I’m working on developing a new module that is a physics-based flute module. You can change things like the exciter, tube dimensions, bore, and tune the blowing noise and chiff sounds. So far so good, but the issue is that when you play it, it is quite difficult to play, perhaps like a real flute.
So, now I’d like a second opinion. or three. ha ha.
The instrument has 2 v/oct inputs. Yeah 2!
First you can set the pipe length, and second you can set the fingering.
The idea is that the Finger input can open a hole in the flute with v/oct control. If you finger a note that’s above the range of the flute you can overblow to play them and they sound different. BUT you just cannot play anything lower than the fundamental of the tube. It will output some groaning unstable sounds if you go there. This results in your melodies getting baselined at the fundamental.
A compromise might be to either allow the flute to play lower notes by automatically increasing the length, or by automatically transposing things out of range back into range of the instrument. However, I feel that either of these approaches just result in the instrument becoming more boring.
You can always drive it directly with v/oct of the pipe length, but this is then like a pan pipe, where you don’t get much character difference between notes.
I think this will become a future free module for CVfunk.
What do you think? I will post a demo module to the nightly once it’s a bit more tame!
I think making this a free module is a great idea! The prospects of flute physical modeling for EVERYONE and not just owners of a swam plugin are very bright, another thing to play with patching into effects and battalion when that comes… hoping you can expand the reach to have more unusual woodwind timbres alongside more traditional flute sounds
Oh wow. I did not know of this controller, yeah that would be really neat for sure!
Looks like WARBL is just a MIDI device? So you could totally patch it to the Breath input and use that, I bet it would work nice. There are quite a few expression sliders that could be patched, so that controller would be quite awesome actually.
yep, it’s highly configurable. You can define custom fingerings and there’s quite a few that it comes with, for emulating various traditional pipes. Can do MPE: has note, pitch bend, breath (breath cc or MPE pressure), pitch, yaw. Quite a versatile controller in a tiny and affordable package. Pairs nicely with an EaganMatrix Micro, where people have been actively designing new patches.
OMG - the Warbl 2 is amazing, and it works with the Eagenmatrix Micro! I have always wanted to use the Eagenmatrix, but I have zero keyboard skills, so the Haken Continuum was a total non-starter, even if I could afford one.
The price of the Warbl 2 plus the Eagenmatrix Micro is totally reasonable!
Thank you for posting this. I believe I now know where some of my Venom earnings are going!
SWAM has a new-ish flute thingie which I have, and quite like. In that one you have to choose whether pitch is altered by changing the tube length slide whistle-stylee, moving to a different fixed length tube like in pan flutes, or opening holes like most flutes. Allowing more than one of those at once seems really neat!
In the digital world you can get accurate tuning at any altered length etc etc, which can’t be realistically achieved on an acoustic one. Seems fun!
If I may offer another point of input from experience of working with the SWAM one: Having a separate trig input for the chiff sounds (ideally with a strength control as well) would be great. Retriggering those during a sounding note for flutter- or double-tongue style playing is fun!
I’m all about physical modeling, and the physics of a resonant tube is fascinating stuff. I will definitely be following your progress on this project.
holy moly!
Thanks for sharing that. That glissando headjoint is just wow. Very neat stuff!
The module is going to be called Aulos. And in my research I found that the Aulos originally had bronze sleeves over the tubes so you can adjust their lengths. I liked the idea of building a really basic flute model where you can modulate a lot of the physical parameters and get interesting timbres.
So like the ancient Aulos, this design has L/R flutes that you can detune from each other by up to an octave. And it has a button so you can set the 2nd flute to either drone or follow the main one. The idea is that the two flutes could cover for each other and let you play over a larger range.
I also added the ability to morph between a short open ended pipe and a 2x longer pipe that’s closed ended. And then I ran into a design edge.. there’s just no way to morph between the flute and clarinet without having some kind of pitch-shift in-between region. Also, the resonances change a lot, and I keep finding places where the instrument wants to blow up because of some funny energy build up I didn’t expect.
I’ll provide some more images soon. The test panel has no labels yet, so it’s not super helpful! It’s like parameter city right now… Some of the knobs might have to go into the context menu eventually (attack and release times I think will go away)
This project sounds amazing - I will be very interested in the results, doubly so if I get the WARBL!
Very interesting! I presume (hope) you are altering the harmonic content - odd and even for open, only odd for closed.
I enjoy playing overtone flutes (narrow bore with no finger holes) where increasing velocity/pressure marches up the harmonic series. My flutes are constructed so I can keep the end open or closed with my palm or finger (depending on bore size). It is fun to play with the different series of the open vs closed tube. I especially enjoy my contra-bass overtone flute with a fundamental of B below the cello’s low C. Partially closing the end allows bending of notes. There are lots of instabilities that give an interesting sound - It would be amazing if you manage to model this in a convincing way.
Another potential thing to explore are the percussive sounds of the tubulum, like played by the Blue Man group. That is probably the most exciting performance technique with my contra-bass overtone flute - combining the tubulum sound with the overtone flute sound by banging on the end of the flute with my palm.
One thing I like to do is sometimes sing into my flutes, the interaction of the voice and flute giving a really cool timbre. If you could model this somehow, it would be cool if you provided an audio input to the flute.
Rather than a single “do it all” module, maybe this could evolve into a series of modules, each with its own features and strengths?
Yes exactly! It uses an internal waveguide to reflect the waves and cause the interference. So the intention is to control the even/odd harmonics. Some of the sounds are convincing, like when overblown, and many are not (my reed sound)… it’s very difficult to get this right. I appreciate your explanations, the dynamics interaction are a big part of how the sound could be made more convincingly. The ‘Damp’ control adds a low pass filter into the chain and that dampens the whole system. There’s also a ‘Lip’ control that essentially is like how tight your lip is onto the resonator.
So ‘Bore’ is currently setup to morph between open-end tube and closed-end conical tube. It’s very expressive, but at the mid-point there’s a total dead-zone where the sign inverts. So I’m struggling with how to deal with that.
Perhaps I’ll explore linking the bore-size to the reed/flute exciter, and then having this control be more specifically simulating putting your palm over the end.
The previous prototype has an audio input port, but it was not very impressive sounding in my initial testing… as the little flute resonator is very tiny. But I didn’t explore it much, as yes you can get interferences and things. I can add it back!
This seems great for your plug and the rack, I think creative digital modeling of physical resonant objects is an area we could have more voices and designs in for sure.
Alloy is really cool karplus style resonator, looking forward to the reverb release and this foray into “flute” is fabulous.
I love the sound of many interesting handheld percussion instruments, we have great samples but modeling is often more exciting in its applications and modulations. I would love to see a little orchestra of percussion avaialble that is modeling based ( cabasa, maracas, vibraslap, flexatone, guiro, cuica, claves, wood blocks, castanets, tambourine, chimes). Hopefully someone will get interested in these simple little sounds that could be fun to model.
Looking forward with much interest in the flute module.
Brass modeling is definitely another necessity for VCV or anywhere really, but we REALLY need physical models of electric guitars and especially clean ones so I can put guitar effects on them without needing a real guitar everywhere which may not fit all the time or gigabytes of sample libraries
Now it’s mostly working, but I feel like it could be more sweet spot and less dead zone.. so now begins a process of tuning the sliders in fun non-linear ways to make it more dialed in.
I added back the external input on the suggestion of @DaveVenom , but now with a CV and gain.
I should probably try to explain how the model works.
We have a pair of flutes, that you can control as one. They can be either in unison, or detuned up to an octave apart. And the second pipe can be set to drone, follow, or pair.
The flute model is build around two coupled waveguides that represent the air column. The delay length is set by the ratio of the pipe pitch to fingering pitch, so that the model can continuously vary the active acoustic length in a fixed physical tube length.
You can excite the system with a blend of two different models, the first is a reed-style pressure source consisting of breath energy, turbulence noise, and saturation. The second is a flute-style jet model where delayed acoustic feedback from the bore is passed through a cubic jet nonlinearity (y=x-x^3/3).
The flute design uses several nonlinear stabilization stages inside the feedback path. It has amplitude-dependent ADAA tanh soft-compression limits runaway oscillations, and energy-monitoring system adds dynamic damping when the resonator gets out of control.
Transient attacks are enhanced with short-duration noise burst (Chiff), while continuous breath (noise) provides turbulent excitation. The final output is the sum of primary and secondary resonators, followed by DC removal and anti-alias filtering.
