Does windway size affect tone?

Actually, most of this is totally irrelevant, IMO. It is the distance between the blown edge and the bell end that gives you a correctly tuned bell note. Trying to guess exactly where that blown edge will end up (by finding center of hole) when you have actually achieved it is a little like shooting into darkness. If your whistle is to be of a non-tunable design, then once your mouthpiece is finished enough to generate sound (don’t completely finish it, you may want to disassemble and adjust something when voicing.), you simply warm the entire piece and then cut the whistle to length using a tuner. Your tone holes are then laid out from the cut edge. To carry this a step further (literally), and in addressing tunable whistles, my tunable whistles each begin their lives as a non-tunable whistle, and is then made tunable by cutting and adding an appropriate sleeve.


As to using Pete Kosel’s java calculator, embouchre diameter is important. On a high whistle, you will want to use .25" in your calculation, rather than .500". On a lower whistle (say between middle C and low D), using .3750" may work better depending on your ID. Next thing I do when using Pete’s calculator is to drill all of the tone holes under-size. That is, I centerpunch each hole, then start drilling small and step each hole up slowly. When they are all within .0625" of the diameters given by the calculator, I stop drilling and start playing, and voice the whistle for tuning from there, focusing mostly on the upper octave.

This is a simple, but effective and reliable method that I have been using for years now. Probably the best advice for anyone who is serious about making whistles is to keep notes on each one. Over time you will be able to develop a “formula” for each one that produces a good playing and sounding whistle consistently.

To compensate for the Trapezoidal shape of the voicing window you must “average” the diameter before you “plug them in” to Flutomat (or other) calculators.
(La + Wa/2) + (Lb + Wb/2) /2 = Davg
"The average diameter at the bottom of the square window plus the average diameter of the top of the square window, divided by 2 equals Davg " or the Equivalent of a round embouchure. QED

Recorders use the tapered angle of the bore to “shift” the nodes/anti-nodes and “tune” the first 2 registers. The drawback to this is a limited register range (2.5 octaves). Most pennywhistles keep their range and compensate for flat pitch in the upper registers by increasing the windway velocity (more breath pressure) to sharpen/compensate for this “phase shift flattening effect” and maintain correct pitch. This requires the player to have a good “ear” for pitch though. Some whistle makers (Generation) use a tuned resonator cavity, like a flute, to tune interregister harmonics.

Actually, most of this is totally irrelevant, IMO. It is the distance between the blown edge and the bell end that gives you a correctly tuned bell note

If you want to calculate the length of the whistle it is the distance to the center of the hole or window that is used in the calculations.

It is the distance between the blown edge and the bell end,
and …
the window size
wall thickness
bore id
perturbations in the bore (ie covered holes, tuning slides)
air temp
humidity
phase of the moon (not really)
the way the player plays (really)
and … I’m sure I missed something … actually I know I have not included some things, but you get my drift …
that gives you a correctly tuned bell note.

And then you have to worry whether the octave ‘bell note’ is in tune - probably will not be …

Yes Highwood, BUT…there is a “false bore extension” to the aircolumn at the voicing which must be compensated for. This aircolum extention at the voicing is critical to Keynote pitch and sets the location of anti-nodes for tonehole placement. Flatness of pitch multiplies with successive registers.

(D/Eavg)squared X ({2/3} X {E + T} = False Bore Extention.

D = Bore diameter at the voicing
Eavg = The averaged Emboucure/Voicing Diameter
T = The Thickness/Depth of the Embouchure/Voicing

Hmmmmm, I was wondering whether the walls of the voicing window (i.e. the cut edges of the whistle tube) should be cut on a line pointing at the centre of the tube (a radius as it were) giving a hole smaller on the inside of the tube, or cut vertically down giving a constant width hole. Perhaps this choice also matters…

I’m skeptical. According to Arthur Benade, vortex shedding has “only a secondary influence on flute-type sound production.” In any case, a Von Karman vortex street would require interaction between the two streams of air on the other side of the lip: the long body of the instrument prevents any interaction until it is far too late.

I don’t know what the effect is of rounding the lip edge, but I don’t think it reduces the air velocity required.

It may be technically incorrect so say vortex street, since it’s only behavior at the fipple edge that matters. And if Benade means that downstream effects are secondary, that’s probably true, too (I’ve not read Benade yet). But it’s my layman’s understanding that vortex oscillation at the blade is what drives the air reed. So that really seems like a technical quibble, unless the mechanism is actually something completely different.

As for a blunted blade … And trying to get my head around the [u]Strouhal number[/u] from Wiki … It seems that flow velocity is proportional to blade width (effective diameter) for a given frequency. Which implies that a rounded blade needs a higher (not lower) velocity to maintain the same frequency, no? This seems empirically right, too. It’s easier to create a whistle noise by blowing a jet at a card edge than, say, over a cylindrical pencil. And blunting the blade calms a skittish whistle by requiring more breath (velocity) to excite the upper components of the pre-resonant-coupled noise generated at the blade edge. So I suppose this should be testable experimentally.

The response time for a rounded lip labium is faster, ergo, it requires less breath velocity to begin oscillations. The rounded lip creates the V.K. Vortex Street with less,not more breath pressure. I submit the “Flauto Dolce” (quiet flute) organ pipe as my proof. This voicing requires low wind pressure to oscillate.

http://www.organstops.org/f/FlautoDolce.gif

Two things I have learnt from this thread: I have lots more to learn. :smiley:

And two questions to keep the discussion going:

How much do whistle makers rely on the theory, and how much is done by experience/trial-and-error? Hole placement would seem the most exact part of the science of whistle building, but not even that seems mathematically definable. Making the whistle head seems to be far less exact mathematically. Equations are brandished, but are these used merely to understand what happened when a particular change was made? Or maybe the jist of the equation guides us towards a different set of dimensions? Or is there some secret set of formulae for completely defining a whistle and its properties on paper before work starts?

And secondly, am I right in my observation that a shorter voicing window favours the higher notes, while a longer one favours the lower notes? I made a head yesterday with a short window, and the upper notes were very sweet, and easy to play - up to the third OOOOOO, and even the third XXXXXO. This was very unusual for my whistles. The lower notes were very weak, and as I altered the labium edge to strengthen the lower notes, the higher ones got harder and ‘scratchier’. By the time I had finished I had a strong low end and a harsh top end, with a longish window.

Just my thoughts here. I haven’t found the secret formula. Theory is just that, theory. What we make tests the theory. One could take an existing whistle, measure it and knock it off. But there may not be much learned from that exercise. We could make whistles by trial and error and get a great result and learn a lot through that process. That could also be a great deal of fun if you are the curious type. But folks have been making whistles for millennia so why not start with the common knowledge that has been accumulated to this point. We have many whistles made by many makers to play, measure and examine. We have many players to hear and we can talk to them to gain their points of view. And we have the work of the acoustic engineers to review. The acoustic physics is far more exact than my(our?) whistle making skills. But the calculations help test my design concepts. They help me build a first whistle that is “close” to what I intend(assuming my application of the theory is correct). That results in less trash but perhaps fewer practical lessons learned. So I make the first whistle, examine it and make adjustments from there without looking back at the calculations. When I get to a whistle that I like I see what other players think. I might make more adjustments based on their observations. Make 'em and play 'em.

Yes, that’s generally right.

I’d say that makes three things you’ve learnt.

Feadoggie

Phill, the equations I offered were adjustments to functions used in Pete Kosels Flutomat calculator. I arrived at them not because I understood something, but from observations. Much of acoustic theory is derived from observations. I found it necessary to adjust the calculator for it to give more useful results. And it does. This is a pragmatic approach. I use the calculator for hole placements, and it is pretty close to right. The window/embouchure placement calculation is not as precise, but it does not matter so much if the design incorporates a tuning slide. For windway/window/fipple I am not using formulas, but what worked for me. Not quite true: I may calculate windway cross section, as there is some dependency between cross section and flow/back pressure.

As to tone: I have some guiding ideas, but no overall theory.

Regards window length: I think you got the right idea there. I’d say shorter windows are characteristic for recorders, while longer windows are used in whistles.

Thank you, Thomas-Hastay, for the fascinating picture, and the other useful information you’ve given on this thread. A quick experiment suggests that something very interesting happens when you blunt the lip. I wasn’t suggesting this wasn’t the case, but I’m quite certain that whatever is happening isn’t because of a Von Karman vortex street. In designing cylindrical structures, a good way to prevent the formation of a V.K. vortex street is a longtitudinal fin extending down-stream from the cylinder. If you look at the cross-section, that’s exactly what the labium ramp looks like: the blunted lip is absolutely safe from the formation of a V.K. vortex street. The sound of the whistle comes from oscillation of the “air reed” blowing across the window (http://www.phys.unsw.edu.au/jw/fluteacoustics.html). Any vortex phenomena that might be happening at the lip would be much higher in frequency than the oscillation of the air reed. (This means, MTGuru, that you don’t need to get your head around the Strouhal number.)

So something else is happening here. I’d love to find out what.

I highly recommend Benade. “Fundamentals of Musical Acoustics” is still available in a Dover reprint, for $20 or so. Although it isn’t the last word on the subject, it’s well worth the money and reading time for anyone making musical instruments.

Thanks Hans, Feadoggie, you confirm what I suspected on both counts.

Oh dear, how do I know if I have accidentally made a recorder? Is my C&F membership get automatically cancelled? Do I get leniency if it is a really bad recorder?

Seriously, I shortened the window (easier with plastic than with metal or wood). The bias toward the upper register returned. I guess I could have two heads and choose based on the tune I want to play. :smiley:

Someone else who shares my quirky sense of humour…

Seriously, I shortened the window (easier with plastic than with metal or wood). The bias toward the upper register returned. I guess I could have two heads and choose based on the tune I want to play.

Maybe you could design a whistle with an adjustable window height - there are whistles with adjustable window width. In fact I have given this serious thought - not for a performance whistle, but as a learning tool - but have not put into practice.

theory v practice - obviously practice is what matters, unfortunately (or maybe fortunately) there is not enough time in even many lifetimes to try out all combinations in practice. Theory is great to point you in the right direction - just don’t let theory stop real progress because it may be wrong(!) or at least simplified - and the interesting stuff is almost always in the details (IM not so HO).

hole position can be predicted well with theory - which does not mean there is one answer - you can optimize for different things there is most certainly no right answer, of course there are wrong answers!

head design is a whole different game - I am not aware of much theory that will help out. It is all about practice and perhaps learning from the lengthy tradition of organ pipes, though organ pipe windows do not have to work over two octaves!

as they say just my 2 cents YMMV and most certainly will