Physical modelling drum product for 2018?

[yinbell]

Hi U-he. Looking forward to the physical modelling drum thingamagog you are building (hopefully). I am in the process of buying the Roland V-Drums TD series drum module, but if the thingamagog is going to be released in the near future, I might save some money on buying your software (I hope it's standalone?). Keep up the good work!

[sascha]

Well, considering the past lifecycles of the Roland flagships (TD 10,20,-X/TDW, TD 30, now TD 50), these were all like 5 - 8 years, so chances are we can make something within the TD 50 life span...
But it won't be anything we know from current e-drum modules, which are all mainly sample-based paired with some DSP processing on top. We're definitely trying to walk on unknown grounds, so we can't foresee at this time where we'll end up, and when, and in what shape it would come.

[aMUSEd]

A physical modelling plugin would be great but just limiting it to drums is boring, please make it do a wider range of sounds.

[pdxindy]

A physical modelling plugin would be great but just limiting it to drums is boring, please make it do a wider range of sounds.

If it could just do drums really well... that would already be quite something!

[EvilDragon]

It's not boring if it's done really really well

[aMUSEd]

Yeah but I don't really use drums and I just find the over focus on things like hats and kicks etc tedious

More exotic percussion gets a bit more interesting - hangs etc

But I would vastly prefer tonal instruments

[zvenx]

Is there a market for physical modeled drums?
rsp

[Scotty]

The Roland TD range is sample/physical model hybrid and I would think with some positioning a quality virtual drummer wih PM would be appealing . Considering that I recently paid $500 for an 8 year old TD 12 engine and a TD 20, 30 or 50 engine is out of reach for me there might be some market there. You would need the I/O for the inputs but that could be done cheaply enough given that they don't have to be high quality pre amps to act as triggers.. I know this is niche sounding stuff ; likely too narrow. Even without external triggering a good engine based on or augmented with PM characteristics might appeal. My musings.

Is there a market for physical modeled drums?
rsp

[zvenx]

Almost everyone I know who owns a Roland TD uses it as a percussion controller to trigger other stuff.
Maybe I just know the wrong people

I must imagine this would be a clique product mostly used by people who have days to tweak and experiment with drums.
Of course I have no data to support this , just my own intuition. I certainly imagine physically modelling technology would be more 'appreciated' indeed with not just making a drum modeling software.
rsp

[Scotty]

I use the engine both standalone and for triggering other sample libraries. You are right and probably most electronic drummers use the in-built sounds for live work and practice and when in the studio go for the big libraries and hundreds of kits available using Superior, Addictive or something else ; I can't argue with your main point. Oh well ... probably too niche but I'd like it.

Almost everyone I know who owns a Roland TD uses it as a percussion controller to trigger other stuff.
Maybe I just know the wrong people

I must imagine this would be a clique product mostly used by people who have days to tweak and experiment with drums.
Of course I have no data to support this , just my own intuition. I certainly imagine physically modelling technology would be more 'appreciated' indeed with not just making a drum modeling software.
rsp

[yinbell]

Is there a market for physical modeled drums?
rsp

I think there is.

The latest trend is to beef up live shows of DAW-created music with live drums to make it... you know, more LIVE, and I believe that this will create a demand for real drumming in the future.

If you look at new top of the range e-drumkits like the Roland TD-50 or Alesis Spark, the border between a real and a virtual drumkit gets smaller every day. And why would you buy a drumkit with only one type of sound in the future?

Although I'm a bit saddened by the news that the U-he product is so far away, I am glad you will spend some time making it right. However, I think there is something to be gained by releasing a beta product that can be tested by real drummers that can provide some feedback on the sound, because the sound is EVERYTHING in such a product, since the tech itself is already established.

If I am allowed to make a feature request, it would be to be able to import a sample and resynthesize it using the physical modelling engine.

Thanks for the clarification on the Roland TD-series being samplebased too. I thought the high end models were only using a type of physical modelling I can't remember the name of right now and the not-so-high-end models uses the Supernatural engine.

Now by the time I have saved up for a Roland TD-50 drum module (I have a project that will use only physical modelling instruments), U-he might have released the project anyway... But if I was U-he, I would make it more affordable than the drum modules on the market and I would cater to the drummers. For non-live drumming, sample libraries will always win because they don't make you stuck in a certain sound or feel that the earlier physical drum VST's did and flopped because of.

[sascha]

It's not the 'Alesis Spark', but Alesis Strike (or Strike Pro) . As far as I read (haven't tried it personally yet), the most comprehensive module atm. is the Pearl Mimic Pro (powered by a proprietary Slate SSD5 library on a 128GB internal SSD).
There you have two opposite architectures: Roland's type of sample manipulation (aka Behaviour Modeling) takes use of tiny fractions of a sample (like attack, decay, loopable areas, additional overtones etc.) and squeezes quiete a variety out of the underlying sample lib. The good thing is, it's superfast to load. Their endorser Michael Schack is a notorious real-time preset switcher (foot-switching in the middle of the action). I personally don't like the signature sound that much, but that's, inargueably, taste.
The other extreme is Mimic, or Strike, or 2box, or ATV AD5 which have 'VST-quality' sample libs, but you can't switch sounds in realtime. And the perceived realism is highly dependent on the number of velocity layers and round-robin or random samples per layer, to counteract the machine gunning. Quite a number of presets on those modules take several seconds to load, or up to 30secs or so. I personally use a MarkDrum Yes (we had that at Superbooth last year for people to trigger our drum-thing baby steps), and its bigger kits need around 8sec for a 200MB preset. While it still takes my guitarists ages to change tunings and pedalboard settings between songs, it's still annoying to wait that long. I'm aware of the cause, but some not-so-techy drummers might be really annoyed by loading times, I guess.
A PM approach would certainly be super-fast, since it's basically just about number-crunching. Well, easier said than done...

We hear you on the 'realism' thing. That's our driving force, and it's the most critical one. Along the way we found out that we need new ideas and approaches to the topic, and there's no easy solution. All currently established ways to create drums algorithmically are more or less simple linear systems and discrete building blocks. But the reality is, drums are complex nonlinear systems, so you just can't take a couple of simple, decoupled LTI building blocks and expect the beauty of complex interaction, buildup and feedback to happen like what happens in reality. This all demands a much greater portion of research on nonlinear systems, and completely different mathematical descriptions of the problem. We're not even fully sure if the usual suspects of piezo- & FSR-based trigger devices would be fully supportive, since the drummer's action basically remains one-way, so the 'drum pad' or whatever is used doesn't know anything about the resulting sound, therefore the whole act remains incomplete. We've decided to let some ideas mature on the side, decoupled from our regular dsp business, and get our heads around it whenever possible. But it's a tough thing. At least for my tiny dsp horizon, the term rocket science fully applies.

[Scotty]

I've often thought that the positional aspect of the electronic drums could use a boost. I imagine some kind of kinect like sensor field aimed at the drum or cymbal surface that could communiate the subtle positional information to the engine. Some kind of retro-fit kit that could be attached to the pads and midi merge with the engine . Some virtual drummers have numerous midi note triggers or control numbers to mimic positional information of any given drum using controller information or they randomize the playback to give it more authenticity... I just don't think we should be spending $9 000 USD on a TD 50 to get that kind of tracking. But I dream.

[himalaya]

A physical modelling plugin would be great but just limiting it to drums is boring, please make it do a wider range of sounds.

Extending the parameter range beyond what the acoustic model would dictate would be definitely cool.
But I think this is more suited to percussion sounds rather than a drum kit. For example, it's great to shape cool, imaginary membrane percussion sounds in Chromaphone. Lots of very usable sounds there, but I wonder how it would translate to, say, snares, kicks, toms and hats and such. People usually have a definite idea of how a snare should sound, for example. Whereas there is a lot of scope for new percussive sounds, which appear to be acoustic but new and unheard of before.

But you never know. It's certainly an exciting area to explore. Especially if the dynamic aspect of the sounds are emulated very well, then creating new hybrid sounds with this emulated expressive dynamic response, would be very exciting.

I've been working with drum synthesis a lot lately, mostly using subtractive+FM+additive as a means of obtaining real acoustic sounds. Lots of fun.

Give us more toys, guys!

[justin3am]

I recently played with Korg Wavedrum (the original WD1 with RE1 controller). It's an amazing instrument but much of what makes it interesting is in how it's played. It's so responsive to the slightest brush of the drum head and tweaking parameters while playing results in some very dynamic sounds. However, it seems to me that a lot of that experience might be lost if the physical interface isn't there.

For example, the resonance of the Wavedrum can be dampened by putting your palm on the drum head while striking it. The more pressure you apply with your palm, the less the drum model resonates. And putting your palm in different positions changes the character of the sound.

With that considered, it would be interesting if a VST drum synth would allow alternatives to being triggered via MIDI. That could be a means of sending the output of piezo mics and FSR sensors directly to the instrument as audio/CV, rather than being converted to MIDI first. You would need a multi input audio interface to get the most out of it, of course but then you could then use anything you would like as an exciter for the drum model. I hate to bring it back to the the Eurorack thing, but I wonder if Urs has had a chance to use a piezo (like Ears or Microphonie) to trigger his Rings module. It's not quite a Wavedrum, but it's a lot of fun!