New style whistle heads

[stringbed]

If I understand correctly, in some of these you are saying that the entire inner face of the block slopes, with the top (under the windway) starting before the end of the windway, but with the bottom of the block protruding significantly into the windway. An extreme and steep chamfer, if you will.

Correct. On all my Shaw whistles, the surface of the block that forms the floor of the windway ends just a smidge short of the end of its ceiling. The surface of the block that delimits the upper end of the air column is not perpendicular to its axis and slopes to a point opposite the middle and lower end of the window.

I've wondered about that approach, from the perspective of what happens to the returning pressure wave front as it approaches the back of the stopper. Just slams into it, and then looks around for a way out of here? "Hmmm, I'm going to have to shove that pesky air jet out of my way..."

The air reed and air column form a coherent oscillating system. I can’t see how the one component can, much less need to, exert force counter to the other. Additionally, the Shaw whistles are strongly tapered. On the high D one, the area of the window is 25% larger than that of the bell. If the vibrating air column were detrimentally constrained by the upper aperture, it would be even more so by the lower one. We would then be looking at a good reason for reducing the taper. For reference, the area of the window on a comparable but cylindrical Generation is 5% larger than than the cross-sectional area of the main bore.

What would happen if we gave it some help and guidance? A slope at the back of the block as I think you've described to redirect it upwards? Or would it be better to offer it a curved (concave) slope to coax it around the corner and help it redirect its energies into shoving airjets rather than generating heat? Or am I fantasising and anthropomorphising?

Many musical instruments take offense at being anthropomorphized (ask any so-called “baby grand” piano) but we should be able to avoid that risk here. And rather than speculate about the internal aerodynamics of a whistle perhaps we can stipulate a workshop rule of thumb that is amenable to informal assessment. The one here would be that any detail in a whistle’s structure that may potentially cause undesirable turbulence is worth trial modification. This is well instantiated by the suggestion that:

We could probably test these questions by sticking shaped block extensions into the back of flat-blocked heads.....

[Terry McGee]

I have been holding my breath, and realise that it may be the wrong thing to say in the company of craftsmen and engineers, but would a 3D printer have roll here. It might not make finely tuned whistle heads but one change at a time would be possible. Results could be hooked up to the blowing machine.

I reckon once we can crack whistle (and flute) computer modelling, 3D printing should have great potential at least in prototype development. But I don't see the likes of me getting that far. We have to leave something for the next generation to do!

Interesting to look back on the last 50 years. When I started making, there were no wooden flutes being made apart from those by a few baroque-flute makers, and just a few whistle makers, all well-established names. Look now at the C&F lists of flute and whistle makers!

Pretty much all of what has gone before was achieved through experimentation and selection - acoustic theory has not played a big role. That's where I'd be looking for the next big developments.

[MadmanWithaWhistle]

I reckon once we can crack whistle (and flute) computer modelling, 3D printing should have great potential at least in prototype development.

Do you mean modelling as in CAD, or modeling as in simulating fluid dynamics? Because I do the first, and I know the 2nd is possible... I just can't figure out how to get my model into this program: https://sim-flow.com/

All of my whistleheads start out 3D printed, and I finish the windway floor, ceiling, exit face, and airblade with needle files and sandpaper. Less cleanup / voicing is required on resin printed whistles, but adjusting the critical surfaces and edges is still essential. I don't resin print production models simply because there's no way to make them look nice/maintain durability.

At this point I bet it's quicker to feed sound samples and a CT scan of the instrument into a machine learning program (aka AI) and have it predict what a flute producing a desired tone would sound like. It may not give the causal information regarding the impact of the design features, though. I don't know enough about ML to say whether it's possible to look 'under the hood' for a program like that.