Better Tubas

[Søren]

From reading on tubenet and talking to Imperibari, I have learned that one of the things that make a tuba play better is to minimize air turbulence. From my understanding, this problem alters the intensities of the frequency spectrum, and may introduce false ones.
Since I am studying for an exam right now, I was looking in my old introductionary physics book and stumbled on a section on fluid dynamics. I have talked to some people, and it seems that one could, more or less, use these calculations on airstreams too. The trick is to introduce a parameter so that one can compress water(if I had been reading a more advanced book, these parameters would have been present, since there is always some air in water).
One could just set this air parameter to maximum, and the calculations would largely apply to airstreams.

It seems from my book, that you gain a lot from having a smooth surface in the tube where the waterflux is. And my thought was that this conclusion would also apply to airflux in a tube/tuba.
All of my tubas have a smoother outer surface, than the inside surface. I can imagine that this it partially due to the massive amount of work it would take to polish the inner surface of a tuba(if it could be done at all!!).

But imagine that one could smoothen the surface of the first half of the tubing on a tuba(where the airspeed is highest), and taking that all joints and all soldering is perfect. This would give a lesser amount of airresistance, that I imagne would be a good thing. But I guess that it would also minimize air turbulence. And THAT would significant better the playing characteristics, I guess…


On another note: I know that doing calculations on a straight tuba is far from a real tuba, but I do not see why it can not be done. It seems that everybody is calculating everything, in order to make things work. EVERYTHING is scientifically worked though, BUT brass instruments. From what I have learned, tuba designing it is considered a craft not a science. Why is nobody computersimulating a tuba. Tell the computer what sound (frequencies) you would like the tuba to produce, and then let the computer calculate what dimensions that tuba should have?
I know that these calculations would be almost endless amounts of differential equations, but I guess it could be done. Why is there not an instrument maker, who would use the money to have an acoustics Ph. D to make such a program?


And last. Does anyone on this forum have a small computer program that is capable of doing Fourier Transformations on a wavefile? I would like to record my tubasound, and see what frequencies I produce with different tubas and mouthpieces.

Any respond and insights is appreciated.

[Rick Denney]

"Smooth" is not the same thing as "polished". It turns out that a slightly matte surface actually flows better than a polished surface, because the turbulence at the microscopic level on the matte surface lubricates the boundary layer. It's a little like rolling on marbles or ball bearings. When the turbulence grows too much it creates more disadvantage than advantage.

Of course, fluid flow is usually linear, not oscillating, so the calculations are different. In the frequency domain, the reflections from the features of the instrument (including the bell opening) either inhibit or help sustain the buzz. Those pulses are pressure waves that move at the speed of sound and superpose on top of each other. Some frequencies are unresonant and are attenuated out. Some are resonant and are amplified. Both of those processes work by allowing the pressure waves to move freely through the instrument.

Moving pressure waves through the instrument is not the same thing as moving air through the instrument. The air is there already. The effect of turbulence is to create its own often undesirable pressure waves more than to inhibit air flow.

In fluid flow, one worries about the frequency domain to make sure that turbulance and resonance don't create vacuum pockets that cause cavitation, so the goal is to eliminate turbulence if possible. In a tuba, that is an effect rather than a cause. The goal in a tuba is to resonate useful frequencies to a greater extent and attenuate unuseful frequencies more completely.

Rick "noting different requirements" Denney

[Søren]

I see. Guess I should have looked in the section on sound and waves also.

[Alex C]

Some tuba manufacturers are already designing tubas on computer. The Miraphone 1291 is supposed to have been computer designed before it was built. I've read that the asymetrical design of the bows was a direct result of computer/engineer input.

If I'm not mistaken, the Holton Phillips model tuba was first designed on computer, though I'm not sure this included any acoustic information. It may have been for design and manufacturing purposes only (or it seems that way from reading the article).

Interestingly, tuba manufacturers are using an artificial tuba player to determine many of the playing characteristics a completed design has. Apparently there were too many variables using real tuba players.

[circusboy]

Interesting stuff, SØren and Rick. Thanks.

To me (one who knows very little of tuba manufacturing) it does beg the question: Why haven't they figured it out yet?

Look at optics, for example. When they started using computer modeling in earnest for lens design for consumer-oriented cameras (1980s?), the lenses improved dramatically AND the cost to manufacture went down dramatically.

With what we know about the nature of sound and air it seems to me that there should be more (some? one?) acoustically perfect horns out there. Or is the problem that there aren't enough perfect wind/buzz makers out there? Or are the demands (x # of notes over x # of octaves) just too high/unrealistic/unobtainable? SHOULD it remain more a craft than a science?

[Donn]

With what we know about the nature of sound and air it seems to me that there should be more (some? one?) acoustically perfect horns out there.

Leaving aside for the moment whether this is true, I suspect the answer to your question is that tuba players don't particularly want a perfect tuba. Even just in terms of tending to play in tune with itself across the 2nd/3rd/4th partials. If true, that would reduce the incentive to invest big money in this quest.

[Chuck(G)]

Some tuba manufacturers are already designing tubas on computer. The Miraphone 1291 is supposed to have been computer designed before it was built. I've read that the asymetrical design of the bows was a direct result of computer/engineer input.

With very few exceptions, computers design nothing. And thank heavens that we haven't devolved to that level yet!

When we used to care about accuracy in terminology, we called it "Computer Aided Design". That is, the computer gives us some tools that might not have been practical heretofore, such as mathematical simulation.

But it's still some human sitting at a desk somewhere who's calling the shots when it comes to design. That human simply has more information now that s/he had before.

[windshieldbug]

... and in-tune with itself is, itself, a moving target. Well-tempered? Pythagorean? Just? In what context, and with what other instruments?

There are many (and I count myself one) that say that a 'perfect' instrument would allow adjustable intonation and tone quality, not just the one...

[evan]

And last. Does anyone on this forum have a small computer program that is capable of doing Fourier Transformations on a wavefile? I would like to record my tubasound, and see what frequencies I produce with different tubas and mouthpieces.

You can do this in Matlab, of course if you're not at a university or some kind of reseach institution you probably won't have access to matlab. There are a couple of open source (free) programs that are like Matlab, but I really haven't played with them much. You could try Octave or SciLab. If you just want to look at a file or two, you could send them to me and I can FFT them in Matlab

----

Also, it seems like with some kind of Computational Fluid Dynamics or Finite Element software it would be possible to model a tuba. With the number of engineers working in this area over the past couple decades, I bet the software already exists -- you'd just have to input the precise geometric structure of a tuba & model the soundwaves produced by buzzing lips. No problem! (hah)

-Evan

[sloan]

Of course, fluid flow is usually linear, not oscillating, ...

Do you believe that air flow through a tuba is oscillating?

[Rick Denney]

Do you believe that air flow through a tuba is oscillating?

No. Just the pressure waves.

And, yes, I know there are pressure waves in liquids, too. But in the fluid dynamics classes I took, we were not too concerned about that compared to flow, head, and cavitation. With the exception of some aspects of cavitation, most of those issues can be modeled without delving into the frequency domain.

Rick "noting that flow is not especially important in a tuba compared to vibration" Denney

[Lew]

It is interesting that Conn used to advertise (as early as the 1930s) that they used hydraulic expansion, i.e. high pressure water through tubing in a form, to make their instrument branches, claiming that it resulted in perfectly smooth and uniform thickness tubing. Do manufacturers do that today? Would it matter?

I do find that Conn "Crysteel" valves seem to exhibit less wear than other valves from the same era. I have seen "crysteel" valves on obviously very used 20Js from the 30's or 40's that looked, and played like new. Was there really anything different about these valves from standard nickel plated valves? If they were better, were they that much more expensive to make and therefore no longer economical? I know that I would spend a little extra for valves that would last forever, although I expect that the valves on anything I own now will not need any work until long after I'm gone.

[Rick Denney]

Look at optics, for example. When they started using computer modeling in earnest for lens design for consumer-oriented cameras (1980s?), the lenses improved dramatically AND the cost to manufacture went down dramatically.

Yes. There were also advances in molded aspherical elements and the use of exotic low-dispersion glass that made previously difficult things possible.

That said, modern prime lenses (as opposed to zooms) are very little better now than they were in the 30's, with the exception of improved anti-reflection coatings and better close focusing by using floating elements. The Zeiss Planer, one of the best-performing normal prime lenses on the market, was designed well over 100 years ago. The storied Zeiss Sonnar was designed in the 30's. The Schneider Xenotar was designed in the 40's. My point is that computer modeling didn't do anything we didn't already know how to do, but lacked the computational convenience to do it for consumer products. Now, even the most complex lens modeling algorithms can be conveniently applied to cheapie built-in zooms in point-n-shoot cameras.

The computer tools can only do what we know how to do. The breakthrough with lenses was thinking of a new way to characterize their quality, so that we could optimize to a more useful objective. This occurred when lens designers started optimizing to improve the modulation transfer function instead of merely resolution.

Another thing is price. Leitz makes the sharpest lenses available for the Leica cameras. A 50mm Summicron, for example, is a simple Planar design. But Leitz gets $1100 for one. Canon's similar optics (but in a cheap plastic barrel) gets $75 brand new. What's the difference? The Leitz barrel is beatifully crafted, to be sure. But that's not the big difference. The big difference is that Leitz is willing to throw away the lenses that don't meet their standard. Their manufacturing skill is little better than Canon's (Canon lenses can be beautifully made when they want them to be), but their line for rejection is much higher.

How many tuba manufacturers can afford to throw away every third or fifth Blattophone that comes out of the shop just because it isn't quite up to snuff? None could afford to, of course. The total production of high-end tubas is still only a handful, and rejecting one would likely drive the price up on the others by very significant amounts. So, they improve their ability to tweak their instruments up to a level of competence, trusting that even the dogs will find someone, as some price, who is willing to deal with the doggish issues.

But the real issue is in your second point:

With what we know about the nature of sound and air it seems to me that there should be more (some? one?) acoustically perfect horns out there. Or is the problem that there aren't enough perfect wind/buzz makers out there? Or are the demands (x # of notes over x # of octaves) just too high/unrealistic/unobtainable? SHOULD it remain more a craft than a science?

What is "perfect"? I submit that the reason there is no perfect tuba is as much because we don't agree on perfection as anything else. Some people listen for sound qualities that most people can't hear. And they are listening for different qualities than the next guy. Some people focus on intonation to the exclusion of sound. Some put their attention to ergonomics. Some want flexibility, while others want forgiveness. Most want some combination of sound and feel, with some players (rightly, I think) focusing on sound and others (like me) focusing on feel. Then, there's size, responsiveness, loudness, propagation, projection, efficiency, portability, finish, build quality, price, cachet--a whole range of competing optimization objectives. No two tuba players have the same mix of these objectives in mind when they choose an instrument, if indeed their mix of objectives remains the same for more than five minutes.

The best computer models that I know about can model the intonation of the instrument given the taper design, but even those models are quite simplistic. They are sort-of like the models used by the big car manufacturers for wind-tunnel testing: They are not in full scale, they don't have the same irregular surfaces as the production version, and they don't have all those messy things like mirrors, wipers, and antennas hanging off them. But those models don't model sound, and we still don't know how to define the most desirable sound.

In these days of computational convenience, we can optimize better than ever. But we need a model of behavior that is realistic, and developing such a model is not trivial, and doesn't have anything like a big enough market driving it. Why would, say, Hirsbrunner spend hundreds of thousands of dollars on research so they can marginally improve (maybe) an item of which they make only a handful each year? The only motivation for doing so would be to improve sales, and I submit that such research could not possibly pay for itself when applied to high-end instruments.

But even if we had realistic models, we still 1.) have to define the objective function, and 2.) have to be able to build faithfully according to the model.

There's nothing to suggest that even a successful design based on computer modeling will do better than the sum of 150 years of experimentation and tweaking.

Rick "who thinks our young scientists who ask 'why not?' ought to try their hand at doing the analysis themselves before being quick to accuse the industry of laziness" Denney

[circusboy]

The Zeiss Planer, one of the best-performing normal prime lenses on the market, was designed well over 100 years ago. The storied Zeiss Sonnar was designed in the 30's.

And don't forget my personal favorite, the Zeiss Biogon (1938).

I certainly understand your argument about market variables and the cost of production in the obscure little business of tuba-making vs. the enormous consumer base of the camera-making business.

And I'm glad to learn that manufacturers such as Miraphone are using computers to aid in design. (I never meant to suggest that computers actually design things themselves. Give me a litte credit.) From what I hear about the new 281, it sounds like it's working.

I also understand that one player's perfect tuba isn't everyone's--that there are huge varieties of sound presence, coloring, loudness, etc.--largely taste issues (for which, we know, there's no accounting.) My question regarding what I termed an "accoustically perfect" tuba was refering to a horn on which all notes within the range of the horn could be played--even Cs below the scale on F tubas--without pulling or pushing slides, embouchure contortions, etc. So, I guess intonation/general playability would be the most important factor.

Other improvements (not necessarily related to "perfection") that I believe CAD could help with would be efficiency-related, including efficiency of breath needed to produce sound, efficiency of size, efficiency of price/manufacture.

I don't know enough about the tuba-making biz to know just what sort of investment it would take to start putting 20th century technology to work for them, but I believe that those who haven't yet should be thinking about it.

circus "whose camera, tuba and body were all born in about the same year and are all showing signs of wear" boy

[Rick Denney]

The Zeiss Planer, one of the best-performing normal prime lenses on the market, was designed well over 100 years ago. The storied Zeiss Sonnar was designed in the 30's.

And don't forget my personal favorite, the Zeiss Biogon (1938).

...

And I'm glad to learn that manufacturers such as Miraphone are using computers to aid in design.

...

My question regarding what I termed an "accoustically perfect" tuba was refering to a horn on which all notes within the range of the horn could be played--even Cs below the scale on F tubas--without pulling or pushing slides, embouchure contortions, etc. So, I guess intonation/general playability would be the most important factor.

Other improvements (not necessarily related to "perfection") that I believe CAD could help with would be efficiency-related, including efficiency of breath needed to produce sound, efficiency of size, efficiency of price/manufacture.

The Biogon was a marvel of a lens, but without the retrofocus design could not be used on a reflex camera. The Distagon was the first Zeiss retrofocus lens, but even 55 years after the Distagon was introduced, it still doesn't match the Biogon. One of these days, I'll find a Hasselblad SWC cheap enough to buy for fun, assuming that by that time film will still be available.

All of the German manufacturers use computer optimization for improving intonation on their taper designs, not just Miraphone. But that software doesn't model important details like bends, joints, and valves. So, it's useful mostly to provide a starting point.

I have this feeling that if it was easy to make the low C on an F tuba feel like the low C on a C tuba, without taking from the F tuba what makes it essentially an F tuba, we would have one by now. But the Europeans for whom those are their day-to-day instrument don't seem to have trouble with that low C. They want an F tuba to be an F tuba, and they have a very refined notion of what that is. We want an F to be like a small C with an easy upper register. That's quite a different set of requirements.

My Yamaha 621 has excellent intonation and a low C that feels like a contrabass. I don't have to manipulate slides to play that instrument, and neither do I have to lip notes around or even pay that much attention to intonation. But that instrument lacks the projection and propagation of a bigger F tuba.

There's no guarantee that the competing objectives overlap to find a solution space. In fact, there's every reason to believe that there is no solution space. Every instrument is most likely a compromise, with the goal of being really good at some things without being too bad at others.

Rick "whose excellent modern zooms are 2-4 stops slower, much larger, and much heavier than the primes they purport to compete against" Denney

[Donn]

They want an F tuba to be an F tuba, and they have a very refined notion of what that is. We want an F to be like a small C with an easy upper register.

How about, we want an F to be an Eb?

My Yamaha 621 has excellent intonation and a low C that feels like a contrabass. I don't have to manipulate slides to play that instrument, and neither do I have to lip notes around or even pay that much attention to intonation. But that instrument lacks the projection and propagation of a bigger F tuba.

I was hoping someone would bring up the 822. Watch people talk about that model in this Tubenet thread from October, and decide for yourself how the market looks for a perfect tuba. (Not that I'm saying it's perfect - they're saying it's perfect. `darned nice horn overall, despite the extremely consistent sound', for example.)

[Rick Denney]

I was hoping someone would bring up the 822. Watch people talk about that model in this Tubenet thread from October, and decide for yourself how the market looks for a perfect tuba. (Not that I'm saying it's perfect - they're saying it's perfect. `darned nice horn overall, despite the extremely consistent sound', for example.)

There was also "the Yamaha is easier to play; the 45SLP is easier to listen to." And not everyone praised the 822's intonation in that thread.

I have this feeling that European F tuba players have a whole different perspective on playing an F tuba. For them, the low C probably feels right (because they know how to blow it) and it probably feels wierd on a contrabass.

The 822 (and the 621) are great F tubas for what they are, but the 621 has more application for me. It's a great chamber instrument. The 822 is biggish for that, with too much of a C-tuba sound. The B&S has an F-tuba sound that goes well with the small groups, but it also has the projection to stand up to a large ensemble. I'm glad I have both the 621 and the B&S--they complement each other.

As to Eb tubas, I've played several. The Willsons don't have the low fourth-valve problem of the F's, but that's true for both the Willson Eb and the F, and even the rotary F. But they are more like the 822 in terms of a broad, contrabass sound.

I think that, as with the body on which the 621 is based, Yamaha went to school on Besson. The Eb Sovereign doesn't seem to have the color of sound that I heard from top soloists on F tubas. But it does have the weight to support a big ensemble.

Again, those who wanted their F to be like a contrabass with an easy upper register, the 822 would do it for them. The Willson, and various Eb tubas, would do the same. That describes me in some circumstances--I'm making no complaint. But optimizing an alternate instrument for players who spend most of their time on contrabasses may de-optimize it for players who plan to spend nearly all their time on the F.

Once again, "perfect" is both undefinable and unattainable.

Rick "viva l'difference" Denney

[MartyNeilan]

Their manufacturing skill is little better than Canon's (Canon lenses can be beautifully made when they want them to be), but their line for rejection is much higher.

I currently live in the South, the land of "It'll-be-alright." I had a few temp jobs while in college and just recently one making auto parts on the graveyard shift. It doesn't matter how expensive the machines are or how well engineered the product is, if "anything goes" on the line. Most recently, at the auto parts plant (for a FOREIGN branded company!) we were making a certain drivetrain component that had to be within critical tolerences. Every 20th one was inspected with computerized measuring equipment, documented, and plotted on a graph so the production run can be analyzed.
Here is the kicker:
If that part was not within spec, we were told to measure the next few until we found one that was in spec, and use that number. If none of them were within spec, we were told to put it down as being within the highest range of acceptability. Whenever I would show something to a supervisor that looked wrong, the response would always be "It'll-be-alright."

I temped at a different plant 2 weeks one summer, that was covered with "quality" signs everywhere. The regular employees just scoffed at the notion of qualty and didn't give a darn about what left the gates, or even if the counts were correct. At least those particular parts were non-critical and wouldn't leave you stranded on the side of the road within 30,000 miles.

[Kevin Hendrick]

... It doesn't matter how expensive the machines are or how well engineered the product is, if "anything goes" on the line.

I temped at a different plant 2 weeks one summer, that was covered with "quality" signs everywhere. The regular employees just scoffed at the notion of qualty and didn't give a darn about what left the gates, or even if the counts were correct. At least those particular parts were non-critical and wouldn't leave you stranded on the side of the road within 30,000 miles.

Exactly right! QS-whatever certification doesn't mean JS if the employees are trying to dance past it and fake their way through it, instead of making an honest (do I expect too much?) effort to implement it ...

[iiipopes]

Ed Deming, the quality control expert for the ages, would roll over in his grave!

http://www.deming.org