Page 11 of 12

Re: Rockstro

Posted: Wed Jul 27, 2022 9:01 pm
by GreenWood
Thanks for the link to those graphs Terry.

The lower join of 4683 looks like it was straight originally and was then compressed...either that or in some unusual way has morphed into the next one !?

I don't say people shouldn't be using computer modelling and so on, and whatever helps people carry on finding new understandings and ideas is fine by me. I just have an aversion to that approach as far as being creative is concerned. It sort of comes down to being that either I am looking at and via a computerised interpretation, or I am instead in that time dreaming out and imagining freely. I know which suits me better. Say Paddler though, I should imagine he will understand the programs he is writing and how they work, so from that he will also have his imagination and creativity engaged in tailoring them to guide his understanding. So not so very different in that sense. I'm not a programmer though :-) . Paddler, I hope you do not mind my having talked that way, i.e. in your absence.

So far my key idea is working, though now I have to prepare some very specifically dimensioned material so that it is as it should be when being played normally, and then I will see if that all fits smoothly with normal playing. I would be surprised if it hasn't been tried before and rejected for some reason, but then again I don't think it has. It would also make the flute to be fully vented, something I think would be good for top note second octave tuning on cylindrical flute without parabolic end. It doesn't allow choice of placement of normal keyless toneholes, but that is on purpose because the original idea I had when I first thought of it was to keep the feel of the flute as close to keyless as possible, but to still have the extra notes accessible. If it is good, it will really be a new direction for the flute, not that players will nescessarily adopt it. I think some would though because the whole concept is meant to be a very natural evolution from keyless.

Re: Rockstro

Posted: Thu Jul 28, 2022 5:43 am
by Terry McGee
GreenWood wrote: Wed Jul 27, 2022 9:01 pm The lower join of 4683 looks like it was straight originally and was then compressed...either that or in some unusual way has morphed into the next one !?
That one has the classic "strangled tenon" look - the thin wood at the bottom of the tenon trench has collapsed after goodness-knows how many rewrappings over goodness-knows how many years. The top tenons show a little compression,the middle tenons considerably more, and this lower one has really collapsed. But it's the lower tenon on the other flute that is really a bit confusing. The bore at that point flares to larger than the bore at the top of the foot. Why would you do that? We'd need to be able to examine the flute to know if it was made originally like that, or if this has been caused by some later repair, damage or (gasp!) "restoration".

I think what these data are telling me is that we are relying on thin ice. These were top-of-the-range flutes from their era, but we know that they were made by essentially cottage-workers. They weren't responsible for improving the product, their job was to turn them out at the lowest price they could. Nobody was responsible for improving the product until something happened that demanded urgent improvement. Like Clinton, or Siccama or Ward or Boehm or Pratten or some other upstart coming in to mess up life as we knew it.

The flutes will have needed maintenance since then, possibly numerous times, and again, we have no way of knowing how knowledgeable those various maintenance people were. Their owners will have done some of the maintenance, and again, we can't assume much skill or understanding. Tenon has come loose, whack some more thread on it. And since it had come loose, make sure to put the thread on tight! And sure, feel free to bring a flute built in cold dank London out to the Californian and Australian goldfields. What could possibly go wrong? (I've dealt with those personally. The answer is "lots"!)

Now, of course, I'm on thin ice, Greenwood, when I promote computer modelling rather than "successive approximation" approach - make one, then try to make a better one. Because that's been my approach, and probably every other maker's to date. And has served us all well. So you should feel very free to have a go! I just feel that to get "to the next plane" at some stage we are going to need to try to get a better understanding of the issues that remain and their possible solutions. I probably should do the decent thing, and dedicate my twilight years to doing just that. Unfortunately, I agree with you - making stuff is far too much fun. Just finishing a 6-key today and enjoying the fact that my standards are still rising. It seems every few days lately we've learned of another great Irish musician/academic has passed on. Today it was Mick Malony, just four years my senior. This is getting personal....

Re: Rockstro

Posted: Thu Jul 28, 2022 1:40 pm
by Conical bore
Terry McGee wrote: Thu Jul 28, 2022 5:43 am Now, of course, I'm on thin ice, Greenwood, when I promote computer modelling rather than "successive approximation" approach - make one, then try to make a better one. Because that's been my approach, and probably every other maker's to date. And has served us all well. So you should feel very free to have a go! I just feel that to get "to the next plane" at some stage we are going to need to try to get a better understanding of the issues that remain and their possible solutions. I probably should do the decent thing, and dedicate my twilight years to doing just that. Unfortunately, I agree with you - making stuff is far too much fun.
Don't we already have a halfway solution between woodworking trial-and-error and pure computer modeling in the form of 3D printing?

If you're interested in different chambering effects within the flute bore without modeling it acoustically, just try a bunch of different shape ideas with a 3D modeling program, print it on a 3D printer and see how it performs. It seems like a logical next step for anyone interested in trying something different from what is already proven to work well. It could satisfy the "making stuff is too much fun" itch because you actually are making stuff. Well, you and the robot. You could try dozens of different ideas very quickly, just for the cost of the electricity and the filament material.

Of course you might end up with something that works well in a bore shape that would be very difficult to ream out of a wood blank, but 3D printing with wood filament is available now. Even methods that recreate grain, although I don't know how much of that has trickled down to home/hobbyist 3D printers. At least the acoustic and weight properties should be similar to resin-impregnated wood flutes.

Re: Rockstro

Posted: Thu Jul 28, 2022 11:55 pm
by Terry McGee
I'd agree that 3D printing has a place in the development but ideally I wouldn't want to start there. You'd need to make a lot of trial bores before you could be sure to have cracked the best approach. Ideally, I'd want to explore the possibilities and alert myself to potential pitfalls using modelling. No point for example in coming up with a design that gives a stunning low E but then finding it takes out 2nd octave A! Then perhaps make a 3D Printed version to confirm that the modelling is working. That might refine the model but then require more tweaking and another printing. Then finally look at the practicality of making it in wood. I'd be pretty confident that, whatever the challenges, we would be able to crack that. (Metaphorically speaking of course!)

It's interesting to consider what we might hope to achieve in such a project. I'd be interested in:
- exploring Boehm's notion of a flute with a 20mm bore head (instead of our traditional 19mm, 3/4")
- exploring Ward's notion of a linear tapered bore, tempered with the handicaps placed upon us by our insistence on open holes that we can reach and cover!
- exploring closer fingering to assist oldies like me to keep playing the flute longer without suffering dodgy notes (I know, self, self, self!)
- exploring bore-based solutions to counter our known weak notes, eg Low E, ideally without loss to our known good notes!
- develop computer modelling to make it a practical tool to guide any future directions fluters might wish to explore
- use the model to vet and optimise other potentially interesting historical models (eg the larger holed US-made flutes)
- use the model to explore just how accurate we need to be in various parts of the flute
- see if the model can help answer historical conundra such as was chambering a real solution to a real problem, a bit of Rockstro puffery or even a misreading of Rockstro?
- see if a keyless flute can simply be a keyed flute without the extra holes drilled, or can/should we optimise the two designs separately?
- attempt to follow Nicholson the Elder's path in "opening up the holes on his old Astor" ushering in the "Improved era". And with it, the flat foot?
- and that's just for starters!

Feel free to add to my list!

Re: Rockstro

Posted: Fri Jul 29, 2022 2:29 am
by paddler
You can only use computer based modeling to answer the kinds of questions listed above when the models are
accurate enough to give answers that can be trusted. They are not really there yet. I say this because when you
take accurate dimensions from an existing flute, plug them into a computer model, and predict its tuning, the
predictions differ from the actual tuning. Maybe not by miles, but by enough to make it difficult to definitively
answer the kinds of questions listed here. And here I'm still just talking about basic pitches, not the complex mix of
harmonics, which influences the quality of tone that is desirable in a flute.

If you succeed in tweaking model parameters to improve its accuracy, you still need to know whether you have improved
accuracy in general (i.e., for all possible flute designs, or some subset), or whether the model is now accurate just for the
specific flute you tested it against. It is not an easy task to produce an accurate computer model, especially one
that is accurate over a range of hypothetical flute designs. So there is always going to be a danger in assuming that the
output from a computer model is truth. I don't think you should underestimate the difficulty in building accurate
computer models, or prematurely place your trust in the predictions of computer models that are not actually ready yet.
Verification will probably require the construction of a bunch of physical prototypes anyway.

Re: Rockstro

Posted: Fri Jul 29, 2022 5:21 am
by Terry McGee
Oh, I agree Paddler. Which is why I thought the intermediate step of using 3D printing to build a prototype and then, if necessary, use the results to refine the model, and print again if needed were important steps. Essentially we would have to train the model before it would be any good to us, and then test it to see how adaptable it is before trusting it in any new applications. It would be a really nice project for a PhD thesis....

Re: Rockstro

Posted: Fri Jul 29, 2022 9:34 am
by GreenWood
It does doesn't it, the bore flare wider. Adding to the very reasonable want list you provide, we might seek a "wood technician" , that is someone who is able to study a historic wood form in detail and provide an interpretation of just how it achieved that eventual shape. That would be by knowing how to study changes in grain direction and compactness and so on.

What is funny, and it is just maybe that when a bore compresses it tends to take the form it does on 4683 by overlapping the potential chamber, but that bottom join looks exactly like the lower join on

Image

to me, your fake Rudall. I don't imply 4683 has a fake footjoint or anything, just that both seem to have acted the same.

Conical bore

"Of course you might end up with something that works well in a bore shape that would be very difficult to ream out of a wood blank"

I think if any design is that good, the extra effort will be taken or an easy way to achieve the desired shape will be found.

I think part of the problem though is that already we are presented with such a diversity of options, of flute types, that one of the hardest parts is knowing what we are trying to refine and why. Tuning is about the only obvious parameter I am able to think of, the rest are degrees of subtlety and not always easily agreed on.

I know if I had a 3D printer it would be in continuous use though, I would be using it to rough out ideas. Terry's suggestion of using 3D printed flutes to train mathematical models seems good also.

Paddler, I would say I pity those trying to build computer models, but I know that usually they find that both challenging and rewarding. I would soon get frustrated though trying to make the program fit the physics fit the flute fit the physics fit the math fit the program and so on.

So I look at say obvious physical features and am just happy enough discovering what any new flute sounds like, and trying to understand why or what is noteworthy in a much more mundane way. Not to put you off at all, I know that computer modelling is extremely difficult and complex. Added to that, there is the whole facet of player style, embouchure and flute placement, which would add another whole new layer of complexity. Just one reason I dread computer modelling, another being that it potentially would eventually be good enough to replace learning flute design by a more intuitive process, good enough but not the same as, and there we are talking craft and experience vs reliance on programmed output, which is a very different topic. I think there is room for all approaches to be complimentary to each other though.

Re: Rockstro

Posted: Sun Jul 31, 2022 12:04 pm
by David Cooper
I don't know how much it would cost for a 3D printer that can handle something as long as a flute, but there is a much more affordable way to produce flutes reasonably quickly with varying bores at very low cost and with beautifully smooth results. I mentioned in another thread that I'd made a quena out of craft resin (a food-safe type of epoxy), though that one had an octagonal-section bore due to the difficulty of making such a thing out of liquid without a mould. However, I have since made one with a cylindrical bore and it's the best quena I've made by a mile, while the method of construction is likely a new innovation. I got the idea from seeing cup turning: https://www.youtube.com/watch?v=7xYubHf ... rpleStitch but I took it a step further by making the inner mould out of silicone rubber in the same manner by pouring the liquid onto a rotating rod. So, I bought a motor for about £8 which turns five times a minute, then attached a rod to it (and to a thing that rotates at the opposite end - don't know what that's called as it was salvaged out of a broken tape recorder). I then set it rotating and poured the silicone rubber liquid onto it, building it up in thin layers (you can't do thick layers as too much weight would make it drip off the tube). Each layer takes a couple of hours to solidify - you mix the stuff out of the content of two bottles and then have about an hour to work it before it gets too thick. I was aiming for a cylindrical bore, so for the final layer I used a long piece of plastic like a ruler to press in against it to get the viscous liquid exactly where I wanted it, but that's the thing you can use to vary the shape and make it conical - cut your ruler-like shaper (let's call it a stencil) to the shape of the required bore and it then shapes the silicone mould you're building. You can get the shape approximately right without pressing that against it just by putting more layers on in the parts that need to be thicker and only apply the stencil for the last layer to get a smooth and perfectly shaped finish. (You could also use a conical rod to start with.) You then build the flute tube on top of the silicone mould simply by switching to pouring layers of craft resin onto it, again building it up in layers and putting more of it where you want thicker walls. You end up with a flute tube with the density of tropical hardwood which sinks in water, and if the walls are more than a couple of millimetres thick, it's very rigid too.

It's important to use a reasonably wide rod to build your mould on, because when you pull it out you want to create enough space to be able to extract the mould from inside the flute tube. You also need to wrap a layer of plastic film round the rod before building the mould on top of that so that you can pull the rod out - the silicone rubber may not stick to things, but it grips it firmly sideways with very high friction. With the rod removed, you can separate the mould from the flute tube using something like a knitting needle to prise the mould into the space left by the removal of the rod, and you can spray in silicone liquid to help keep it free. The mould then pulls out and can be reused to make many more flutes of the same design, or you could make repeated modifications to it to experiment with a range of similar bore profiles.

Re: Rockstro

Posted: Mon Aug 01, 2022 7:10 am
by Tunborough
:wink:
paddler wrote: Fri Jul 29, 2022 2:29 am You can only use computer based modeling to answer the kinds of questions listed above when the models are accurate enough to give answers that can be trusted. They are not really there yet. I say this because when you take accurate dimensions from an existing flute, plug them into a computer model, and predict its tuning, the predictions differ from the actual tuning.
I agree that computer models (those I'm aware of) are not going to accurately predict the tuning of every note on a real, physical flute. No, they aren't there yet. But then, if you attempt to build an exact copy of a physical flute from its basic dimensions alone, how close will the tuning be? Closer than a computer model, yes, but not identical, I'll wager. For that matter, if you hand your flute to another player, will their tuning precisely agree with yours?

Computer models aren't at the point where we can plug the results into a 3D printer and get the flute we expect. They may never be. But they are good enough to indicate tendencies, such as which notes will be hard to play in tune, and suggest what direction to take a design, as in experiments like this, viewtopic.php?p=1212419#p1212419.
paddler wrote: Fri Jul 29, 2022 2:29 amVerification will probably require the construction of a bunch of physical prototypes anyway.
Yes, definitely, but perhaps it can cut down on the number of physical prototypes. In the time it takes to clean the shavings off a lathe, a computer model can try out thousands of permutations of a design and pick the best one it finds.
paddler wrote: Fri Jul 29, 2022 2:29 amSo there is always going to be a danger in assuming that the output from a computer model is truth.
The computer model is only a tool, and maybe a blunt tool, like a 3D printer is a tool. Not useful to every flute builder or designer, and to be used with due awareness of its limitations.

Re: Rockstro

Posted: Mon Aug 01, 2022 8:38 am
by Tunborough
Tunborough wrote: Mon Aug 01, 2022 7:10 am ... with due awareness of its limitations.
I forgot to point out one big limitation... I don't think any model has attempted to cope with keyed flutes, beyond approximating them as keyless flutes for 16 fingers. There are too many additional variables to account for.

Re: Rockstro

Posted: Mon Aug 01, 2022 7:21 pm
by Terry McGee
Fair enough, Tunborough, I don't think that's a practical problem at this stage. If we can use it to help optimise keyless flutes (which is the logical starting point, given that they are much easier and cheaper to make), and then find the equivalent keyed versions are deficient in the harmonic line-up, it's telling us that the additional closed holes are something we need to allow for in the bore shaping. Armed with that proof, we can find ways to tweak the bore to minimise the impact on the harmonic support.

My inclinations would be to embark on some quite new direction and see how far we get. Boehm's idea for a 20mm bore flute still haunts me. From memory, he was well pleased with the fuller body it produced, but put off because it made notes above the middle of the third octave harder to get. Doesn't sound like a killer blow to me!

Or am I already too late? Are 20mm bore flutes already littering the marketplaces in downtown Miltown Malbay, and being redeployed as tomato stakes in Lisdoonvarna? Or have they been banned by the European Union health authorities for their impact on the hearing of nearby tenor banjo players?

Re: Rockstro

Posted: Sat Aug 06, 2022 5:03 pm
by GreenWood
I'm still haven't made a cylindrical bore over 16.5 mm , am centuries behind, listening to a Radcliff flute of the future and feeling there might be some merit to 19 mm. Will probably find everyone playing 25 mm by the time I get to 19 mm. That's the thing though Terry, you have to think BIG, say 23 mm. That way you have room to bargain with the banjo players if it comes to it. Got to haggle you know.

Re: Rockstro

Posted: Wed Aug 10, 2022 1:48 am
by Terry McGee
Heh heh, problem is you're haggling with the Laws of Physics, and they take no prisoners! Interesting to note that the Low Bb band flutes have a head bore about 20.9mm, although again, being "outdoor instruments", they are intended to be played up into the third octave or so.

Some time back I had a binge on piccolos - flutes in D, half the length and therefore one octave above our normal D flute. I was working from 19th century English originals but I quickly decided I didn't like the relatively shrill nature of the sound. If you're already up an octave, you don't need to be shrill too. (Look at me, look at me!) So I decided to experiment with larger bores, and it's instructive what I found.

The originals tended to have a bore around 11mm, and, as I said, were too shrill.

When I was starting out making flutes, I helped sustain the income by making cheap and easy 6 hole cylindrical piccolos I sold at fairs and festivals, mostly for kids though some adults bought them too. They were 1/2" diameter bores, and I thought sounded well, although tending flat at the top of the second octave as cylinder flutes tend to. So I developed a 1/2" (12.5mm) bore conical, and was really pleased with it. It's become my main piccolo. So much nicer than 11mm!

Drunk with power (or maybe something stronger?) I decided to go further, and cooked up a 14mm bore conical piccolo. This sounded fabulous on the low notes, but really struggled about halfway up the second octave. I made a few as "Busker's Piccolos", to capitalise on the power at the bottom end, but I never felt that they had much to offer the world of Irish music. We need at least two good octaves, and a few more notes can come in handy when pressed!

So you can go too far with bore diameter, and it's the top end that will suffer. The question is was Boehm right about a 20mm bore flute being richer in the bottom and OK up into the third octave? There's some confusion in his book, which I suspect comes from the translator, Dayton C Miller (the collector whose flutes we see when we visit the Library of Congress). He quotes Boehm noting in a letter in 1867 that he could still play a 20mm bore flute up to C6, but from F#5 upwards the note were "sounded with difficulty". Given the note at the bottom of the flute is C4, C6 is only the bottom of third octave, and the F#5 is our second octave F#. That would be tragically bad, and seems improbable that a mere 5% increase in diameter would produce such a wipeout. So I can only think Miller got the numbers wrong, and that Boehm meant it was fine up to third octave F#, F#6. That I reckon wouldn't be a problem for us.

Re: Rockstro

Posted: Wed Aug 10, 2022 3:39 am
by paddler
The value of a wider bore was well recognized by at least one baroque flute maker as early as the mid 1700s.
Johan Joachim Quantz (1697-1773), a hugely influential flute maker, was famous for such things as inventing
the tuning slide, and designing flutes with two separate keys for D# and Eb, but his flutes also had a larger bore
than was common at that time, or today, or any time in between.

I have a set of detailed plans for several of Quantz's flutes, and notice that he used a head bore of 20.4 mm!
This particular flute had several corps de rechange (interchangeable upper body sections for use when playing
at different pitch standards), and the widest point in the bore for each of the body sections varies from 20.3 mm
for the smallest, to 20.5 mm for the largest. This is for a D flute, but it is pitched lower than modern standards.
The intention seems to have been to optimize the bore for something close to A=392, which is more like a
C# flute at A=440 hz, but the interchangeable body sections allow the pitch to be changed up to A=415 hz for the
shortest section. Whichever way you look at it, this is proportionately a very large bore.

Also, note that this is a conical bore flute, not a cylindrical bore, parabolic head, Boehm-style flute, so it is quite
applicable to the conical bore Irish flute designs we use today. Quantz flutes are reported to play very well across
the full range of notes used in baroque flute music, so definitely beyond what we typically use in ITM in the upper
registers.

Given the extraordinary level of sophistication and attention to detail in Quantz's work (it could even be argued,
far surpassing anything since!) I think it lends considerable weight to the argument that there is merit in a larger
than standard bore. For a deep dive into Quantz's mindset and level of mastery of all things flute-related, you can read
a translation of his original 1752 treatise "On Playing the Flute" (400+ pages). It contains an absolutely mind-boggling
level of detail about flute playing, performance practice, tuning standards, etc. After reading it I was left with the
feeling that almost everything we think of, and a lot that we have not thought of, has already been thoroughly explored
hundreds of years ago ... and written up, published, patented, built and sold. :boggle:
I was also left with the strong impression that Quantz didn't do anything by accident.

Re: Rockstro

Posted: Wed Aug 10, 2022 6:14 pm
by GreenWood
Terry

I'm still trying to understand how such a small change in bore makes such a difference. Not the physics as such, just from the sound of the various cylindrical flutes I have made at around 16mm. The toneholes on these are slightly different from each other but not enough to make the difference in sound I notice. I say around 16mm, because the final bore size just depends on shrinkage if wood is green, sanding and so on. So 15.5 mm on a D flute I find gives a very clear sound and a bass that is contained but correct. 16 mm gives a resonant bass and slightly rounder overall sound. 16.5 mm has full but rounder bass and second octave similar to 16 mm, but the lesser resistance of the bore allows it to be played more like some conical bore flutes, with a lot of note variation at choice with air pressure (sort of towards Sligo style ?) . This is just one design of flute though, with one size of embouchure, so I could not say it all transfers to other flutes but it has given me a clear example of how much difference bore diameter can make.

I don't know much about fifes and piccolos, I made one and figured it takes a certain character to play them ? I like bass, and already second octave seems high sometimes on a flute... I remember a discussion somewhere where someone was trying for #7 or 8 or something and was told to buy a dog whistle. My hearing is sensitive though, so say on guitar I tend to hear all the treble in even a bass note, which might explain my preference.

I still think conical bores are strange Terry, I haven't figured them out yet but hope to in my own way. Where it comes to bore diameter it gets even more confusing because of the shape. You would think they would start thin and then widen ? I might yet try a flute with a gentle gradient that way, but at the same time it makes sense to keep the sound "in", i.e. a clear definition of chamber, which conical bore offers. I think most of the sound comes via embouchure, so tonehole size isn't immensely important in that respect, in terms of sound volume more by allowing different pressures of playing (more air and louder for example) which wider bore does also. I'm just speaking from what I sense so far, so I hope it is taken as a form of interpretation only.

Maybe the 14mm piccolo would have liked a different shape bore, slightly flatter, to help top of second octave ? There are so many possibilities with bores I wouldn't know where to start trying to figure it out beyond trial and error, then learned intuition.

That is the thing with Boehm's 20mm also. Admittedly a Boehm system 20mm flute is a large project, but to make one with all holes drilled and no key system, just stopping toneholes as nescessary to explore the sound is not so much...it's always easy to say to someone else to do the work though. I would be surprised if no one has tried that, but obviously no one has written up on it.

Paddler

With a conical bore, I always have the impression that the bore volume is secondary, that the main characteristic comes from resistance (i.e. total venting). So larger volume (wider bore) would seem to me, within reasonable parameters, to be able to accentuate qualities found via the resistance. Conical bores I picture as having two components, where the cylindrical head is one, and the conical part acts as tuner to that. I know the whole sound is continuous between the two, but somehow it seems to me to have these components. I will have to read up on what Quantz had to say. If his flute design was thought out, then he must have found a very clear and accurate way of visualising the whole.