planning to invest on R&D on physical modeling technology for drums, certainly something that hasn't been done that much.

Nevertheless I think I would like to see an Analog drum machine emulation in the spirit of Diva, there is nothing on the market with u.he quality.