I was having a discussion on here maybe a month or so ago about hi hat synthesis, which a number of people, myself included, seem to be frustrated by. I’d love to hear how people approach it.
What I’ve seen normally is something like this:
Generator(s) → Filter(s) → Env (short A, relatively short R)
Where generator(s) are noise and/or a number of relatively high frequency oscillators - usually sine or square and filter(s) are either or both in some order of HPF/BPF. Occassionally I’ll see FM/AM/RM instead of some kind of additive approach to the generators, but it’s normally still this general signal flow. Often comb filtering will happen toward the end of the signal flow as well.
I’ve also seen far too many posts (not here) saying “open hat is just closed hat with a longer tail” but I’m really not satisfied with that answer. It might be related to spectral components and decay, but there’s a timbral difference between open and closed hi hat. Putting a longer tail on the envelope isn’t going to get there. I’ve also yet to see anything about pedal hat synthesis.
What I have come across involves a lot of weeding through bad tutorials on “how to make a closed hi hat with FM8” or “how to improve your MIDI sequencing in Logic to get more realistic hi hats because this is how they work in the real world and you’ve maybe never seen a hi hat before”
Have you tried a very quick downward sweep of a sine osc as excitation
signal of a resonator, possibly followed by AM to get more
inharmonicity, with bandpass filtered noise added?
What I have come across involves a lot of weeding through bad tutorials on “how to make a closed hi hat with FM8” or “how to improve your MIDI sequencing in Logic to get more realistic hi hats because this is how they work in the real world and you’ve maybe never seen a hi hat before”
Yes, that’s the internet.
For generating knowledge yourself, rather than relying on search engine
results, you might wanna try reading something like Andy Farnell’s
“Designing Sound”, which, although it doesn’t cover HiHats, will give
you a good general understanding how to synthesize physical objects.
And have you checked the schematics of analogue drum machines? A quick
startpage search
yields this peer-reviewed paper, which could give you a first hint
albeit being about the cymbals:
The TR808/909 service manuals have some incredibly detailed schematics I’ve been looking at.
I guess part of what I still am intrigued by is the difference between open/closed/pedal
Yoshinosuke Horiuchi made a very convincing SC 808 replica a few years ago, I see it got a bunch of new features now: SC-808, the SuperCollider TR-808 recreation, now has an Advanced version - CDM Create Digital Music. I read a while ago that he went pretty deep into the schematics of the 808. I just took a look at the code - he uses 6 LFPulse Ugens. The open and closed hats are basically the same with different decay times and slightly different filtering.
The synthdefs aren’t good, they are exceptional. It is a treasure trove of perc techniques. The clap, snare, hi hat. I would have never figured these out. The code is not great sc code, but can easily be fixed up to your liking.
The quirky thing about the SC-808 synthdefs is that they all use envelopes with really long release times, but with very steep curves. For example, the low tom has a release time of 20 seconds, with a curve of -250. I would have never thought to do that, but it actually sounds pretty great. The problem is that synths start to pile up on the server very quickly, so I ended up changing the release time to 1 sec, with a curve of -12.5, and it sounds indistinguishable to my ears.
One needs a nicely tuned system to handle that, I was wondering why the computer was getting hot and the sound cracking while I played around with those SDs