Blowing machine

The Ultimate On-Line Whistle Community. If you find one more ultimater, let us know.
User avatar
Terry McGee
Posts: 3330
Joined: Sun Dec 12, 2004 4:12 pm
Please enter the next number in sequence: 1
Location: Malua Bay, on the NSW Nature Coast
Contact:

Re: Blowing machine

Post by Terry McGee »

Tapered window is 9.43 wide by 5.94 long. Foot diameter is 12.15 (having flared a bit from a minimum of 11.36mm). The "Equivalent Diameter" is 13.25, so a little smaller than the Cylindrical whistle.

Highly Tweaked Soodlums Mellow D is 8.6 by 5.72 long. Bore is cylindrical at 13.7 all the way.

Their windways are very similar in height from 2 down to ~ 1.28. Width of Tapered is bigger at the blowing end, 9.65 compared to 8.66.

Nobody knows as much about loudness as broadcasters. "Charlie, there's a lady on Line 1 that wants to know why the ads are always louder than the music."

In the old days we had the VU (Volume Units) meter to help gauge such matters. They were made intentionally a bit sluggish so that they ignored narrow peaks and went with the flow. You needed a burst of 0VU signal lasting 300mSec to show 0VU.

Then the BBC invented the Peak Program Meter with an integration time of around 10mS. Really good at preventing you overloading the tape or transmitter or whatever you were feeding, but hopeless for gauging loudness. You had to set the music to one level and chat to another.

Finally the European Broadcasting Union spat the dummy and came up with the Loudness meter. It started as a hardware device but has now evolved into software. It's technical specification is introduced at https://en.wikipedia.org/wiki/EBU_R_128

You can now even get free on-line versions, eg: https://youlean.co/online-loudness-meter/

I've used that system to quantify problems with bell clappering. Normal sound level meters are fooled by impulsive, distorted sounds, but the Loudness Meter can hear through the artefacts and react in the same way we do.

Now that we have access to such meaningful metering, why is YouTube now worse than ever at setting audio levels?
Tunborough
Posts: 1419
Joined: Sun Dec 05, 2010 2:59 pm
antispam: No
Please enter the next number in sequence: 10
Location: Southwestern Ontario

Re: Blowing machine

Post by Tunborough »

trill wrote: Thu Apr 13, 2023 6:47 pm For a whistle, what are X, R ?

I'm trying to visualize how the ratio X/R (reactance/resistance) looks.

For a whistle, how is R calculated ? How is X ?
I haven't forgotten your questions. They are quite reasonable, but I don't feel equipped to give a clean explanation. Here's some fundamentals ...

Acoustic impedance Z is the ratio of pressure divided by flow, P/Q, in the frequency domain. If the pressure wave and the flow wave are in phase, the impedance is purely resistive, Z = R. But in whistles, pressure and flow aren't necessarily in phase, so the impedance is a complex number, Z = R + iX. The impedance can also be expressed in polar form, as a magnitude and a phase angle. The magnitude is sqrt(R**2 + X**2), and the phase angle is arctan(X/R). If you're still with me, I'll give some examples when I have time to pull them together.

BTW, I am wondering if the polar form might give more useful results, provided we can figure out that bend around St = 0.3.
trill
Posts: 685
Joined: Thu Aug 17, 2006 8:44 pm

Re: Blowing machine

Post by trill »

Terry McGee wrote: Sun Apr 16, 2023 4:16 am Tapered window is 9.43 wide by 5.94 long. . . Nobody knows as much about loudness as broadcasters. . .
Thank you for the whistle measures + explanation of "loudness".

I asked for the measures because I was wondering about the "available area" for sound emission from the window + bottom (forgetting the holes for a moment).

I keep trying to visualize all the harmonics running around in the tube: all these pressure+displacement waves running over each other. Well, because they are "free-free" modes, they must all synch up at the ends, right ? (in the perfect/simplified world of HS physics, right ?!). I'm particularly focused on the window + jet, flickering up+down. I'm just glad it all works.

Terry, do you happen to know if the Audacity spectrum is the the real part or the magnitude of Re+Im ? I may ask the Audacity forum.
User avatar
Terry McGee
Posts: 3330
Joined: Sun Dec 12, 2004 4:12 pm
Please enter the next number in sequence: 1
Location: Malua Bay, on the NSW Nature Coast
Contact:

Re: Blowing machine

Post by Terry McGee »

trill wrote: Mon Apr 17, 2023 10:22 pm I keep trying to visualize all the harmonics running around in the tube: all these pressure+displacement waves running over each other. Well, because they are "free-free" modes, they must all synch up at the ends, right ? (in the perfect/simplified world of HS physics, right ?!). I'm particularly focused on the window + jet, flickering up+down. I'm just glad it all works.
Yes. Good to remember that, because whistles and flutes are "open" at each end, the longest wavelength it can resonate at is a half-wave. So if we start with a pressure front running down the tube, it's followed by a low-pressure front running back up, pressure front down, low-pressure front up, etc.

So that gives the fundamental, say D5 if we go for xxx xxx. The next partial is a quarter wave - an octave higher, D6. Then a sixth, three times higher than the fundamental, A6. And then an eighth of a wavelength, giving us 2 octaves or D7 in our example. And these can all happily coexist, as we see when we look at a spectrum of xxx xxx.

And here's a tricky one. Imagine you have a really badly tuned whistle, so that when you play the partials mentioned above separately, they come out all over the place. If you then play the lowest note, and look at the spectrum hoping to see all these horribly tuned partials exposed for their disruptive proclivities, you will be totally disappointed. They will all measure spot on! Because the jet-switching action realigns them at the start of each cycle.

So tuning doesn't matter, you venture hopefully? No, it does. Although the jet-switching action appears to sort everything out, badly tuned partials continue to wreak havoc by weakening the overall response. Efficiency goes down. And I reckon we can hear it as a souring of the notes. A good example of this would be the Renaissance Flute. Just a cylinder, and a rather long one. Sounds fine on the low notes, but as you make your way up the second octave, the notes are going flatter and flatter. And the whole thing sounds a bit mournful. We solved the problem during the baroque era by adopting tapered bodies. And then Boehm resolved it by using tapered heads.

So if the tuning of all the partials is good, the whistle or flute will be sweet and powerful. Because all the partial fronts will arrive back in synch to optimise the jet-switching action, not all over the place to weaken it.
Terry, do you happen to know if the Audacity spectrum is the the real part or the magnitude of Re+Im ? I may ask the Audacity forum.
Sorry, above my pay grade!
User avatar
Terry McGee
Posts: 3330
Joined: Sun Dec 12, 2004 4:12 pm
Please enter the next number in sequence: 1
Location: Malua Bay, on the NSW Nature Coast
Contact:

Re: Blowing machine

Post by Terry McGee »

Now, remember I ordered a Clarke Sweetone whistle from a music shop here in Australia. It lobbed in today.

My aim was to get a bit more experience with tapered whistles. Because my only experience was with a Clarke C whistle (and that best-of-luck windway and edge makes that a dodgy experience) 50 years ago (see also, "scarred for life"), and currently with my tapered "found object" pic-whistle.

And remember I predicted two bleating responses?
1. How dare they make a better whistle than I can for only AUD $22! Or,
2. These guys have been making whistles since 1843 and that's the best they can do?

Relieved to find it wasn't No 1. Indeed, it's closer to No 2, except it's not quite that simple.

Received wisdom states that a tapered whistle (or flute) will have its holes closer together than a cylindrical whistle. OK, cop this. Spacing between the centres of holes 1 and 6 on the following whistles:

Killarney cylindrical, 108mm
Highly Tweaked Mellow D cylindrical, 108mm

Terry's tapered pic-whistle, 100mm
Clarke tapered Sweetone, 109mm.

What the? It's supposed to be less than the cylindricals, but it's a smidge more? While my pic-whistle is significantly less.

And the holes are really small, varying between 4.6 and 6.2 (holes 1 and 5). Compare with 4.75 and 7.63 on the Killarney. And very much more on the other two. We can expect small holes at the bottom end where the bore diameter is small. But we'd expect bigger holes at the top end where the diameter is larger than the cylindrical.

Tuning isn't great, and performance is weak. And I'd say "dull" rather than "sweet". And some notes have "shadings". They don't sound clearly.

The good news is that it's tuneable. Not glued. I was worried about that, but turns out for no reason.

So, what have I learned? So far I like them in this order:
Terry's tapered pic-whistle (sonorous, authoritative, but perhaps lacking a bit in the bottom end?)
Highly Tweaked Mellow D cylindrical ( a bit "sizzley"?)
Killarney cylindrical (sweet but too quiet)
Clarke tapered Sweetone (dull, weak)

But hey, I'm a power-crazed flute player, so I would say that, wouldn't I. But the scary thing is it's inviting me to go forward with the pic-whistle, not retreat into safety. Woah! Reason with me, quick! If this all comes unstuck, it's your fault!
Tunborough
Posts: 1419
Joined: Sun Dec 05, 2010 2:59 pm
antispam: No
Please enter the next number in sequence: 10
Location: Southwestern Ontario

Re: Blowing machine

Post by Tunborough »

trill wrote: Mon Apr 17, 2023 10:22 pm Terry, do you happen to know if the Audacity spectrum is the the real part or the magnitude of Re+Im ? I may ask the Audacity forum.
Power is a scalar quantity, it doesn't have a phase, so we use a real number for it, not a complex number.
Terry McGee wrote: Tue Apr 18, 2023 4:46 am [The hole spacing on a tapered whistle is] supposed to be less than the cylindricals, but it's a smidge more? While my pic-whistle is significantly less.

And the holes are really small ...
That's the trade-off. With a tapered bore, you can have the same sized holes closer together, or you can have smaller holes with the same spacing, or something in between. Clarke has chosen to go with smaller holes. I can see an advantage with that: your fingers don't have to learn a dramatically new layout when switching between whistles.
Terry McGee wrote: Tue Apr 18, 2023 2:58 am So that gives the fundamental, say D5 if we go for xxx xxx. The next partial is a quarter wave - an octave higher, D6. Then a sixth, three times higher than the fundamental, A6. And then an eighth of a wavelength, giving us 2 octaves or D7 in our example.
To be clear, at D5, the whistle bore holds a half-wavelength standing wave. At D6, the standing wave is a full wavelength. At A6, it is 1.5 wavelengths, and at D7 it is two wavelengths.
User avatar
Terry McGee
Posts: 3330
Joined: Sun Dec 12, 2004 4:12 pm
Please enter the next number in sequence: 1
Location: Malua Bay, on the NSW Nature Coast
Contact:

Re: Blowing machine

Post by Terry McGee »

Tunborough wrote: Tue Apr 18, 2023 8:21 am With a tapered bore, you can have the same sized holes closer together, or you can have smaller holes with the same spacing, or something in between. Clarke has chosen to go with smaller holes. I can see an advantage with that: your fingers don't have to learn a dramatically new layout when switching between whistles.
Heh heh, I've certainly noted that when I pick up the Pic-whistle, I have to feel for the holes. Posh folk would call it "recalibrating". But once done, I'm fine. Now I'm finding I have to do it with the Clarke as well! But the catch is I'm not getting the reward.
To be clear, at D5, the whistle bore holds a half-wavelength standing wave. At D6, the standing wave is a full wavelength. At A6, it is 1.5 wavelengths, and at D7 it is two wavelengths.
Thank you for saying what I wanted to say!
User avatar
Terry McGee
Posts: 3330
Joined: Sun Dec 12, 2004 4:12 pm
Please enter the next number in sequence: 1
Location: Malua Bay, on the NSW Nature Coast
Contact:

Re: Blowing machine

Post by Terry McGee »

So, made a new head for the tapered pic-whistle today. In Delrin, based on the Sindt approach. Looks better already, not being pink.

Aimed first at a 1.25mm ceiling height in the windway, with the same windway width as the pink Bb head. As expected, too much resistance. Which registered as a bit weak in the low octave, and needing to push too hard for the second. So I dug into the top of the stopper so that I'm almost overblowing at my normal pressure at the bottom of the first octave. I've possibly gone a bit far, as I'm getting messages from the rest of the family. It's "an outside whistle".

The good thing about digging in to the top of the stopper is that the stopper is the most easily replaced part. So I can easily make a new one and pull back a bit on the voicing.

The Sindt design also allows me to experiment with pushing the stopper out a bit into the window space, or pulling it back a bit, which feature we noted with the Bb head. You can't go very far either way without problems arising.

I thought it might be interesting to map the pressures/flows at several fingerings between tapered whistles and cylindrical whistles. But I'd better get a few flutes moving first....
Tunborough
Posts: 1419
Joined: Sun Dec 05, 2010 2:59 pm
antispam: No
Please enter the next number in sequence: 10
Location: Southwestern Ontario

Re: Blowing machine

Post by Tunborough »

trill wrote: Thu Apr 13, 2023 6:47 pm For a whistle, what are X, R ?

I'm trying to visualize how the ratio X/R (reactance/resistance) looks.
This graph shows the impedance of an open-ended cylindrical tube, 266 mm long and 12 mm in diameter, relative to the characteristic impedance of the tube. The blue line is R, the red line is X.

Image

The resonances of the tube occur where the imaginary part (X) crosses the x axis around 640 and 1280 Hz.

For comparison, here is an open ended tapered tube, 12 mm in diameter at the input end, 10 mm in diameter at the output.

Image
trill
Posts: 685
Joined: Thu Aug 17, 2006 8:44 pm

Re: Blowing machine

Post by trill »

Tunborough wrote: Tue Apr 25, 2023 11:15 am This graph shows the impedance of an open-ended cylindrical tube, 266 mm long and 12 mm in diameter, relative to the characteristic impedance of the tube . . .
A few questions:

1) what algorithm is used to calculate these ?

2) what is the name of the software used ?

Also, could you post the data ? I'd like to do a close comparison.

Thanks very much.
User avatar
kkrell
Posts: 4831
Joined: Mon Jul 29, 2002 6:00 pm
antispam: No
Please enter the next number in sequence: 8
Tell us something.: Mostly producer of the Wooden Flute Obsession 3-volume 6-CD 7-hour set of mostly player's choice of Irish tunes, played mostly solo, on mostly wooden flutes by approximately 120 different mostly highly-rated traditional flute players & are mostly...
Location: Los Angeles
Contact:

Re: Blowing machine

Post by kkrell »

Terry,
I've been trying to understand better about the shape of the windway, the plug, etc. I came across this video by Phil Bleazey and thought it might be meaningful to readers of this thread. Phil makes the ramp starting at about the 7 minute mark (that's where my link takes you), and then makes various adjustments, explaining somewhat as he goes about the effects of the chamfers.

https://youtu.be/6LM4anm7vE4?t=405
International Traditional Music Society, Inc.
A non-profit 501c3 charity/educational public benefit corporation
Wooden Flute Obsession CDs (3 volumes, 6 discs, 7 hours, 120 players/tracks)
https://www.worldtrad.org
User avatar
Terry McGee
Posts: 3330
Joined: Sun Dec 12, 2004 4:12 pm
Please enter the next number in sequence: 1
Location: Malua Bay, on the NSW Nature Coast
Contact:

Re: Blowing machine

Post by Terry McGee »

Right, so let's see if we can follow the action.

He starts by cutting the ramp on his copy-router. So that's a spinning end-cutting tool that's following a template he's created. You wouldn't do this for a one-off instrument, but if you're trying to make a living out of this highly improbable career path, and you plan to make lots of these, you desperately need to avoid wasting time (and possibly wood) on repetitive tasks! And as he says, it's only roughing out. He will be personally responsible for the final profile.

I should mention I do something similar - I use a computer controlled (CNC) lathe to rough out my flutes. But, like Phil, then take over and do all the fine work by hand.

His second tool cuts the windway. It's essentially a scraper, hand driven by the lever you can see. It cuts on the pull, the forward stroke merely getting it into position. It's blade is curved, because he wants a curved windway. Not as curved as we see in the Sindt approach I would guess. That is curved to the diameter of the bore. This is less curved - lying somewhere between flat windways (think Generation etc) and the Sindt approach. I'd like to see some study into the relative merits of the flat, slightly curved and bore diameter curved approaches. Maybe someone has already done this?

Note that he's cutting into what we might call the ceiling of the windway. Phil's approach is (I think) similar to that used by fine Recorder makers. (The results, not necessarily the methods.) But I'm happy to be corrected on that.

He shows how he uses his "step gauges" to check the progress of the windway cutting. They appear to be cylindrical rods with a short section at the tip being turned down a little smaller. He appears to be holding the body of the gauge up to the ceiling of the windway, and checking if the smaller diameter tip is coming up to the bottom of the ramp. But others might have a different interpretation.

After a bit more scraping, he tests again and proclaims that "the step inside the instrument is now 9/10ths of a mm". That to me suggests he's looking for an exit height of 0.9mm (at least as a starting point). The step on the front of the step gauge presumably being 0.9mm deep, or 1.8mm less in diameter.

At this point, lets take a step back to where he discusses the design of his scraping machine. He mentions a curved "template" back near the driving handle. It's job is to allow the back end of the shaft to drop during the cutting stroke, meaning (if I understand correctly) the cutting tool end to rise up during the cutting stroke. This would yield a windway whose ceiling bows up in the middle, like we have seen in recorder and organ pipe windways and some moulded whistles. I would expect that to reduce resistance, but it might have other effects as well, eg directing the jet to some extent.

And then he moves to the hand-finishing stage. He talks about finishing the ramp by paring with a nice sharp chisel and mentions leaving a "lip at the end half a mm thick". I'm not sure what he means by that. Is it a flat front to the ramp 0.5mm thick? That sounds a bit brutal, but remember I've only had experience with treble D whistles up to this point, so what would I know? Others might usefully chip in here.

I do wonder if he's left that 0.5mm step at the tip of the ramp with a view to undercutting under the ramp. Remember his intended windway and ramp tip curve is not that of the bore of the instrument. So some internal tweaking would seem to be essential.

He briefly mentions the block, but doesn't go into detail other than saying "finish the top of the block". Since the top of the windway has been cut to a curve lying between flat and bore diameter, I'd expect the top of the block to be similarly tweaked, so that we end up with a constant windway height across the windway at any point along it. Any other interpretations?

So, Kevin, have I answered your question, or complicated it? Happy to take another shot!
Tunborough
Posts: 1419
Joined: Sun Dec 05, 2010 2:59 pm
antispam: No
Please enter the next number in sequence: 10
Location: Southwestern Ontario

Re: Blowing machine

Post by Tunborough »

trill wrote: Wed May 03, 2023 11:12 pm 1) what algorithm is used to calculate these ?
I used the Transmission Matrix Method (TMM). I can't find online a tidy description of how this method works, but you can find summaries in the theses of Antoine Lefebvre (2010) or Paul Dickens (2007).
trill wrote: Wed May 03, 2023 11:12 pm 2) what is the name of the software used ?
I wrote my own Python code, using Numpy for the complex numbers and matricies, and Matplotlib for the graphing. WIDesigner also uses the TMM, coded in Java. You can see most of the interesting formulas in https://github.com/edwardkort/WWIDesign ... /Tube.java, with the data structures in https://github.com/edwardkort/WWIDesign ... ector.java, and https://github.com/edwardkort/WWIDesign ... atrix.java.
trill wrote: Wed May 03, 2023 11:12 pm Also, could you post the data ? I'd like to do a close comparison.
Those graphs represent a couple of thousand data points each. A bit much for a C&F post.
User avatar
kkrell
Posts: 4831
Joined: Mon Jul 29, 2002 6:00 pm
antispam: No
Please enter the next number in sequence: 8
Tell us something.: Mostly producer of the Wooden Flute Obsession 3-volume 6-CD 7-hour set of mostly player's choice of Irish tunes, played mostly solo, on mostly wooden flutes by approximately 120 different mostly highly-rated traditional flute players & are mostly...
Location: Los Angeles
Contact:

Re: Blowing machine

Post by kkrell »

Terry McGee wrote: Thu May 04, 2023 6:13 am He briefly mentions the block, but doesn't go into detail other than saying "finish the top of the block". Since the top of the windway has been cut to a curve lying between flat and bore diameter, I'd expect the top of the block to be similarly tweaked, so that we end up with a constant windway height across the windway at any point along it. Any other interpretations?

So, Kevin, have I answered your question, or complicated it? Happy to take another shot!
I'm not sure that the block DOES get shaped to result in a constant windway height. I was already surprised that the top of the windway seems both curved across the body in 1 dimension, & I guess "scooped" along its length also. It's not clear to me whether the block follows that same pattern, or has it's own unique profile, from which it is possibly further hand-voiced. In any case, my not having any experience like this makes it hard for me to follow at all.
International Traditional Music Society, Inc.
A non-profit 501c3 charity/educational public benefit corporation
Wooden Flute Obsession CDs (3 volumes, 6 discs, 7 hours, 120 players/tracks)
https://www.worldtrad.org
User avatar
Terry McGee
Posts: 3330
Joined: Sun Dec 12, 2004 4:12 pm
Please enter the next number in sequence: 1
Location: Malua Bay, on the NSW Nature Coast
Contact:

Re: Blowing machine

Post by Terry McGee »

I think you're right, Kevin, which suggests that the radius of the cutting tool that scrapes the top of the windway follows that of the bore, albeit at a greater radius. In other words, the difference in radius of the cutting tool and the radius of the block is the height of the windway, presumably at the exit. Since there is a scooping action involved, and the radius of the cutting tool is fixed, the windway height will vary further forward and from side to side. But not much.

Otherwise, we would have seen some shaping of the top of the block. In the old rectangular windway recorder days, the top of the block had to be flattened, from side to side and sometimes from end to end, although sometimes a little scooping was involved along the way.

And he would have had to undercut the bottom of the ramp to follow the intermediate curve.

The bit in Phil's description that still bothers me is the 0.5mm step he leaves at the tip of the ramp when handfinishing it using the paring chisel, and as viewed "in the rear-vision mirror". It seems unlikely that the finished product ends up with a blunt ramp (hey, but keep in mind I've never owned a low D whistle!). Is it possible that he leaves the 0.5mm step in place at that point, but then tidies it up in final hand-voicing, perhaps with a steeper ramp at the tip?

I quipped to an overseas colleague about how sometimes the term "the simple whistle" crops up. There is nothing simple about whistles!
Post Reply