Yes, Roland use physical modelling on their flagship V-Drum modules, but it's a harder problem than you might think. High-quality sensors for electronic drum kits sense not only impact velocity, but position - a cymbal will sound radically different depending on where you hit it. Every part of a drum kit will sympathetically resonate with every other part. It's all massively complicated to model and it only has to be slightly off to feel wrong to a good drummer. Despite decades of serious technical effort, there's still a big gap between the real deal and digital, particularly for drummers with a delicate style who make use of brushes, rutes or soft mallets.
Agreed. Drummer and programmer here. I consider electronic drums to be a different instrument from acoustic drums---especially cymbals---with somewhat comparable musical functions. The advantage of e-drums is that the sound guy can turn the volume down to zero no matter how hard I'm hitting, and that's helpful in places like churches or my apartment. The other advantage is that toms (i.e., the tom samples) are always perfectly and consistently tuned. Apart from those advantages, there's far less traditional drum expressiveness in e-drums (for traditional genres like jazz and rock). The flip side is: with e-drums I can hit the cymbal and make a laser noise.
With more effort I expect that eventually we will be able to model physical instruments just as good as the real thing. However it is an open question if that is good. Sometimes along the way we discover how to make something better. Of course sound is subjective and so no everyone will agree it is better.
There are two distinct problems - modelling a drum kit, and modelling a drum kit with millisecond latency in a way that is accessible and affordable for non-technical musicians. The latter requires a lot more engineering compromises than the former.
Cymbals aren't like speakers, microphones, or rooms, whose inputs are audio and can therefore be modelled with IR via an impulse input. (Well, you can model a cymbal with IR and use it as a reverberator, like a speaker.)
AFAIK, what you're talking about would probably fall into the category of physical modelling, and there's quite a bit of development happening there currently.
Yeah, there is so much more than just hitting these cymbals. How and where and how hard. Anyone who strikes one is not likely to ever forget the experience and it takes quite a while to even scratch the surface of how many different sounds can be produced with each model using various techniques.
I used to crawl around drum sets when I was a baby as my father was a professional musician. I mounted the set as soon as I could and it was the only instrument I was was ever really interested in. Alas, I was heavily discouraged from pursuing music as a child by my mother. She blamed our shit life on my father and his obsession with music and he basically quit after the divorce and went into entrepreneurship (funded by my grandfather's GE stock) to pay our child support. I digress.
>> She blamed our shit life on my father and his obsession with music and he basically quit after the divorce and went into entrepreneurship (funded by my grandfather's GE stock) to pay our child support.
Sounds like there was some truth in what your mom said?
Percussion synthesis is a deep and wonderful practice within electronic music, especially techno and especially in its more experimental manifestations. I’ve never encountered anything specifically claiming to emulate specific cymbals but I’m sure it’s out there galore. In any case, cymbal and hihat synthesis is a classic task in techno production.
A cymbal is not even close to a 1 dimensional sound. Not only do you have intensity of strike, but you have the part of the stick you struck with, your grip looseness, the area of the cymbal you hit, the material of the stick, the material of the cymbal, etc etc etc.
Simulating the sound of a cymbal to high fidelity is equivalent to simulating the physics of the atoms making up the cymbal.
If you want to hand wave away the complexity, sure, Nintendo regularly simulates the sound of a cymbal with a white noise generator modulated by an intensity envelope.
There are several ways to do that, from Karplus-Strong or FM synthesis all the way to solving differential equations (usually with the finite difference method), the more work is involved the more closed is the resulting product. And it's very hard to get it right.
ie impulse response hit depending on intensity scaled 0 to 1 > makes the sound