Tweaking instrument production with software

[graybach]

https://share.google/GawrUd0MqiuGMIOda

Interesting. I wonder if the same thing could be done,
(?or is already being done?) with instruments in the tuba family.

[UncleBeer]

I've posted before about the Brass Instrument Analysis System (BIAS), but this is what you're thinking of. https://www.artim.at/en/bias-for-brasses_basic/ Not only will it objectively diagnose an instrument's playing characteristics, it can also do physical modeling, predicting how specific changes will change an instrument. https://www.artim.at/en/bias-for-brasses_extended/ Also works on woodwinds and strings (with additional components).

[daktx2]

There's also an open source instrument design software for Python available, developed out of a lab in France, called openwind. It doesn't have the cool hardware that BIAS has or even a UI, definitely less approachable, but does have the ability to predict intonation characteristics given a geometry.

https://openwind.inria.fr/

[bloke]

Don't let this response pull the thread off topic but I've noticed (and I'm sure others have noticed) that with huge brass instruments (tubas) it takes the larger outer bows much longer to warm up to playing temperature, whereby if the player plays constantly for five or 10 minutes with their hands and body close to the instrument, some flat pitches will pull themselves up to In tune once the air is warm in the far end of the instrument.

That's one of the challenges to playing in a symphony orchestra, is that so often we sit there with our instrument for twenty minutes and then play like a maniac for about two and a half minutes, whereby the first two minutes of that two and a half minutes is dealing with flat pitches - due to the instrument being cold on the far end of the instrument. (nodes and such way out on that end). me? I'm usually setting the tuba in my lap in such situations and hovering over it with my fat self.

Back on topic, I would assume that these design programs are based on the entire instrument being at an even temperature, yes?

[UncleBeer]

with huge brass instruments (tubas) it takes the larger outer bows much longer to warm up to playing temperature, whereby if the player plays constantly for five or 10 minutes with their hands and body close to the instrument, some flat pitches will pull themselves up to In tune once the air is warm in the far end of the instrument.

I'd wager that even after playing awhile, only the first two or three feet of the tuba warm up to any appreciable extent. So little air actually moves through the instrument (despite what pedagogues try to convince you).

I think the coldest component in the described scenario is probably the player's brain.

[bloke]

I wasn't claiming that it moves much; it's just that (I'm fine with agreeing with it being mostly static) air in the last few larger diameter feet of a tuba (certainly compared to a trumpet or trombone) sort of remain colder until a player has been playing a tuba for several minutes... Certainly this pertains to really large tubas.

[daktx2]

@bloke, the open source software definitely makes a constant temperature assumption. My training is in a different kind of mathematics, but taking a glance at how the math/software works, extending it to account for different temperatures at different parts of the tube wouldn't be too hard, could probably be done fairly quickly with the assistance of Claude Code / another LLM.

I wonder if it will show in theory what I experience (anecdotally) in reality, namely that different partials get warped differently by a cold horn. Might be a fun evening project for me sometime, I will report back.

[bloke]

@bloke, the open source software definitely makes a constant temperature assumption. My training is in a different kind of mathematics, but taking a glance at how the math/software works, extending it to account for different temperatures at different parts of the tube wouldn't be too hard, could probably be done fairly quickly with the assistance of Claude Code / another LLM.

I wonder if it will show in theory what I experience (anecdotally) in reality, namely that different partials get warped differently by a cold horn. Might be a fun evening project for me sometime, I will report back.

My Miraphone model 98 is the closest to In tune huge tuba I've ever played. Of course the fourth partial 2-3 F sharp needs help (just as that corresponding pitch needs help on virtually every brass instrument in existence) but I have a really easily accessible trigger for that. The third partial F starts out about 5 to 8 cents flat, but - once the entire instrument is warm - it's actually 1 or 2 cents sharp (in addition to my tendency to help that pitch, because it takes a while to warm the instrument up, so I find myself helping that F).

I feel certain that they used some of this type of software to design that instrument. It just turned out too well to have not.

[Nemo]

There's a guy working on a physically-based audio simulator to recreate engine sounds (primarily for racing simulations, but possibly for electric sports cars that currently use tinny recordings). It includes air temperature and density as those are obviously pretty important to the way an engine sounds. The simulation also works on wind instruments in theory and can produce some convincing results if you give it the right geometry. I can foresee something like this being used for rapidly tweaking instrument design for characteristics like intonation and resonance without needing to produce physical prototypes for every minor change

https://www.youtube.com/@AngeTheGreat/videos