Why does brass sound different than steel?

[ttonon]

Greetings,

 

Because of questions and comments by Chris and others in the post, Why do Brass Tongues Break?, I took a very practical approach looking for answers.

 

The main question here is, why does brass sound different than steel? The complete theoretical/mathematical solution to the forced vibration problem is too difficult for me, but there may be clues if we look both at the dynamical problem of vibration and the static problem of cantilever bending.

 

The reasoning goes like this. From the dynamical solution for a vibrating cantilever, we can calculate the frequency of vibration, given the beam geometry (length and thickness) and its material properties, Y/rho, where Y is Young’s modulus and rho is material density. From the solution for a cantilever bending in static load, we can calculate the beam deflection, given the beam geometry (length, thickness, and width), material properties, Y, and by assuming some arbitrary force.

 

The reason for looking into what the static solution has to offer is because the first mode dynamical solution for free vibration does not fix the amplitude of vibration. The amplitude is arbitrary, because there’s nowhere in the formulation a mechanism involving amplitude. But it’s intuitive to me that the relative stiffness of the reed tongue should affect the results in the complete, forced vibration case, and so, in how the tongue sounds, and the simplest place to look for such information is in the static solution.

 

With this approach in mind, I find the following results, calculated from these two approaches, in which we compare the behavior of two beams having in all cases the same length and frequency of vibration:

 

1) The beam with smaller Y or larger rho requires a larger beam thickness (for the same frequency)

2) For the beam with a smaller Y and the same rho, its static deflection will be LESS, for a given load, and this counter intuitive result occurs because the increase in thickness required to maintain the same frequency (1) over-rides and causes the beam to be stiffer!

3) For beams having the same Y, the beam with double the density of the other requires 41.4% greater thickness (for the same frequency), resulting in 2.83 more total mass, and it has 35.4% of the static deflection of the other beam

4) For beams having the same rho, the beam with Y one-half of that of the other requires 41.4% greater thickness (for the same frequency), resulting in 41.1% more total mass, and it has 70.7% of the static deflection of the other beam

5) For the same length and frequency, brass has 38.9% more thickness, 50.6% more total mass, and it experiences only 66.4% of the deflection of a steel beam, for any given equally applied static load!!

 

To me, these results are surprising, and I had been biasing myself into thinking the steel tongue would be stiffer. Again, for clarity: The (constant area) brass tongue of a free reed having the same length and pitch as a (constant area) steel tongue is stiffer than the steel tongue.

 

Now begins the conjecturing. Perhaps what this means is that, for the same bellows pressure, the swing amplitude of the brass tongue will be smaller than that for the comparable steel tongue. If this is true, since the frequency is the same, the velocity amplitude of the brass tongue will be less than (perhaps only 66.4%) that of the steel tongue. If that’s true, the steel tongue, traveling faster, produces sharper pulses of air in its motion, and it’s understandable that its tone will have more higher harmonics than the tone of the brass tongue. It might also give argument to why steel might have higher playing volume than brass.

 

There’s also two additional conclusions arising from this conjecture. Since the steel tongue isn’t as stiff as the brass tongue, it will have more tendency to bend in a more complicated shape, as a result of possible excitation of higher vibrational modes, again resulting in higher harmonics. Finally, the higher velocity of steel results in larger aerodynamic forces, which may again mean higher playing volume, and which in turn are better suited to excite the higher modes in a more flexible tongue.

 

In conclusion, if the conjecture is correct, we can perhaps well understand why steel sounds brighter than brass.

 

Best regards,

Tom Tonon

www.bluesbox.biz

[ttonon]

Now begins the conjecturing. Perhaps what this means is that, for the same bellows pressure, the swing amplitude of the brass tongue will be smaller than that for the comparable steel tongue.

After a couple hours break from this topic, I realize that this conjecture is only a wild goose chase. I think it's equivalent to saying that you could predict the relative amplitudes of two vibrating spring/mass systems by only measuring the spring constants of each. No go.

 

Sorry for the digression, but I'll leave the post because it might stimulate more thought and the 5 points I listed are still valid.

 

Best regards,

Tom

[Chris Ghent]

Tom,

 

thanks for this. I was expecting the brass to be stiffer than the steel for hands on feeling reasons, but the concept of the brass tongue traveling less far and therefore for a given frequency not needing to travel so fast seems like a great theory. Much better than trying to imagine the effect on timing and speed of different rate reversals, though these might in fact still occur. And the added partials from a higher passing speed do explain the different sound. Apart from greater deflection in steel reeds allowing greater passage of air it is possible to partly explain the higher volume also in the sense that the higher speed creates more high partials, and higher partials contribute much to perceived volume. Your last paragraph suggests you mean more than this and I will get to this.

 

I kindly mention that your distinction, “fast” and “slow,” I think is misplaced. If brass and steel tongues with the same length vibrate at the same frequency and at the same amplitude, their velocity amplitudes must be the same.

I do find myself not understanding why the concept of fast and slow reeds does not sit well with you. While adding same amplitude to this sentence means they cannot be a different speed, in your example above you talk of a theoretical deflection of brass being 66% of steel, leading to a speed difference which we can see our way through to being the agent for creating the different sound.

 

I asked my friend with the brass reeds yesterday if he had a strain gauge we could use to measure deflection pressure in the brass reeds he has. Unfortunately he does not. It is way too crude relying on reeds that are not in the same place and cannot be compared too well anyway. Dana has a reed deflection pressure measuring machine and might be persuaded to enter this discussion.

 

In the light of your findings I feel the need to further describe the crude tests my friend and I did. Comparing the length, width, profile and maximum deflection of brass v steel reeds, it seemed to me they were all very similar. To elaborate on that, the reed lengths for middle C were within .3mm (12thou) over 26mm (slightly more than an inch, 1024thou or so. I called that the same size as it seemed like not much of a difference. The deflection, now looking like a considerably more important measurement, was 2.9mm in the brass and 3.1 in the steel. Again I thought this to be very little. Very approximately it is about 6% and probably within any accuracy fluctuations in the method, it is a difficult measurement to do accurately without a careful setup, which neither of us had.

 

The profile is the other important comparison, and here things get trickier and I feel the need to be more discursive. Reeds come in three variants, wedges, bars and concaves (my terms). The wedges and concaves are in order to create a compression to the length scale of reeds which would otherwise become to long in the bass and too short in the treble. The bars, very rarely perfectly so, occur around about F to G above middle C. In the case of my concertinas the G would typically be a bar about 9 or 10 thou thick. In my case they have been hand filed to this, so will have some fluctuation, perhaps .5 of a thou, plus possible deviation for fine tuning. Others who use grinding technology, which is capable of extraordinary accuracy, might have in comparison a near perfect bar.

 

The Wedges are above this, from this point F to G above middle C to as high as you like. As you go up the scale they become increasingly thinner at the tip and wedge shaped. I might point out there are two ways to make wedge reeds, 1) the straight wedge with a single line from the root down to the tip thickness, and 2) the wedge that descends to a particular thickness and then this thickness continues as a bar to the end of the reed. It is possible to make a wedge with a commencing bar and then a quicker wedge shape. This is a way of coping with a frame that is longer than desirable for a higher pitch, better to get a shorter frame.

 

The concaves are in the lower register and have a concave scooping out of the thickness between the root and the tip. As the pitch lowers the depth of this concavity become greater and the thickness of the tip builds. There is much room here for different profiles. The narrower the thinnest part of the reed is the more it will bend at that point no matter what the rest of the reed is. ie. if the reed is in thou thickness measured every 40 thou, 12,12, 12, 10, 10, 12, 12, 12, you will get a lot of your deflection from this narrow area, the reed will be weaker and you will probably need to remove more tip thickness to get to pitch. The concavity is sometimes a bar section, as in 12, 12, 12, 12, 12, rather than 14, 13, 12, 11, 11, 12, 13, 14,The different profiles here range between wider thicker concavities or bar segments and narrower thinner ones.

 

Sorry if I am detailing things you already know but I think some good might come of it.

 

When comparing the brass and steel profiles at middle C the thickness of the concavity was very similar, and reasonably bar like at 10/11 thou in both. The brass reed however was lighter at the tip. Not by a measurement that sounds like a lot, 3 thou, but in reed terms it is a lot and perhaps I should not have described the profiles as being the same. This difference would make the reed stiffer.

 

Since the steel tongue isn’t as stiff as the brass tongue, it will have more tendency to bend in a more complicated shape, as a result of possible excitation of higher vibrational modes, again resulting in higher harmonics.

 

And here lies one reason I was so long winded on the subject of profiles. When making a reed I often lift up the tip and look at the deflection curve the reed makes. If it is not a pleasing curve, a very unscientific measure I am afraid, then I will thin the reed in the appropriate place to make it so. Making reeds you are slave to many masters apart from pitch; getting the right tip thickness, thinning the dead zone to remove stiffness, and then this pleasing curve. A non pleasing curve often has a flat spot in it, and it is a matter only of taking material from the flat spot, it could be as little as a thou, and the reed will regain a nice curve.

 

When I look at reeds I made starting out I sometimes come across a reed which disobeys all of my current "rules" but seems to work. The sort of thing I am talking about, and this is to exaggerate, at 40 thou intervals starting at the root,15 (the starting bar thickness) 14, 13, 10, 13, 12,10, 9, 8, 7, 6, 5, 5, 5, 5. As you can see in this "wedge" reed there is a "weakness" at 120thou out from the root. This weakness will induce a secondary vibration (this is an assumption on my part) in the reed, which can be brought to pitch but will be weaker in volume than hoped.

 

I need to bring in another concept. My current model of extracting maximum power from a reed demands a predictable curve from the reed because I want, at the instant at which the reed enters the frame, for the reed to be absolutely straight so the hole in the frame is 'sealed' as much as possible, and if this is achieved maximum compression will create the biggest possible "power stroke". Assessing the curve of the reed when adjusting the set of the reed helps in knowing what shape to make the set. I believe I can do this for any material, ie brass or steel.

 

The reason why I described the reed with the thin part above as weaker is because the reed may not travel through the frame in a pattern which will necessarily produce a straight line at that moment.

 

This brings me to your...

Since the steel tongue isn’t as stiff as the brass tongue, it will have more tendency to bend in a more complicated shape, as a result of possible excitation of higher vibrational modes, again resulting in higher harmonics. Finally, the higher velocity of steel results in larger aerodynamic forces, which may again mean higher playing volume, and which in turn are better suited to excite the higher modes in a more flexible tongue.

 

I think what I am saying is,(and only if I understand you sufficiently, never guaranteed) I am working to keep these higher vibrational modes at a minimum, coping of course only with the affect of my own actions in filing. Perhaps what I am saying is, I find it hard to believe from a hands-on point of view, that material differences will contribute as much seeming complication to vibrational modes as my small slip in filing. For any material filed into a wedge shape the simple bending characteristics can only be related to the thickness and will be obvious to the filer and compensated for in the set. If the curve is linear then any effect the different thicknesses might create must be linear and predictable regardless of the density of the material or its stiffness.

 

I should add here that it is my current belief that only the action of the reed passing the compression point creates partials, and should some odd shivering vibration passing down the reed not prevent the reed from all passing through the frame at the same time then there will be no effect on the sound of the reed. Should the vibration make the reed pass through the frame in a strange pattern then there may be partials produced but they will not be as loud because the reed will not produce the same amplitude and therefore the fastest speed. This is not as clearly expressed as I would like but I cannot devote any more time to it today.

 

Sorry if all of this was long winded, but translating the practical, things maybe thought but not expressed, to the page is difficult. I did it because I feel you have something great here, and I look forward to being able to exploit it in my work. The key phrase from my thoughts is "only the action of the reed passing the compression point creates partials" and I feel your work on the stiffness of the two materials completely explains the difference in sound between brass and steel, providing the sound of the brass reed is in fact a subset of that of a steel reed, and that the difference in maximum deflection between the materials is born out. Phew...

 

I hope some of my reasoning finds favour with you, and look forward to your reply, particularly in regard to "possible excitation of higher vibrational modes".

 

Chris

Edited August 18, 2012 by Chris Ghent

[Chris Ghent]

Now begins the conjecturing. Perhaps what this means is that, for the same bellows pressure, the swing amplitude of the brass tongue will be smaller than that for the comparable steel tongue.

After a couple hours break from this topic, I realize that this conjecture is only a wild goose chase. I think it's equivalent to saying that you could predict the relative amplitudes of two vibrating spring/mass systems by only measuring the spring constants of each. No go.

 

Sorry for the digression, but I'll leave the post because it might stimulate more thought and the 5 points I listed are still valid.

 

Best regards,

Tom

Tom,

 

I cross posted on this, so have replied to your previous post before you recanted!

 

Cheers

 

Chris

Edited August 18, 2012 by Chris Ghent

[Johann]

Hi Tom and all others interested in this subject.

 

I did explore this in the past as well.

 

And my conclusion on it is that Models of explanation are to simple.

 

A lot of research power goes in this days into Laser technology, and if one compares the results and this complicated Models how Light in different substances "vibrates"

then it expectable that mechanical vibration of different beam materials is as complex as different substructure for LEDs are.

 

The main missing part in the extreme simple models we still use for vibrating beams of a tongues is that attenuation, feedback, and interaction of different modes is nearly not even present in this formulas.

 

Also keep in mind that we don't have measurement tools that can measure the different, because or ears are far better in distinguishing the different sound as any technical instrument is.

 

So what ever the experience of the people working with reeds are, trust your "subjective" judgement even you don't have an explanation for it.

 

I hope i could express me well enough, English is not my native language. Most of the new theoretical background I did gain is from the university of Regensburg.

 

http://www.physik.uni-regensburg.de/forschung/schwarz/

 

http://wayback.archive.org/web/*/http://www.physik.uni-regensburg.de/forschung/schwarz/*

 

 

Best regards, Johann

[Chris Ghent]

 

So what ever the experience of the people working with reeds are, trust your "subjective" judgement even you don't have an explanation for it.

 

Johann,

 

I think this is what we all do, but any clues from elsewhere are welcome...

 

Chris

[ttonon]

Hi Chris,

 

Thanks for your interesting response, which provides stimulating detail.

 

Here’s where I think we have similar views: My recantation on my original post questions only the assumption that the brass tongue has smaller amplitude than the steel tongue, and this assumption may indeed be true, but I do insist, still in agreement with my own recantation, its validity cannot be justified solely by calculations of beam bending in static load.

 

If you compare two spring/mass/damper systems in forced motion, there’s no way you could predict how their vibrational response will compare by only looking at their spring constants, which is all that the static calculation gives you. In addition, we don’t even have a complete theoretical formulation of the problem, with both the “forcing function” and “damper” left undefined. It’s thus my present understanding that the amplitude of vibration, an important feature of the reed response, is theoretically inaccessible to us, and we must presently hope only for experiment to provide us with that clue.

 

Thus, I think that it would be very useful to have some vibration amplitude measurements done, and I don’t think it would be difficult to perform them on a steel tongue and a brass tongue, each of constant area, with the same length and frequency, vibrating under the same static air pressure difference, with all other relevant elements (cavity, air passages, etc.) identical. I can envision the two reeds mounted over the same opening from which a blower sucks air. Vibration height can easily be measured by standing a scale next to the tongue tip, which is suitably illuminated. I think such measurements would be more fruitful – concerning musical tone – than static deflection measurements.

 

I appreciate all the points you make concerning profiling, which indeed illustrates how the reed maker must delve into issues that would seriously confound the theoretician. I do point out, however, if the cross sectional area of the tongue changes by no more than about ten percent, and if the thickness varies by no more than about twenty percent, and if these changes are close to linear along the length of the tongue (“wedge” and “concave” shapes), we do have access to a theoretical solution that may provide insight.

 

If, however, we keep things as simple as possible, which is probably best from a theoretical view, the question why brass sounds different from steel can best be answered by working with and thinking in terms of “bars.” This all does lead me to an obvious question: Considering all the profile shapes you work with, is it still true that brass sounds different from steel, and is it still definitely true for bars?

 

I do find myself not understanding why the concept of fast and slow reeds does not sit well with you.

I retract my objection. In writing it, I misunderstood you to say that details within the sinusoidal motion of the brass and steel tongues of same frequency and amplitude cause one to be faster in moving through the slot.

 

When comparing the brass and steel profiles at middle C the thickness… …This difference would make the reed stiffer.

 

I’m pleased that you can experimentally confirm to some extent my calculated predictions. As I stated in my recantation, the five points of my original post are valid, to which I add another here:

 

6) For beams having the same ratio Y/rho, their thicknesses will be the same (for the same length and frequency), their total mass will be in the same ratio as their density, and the static deflection of the one with lower (higher) Y will be larger (less) than that of the other.

 

These six points follow from some very basic relationships, but of course, they apply strictly to tongues of “bar” shape, in which case, you can place great confidence in them.

 

My current model of extracting maximum power from a reed demands a predictable curve from the reed because I want, at the instant at which the reed enters the frame, for the reed to be absolutely straight so the hole in the frame is 'sealed' as much as possible, and if this is achieved maximum compression will create the biggest possible "power stroke".

 

I acknowledge this approach seems rational, although I do not see a way to theoretically prove it. I do recall that harmonium tongues (and possibly free reed organ tongues) are made with a downward curve, so that at rest, the tip is very nearly, or a bit, into the slot, with a region of the tongue between the tip and the root above the slot. I don’t know the reason for this. If you can track down some of these makers, you might share notes. There are these people out there, and I was in touch with one years back, but I lost my contact info.

 

The other point again relates to information we can obtain from static deflection measurements vs what happens in the dynamic vibration. You are making a static measurement here. For a bar, the first mode vibration requires that the tongue shape along its length be straight and level, and this mode is what happens in practice, to first order. For tongues with generally a lot of profiling, this may not be true in all cases, at least theoretically.

 

I am working to keep these higher vibrational modes at a minimum, coping of course only with the affect of my own actions in filing.

 

As you may know, for vibrating bars, the frequencies of higher modes of vibration are, in general, not integer multiples of the fundamental, in contrast, for instance, to (infinitely flexible) strings. Thus, in principle, contributions to the sound from these modes will be, what we normally call, noise. With skill and practice, and by means of techniques similar to your technique of profiling, these higher modes can be tuned close to some musically pleasing multiple of the fundamental, usually one involving whole number (Pythagorean) ratios. Thus, bells, xylophones, and marimbas can be musically tuned. Having said that, however, a little bit of noise can be a good thing, as in the case of these instruments and others, including the piano. So, in general, how do you know that you really want to eliminate all trace of these higher modes? At this point of our ignorance, I don’t think we even know whether these higher modes are necessary for the interesting difference in sound between brass and steel. It thus may be impossible to completely eliminate them, except for the lowest playing volumes. I have great respect for the human auditory system, and so I’m not so willing to state limitations on what it can perceive. I’m of course not objecting to your approach, but rather merely pointing out the complexities you conjure – at least to me - by such an approach.

 

Perhaps what I am saying is, I find it hard to believe from a hands-on point of view, that material differences will contribute as much seeming complication to vibrational modes as my small slip in filing. For any material filed into a wedge shape the simple bending characteristics can only be related to the thickness and will be obvious to the filer and compensated for in the set. If the curve is linear then any effect the different thicknesses might create must be linear and predictable regardless of the density of the material or its stiffness.

 

By “linear” curve, I assume you mean the smooth static shape you explained earlier. Thus, I again comment that the vibrational response can produce shapes that do not conform to conclusions derived from static measurements, and this in itself, I think, proves that material properties are as important as thickness. It’s also true that material properties limit the extent or shape of the profile modifications you make, in order to produce a “desirable” static shape. For a given tongue length, you are restricted on the profiling you can do for a brass tongue, in order to make it have the same pitch as a steel tongue.

 

I should add here that it is my current belief that only the action of the reed passing the compression point creates partials…

 

I’m guessing that by “compression point” you mean the tongue position when it is moving downward and just entering the slot, and that your belief is that this moment is when the characteristic sound pressure pulse is created. If your contention is correct, wouldn’t there then be only one pressure pulse emanated per swing cycle? As can be readily seen from microphone traces, this is not true. In general, there are two distinct pressure pulses created per swing, as can be seen by Figure 3 in the attached document. There are also other “wiggles” in pressure pulses. It’s probably better for you to first read this short paper, before I comment further.

 

… and should some odd shivering vibration passing down the reed not prevent the reed from all passing through the frame at the same time then there will be no effect on the sound of the reed. Should the vibration make the reed pass through the frame in a strange pattern then there may be partials produced but they will not be as loud because the reed will not produce the same amplitude and therefore the fastest speed.

 

Because I think your basic contention here is not supported by experiment, I won’t comment much on this last statement, which I do think has problems, and I again emphasize that conclusions on what the vibrating tongue is doing based on static measurements can be misleading.

 

 

 

Best regards,

Tom

The motion of air-driven free reeds, Cott Lill Reed 1999.pdf

[Johann]

Hi all,

 

You also my look at this ( http://www.ifm-zwota.de/tonzun09.pdf ) document in German language.

Some interesting graphs are presented, but the main conclusion of this document is also that the Axioms used for modelling limit all predictions.

 

best regards, Johann

[ttonon]

Hi all,

 

You also my look at this ( http://www.ifm-zwota.de/tonzun09.pdf ) document in German language.

Some interesting graphs are presented, but the main conclusion of this document is also that the Axioms used for modelling limit all predictions.

 

best regards, Johann

 

Hi Johann,

 

Is it too much to ask you to summarize the main purpose, approach, and conclusions of this interesting looking paper? Thanks.

 

Regards,

Tom

[Chris Ghent]

Tom,

 

I agree vibration amplitude measurements are relatively straightforward, though I would like to see dynamic height measurements for the whole reed, not just the tip.

 

from me

I do find myself not understanding why the concept of fast and slow reeds does not sit well with you.

From you

I retract my objection. In writing it, I misunderstood you to say that details within the sinusoidal motion of the brass and steel tongues of same frequency and amplitude cause one to be faster in moving through the slot.

I am now having problems with my own idea of slow and fast reeds. The problem is every "fast" reed can be played softly, at which point it would move as a "slow" reed and if speed is the major creator of difference in sound then that difference should be apparent when a steel reed is played quietly. Is it?

 

Your point about harmonium reeds is a good one. The twisted and bent harmonium reeds would definitely flout my rules for a well behaved reed. Those harmonium and reed makers were desperate to make their registers sound different to each other. They had access to an excellent air supply, otherwise many of those oddities could never have worked. The ones that I have here have a approx 30 degree twist to the reed about halfway along. This would certainly interfere with partial creation. I did once go looking for harmonium reedmakers but only found people who's basis for work was to copy what once was. They tend to be older people who are not, or were not very on the net, perhaps it is different now.

 

from you...

 

As you may know, for vibrating bars, the frequencies of higher modes of vibration are, in general, not integer multiples of the fundamental, in contrast, for instance, to (infinitely flexible) strings. Thus, in principle, contributions to the sound from these modes will be, what we normally call, noise. With skill and practice, and by means of techniques similar to your technique of profiling, these higher modes can be tuned close to some musically pleasing multiple of the fundamental, usually one involving whole number (Pythagorean) ratios. Thus, bells, xylophones, and marimbas can be musically tuned. Having said that, however, a little bit of noise can be a good thing, as in the case of these instruments and others, including the piano. So, in general, how do you know that you really want to eliminate all trace of these higher modes? At this point of our ignorance, I don’t think we even know whether these higher modes are necessary for the interesting difference in sound between brass and steel. It thus may be impossible to completely eliminate them, except for the lowest playing volumes. I have great respect for the human auditory system, and so I’m not so willing to state limitations on what it can perceive. I’m of course not objecting to your approach, but rather merely pointing out the complexities you conjure – at least to me - by such an approach.

 

Quite right, consider me chastened.

 

from you

I’m guessing that by “compression point” you mean the tongue position when it is moving downward and just entering the slot, and that your belief is that this moment is when the characteristic sound pressure pulse is created. If your contention is correct, wouldn’t there then be only one pressure pulse emanated per swing cycle? As can be readily seen from microphone traces, this is not true. In general, there are two distinct pressure pulses created per swing, as can be seen by Figure 3 in the attached document. There are also other “wiggles” in pressure pulses. It’s probably better for you to first read this short paper, before I comment further.

 

Right, I see what you are saying. First, I did see the compression point as happening twice in each cycle. In my " theory" the partial creation might be greater in the second compression point as the wind speed would be higher when the reed was moving in the opposite direction to the air. Having seen the Cottingham graph and the other pressure changes you mention, the "wiggles", very interesting, and what could produce those? My first thought was returning pulses in the same sense as the returning echo pulses which control tuning in two stroke motorcycle exhaust systems. If you are not familiar with these, it was discovered in East Germany in the 50s that the open end of an exhaust pipe sent a reverse signal, a pulse, back up the exhaust pipe. If this pulse arrived at a bad moment in the cycle it could result in poor mixture in the cylinder. By tuning the exhaust, meaning making it longer or shorter, the pulse could be timed to arrive at a moment when it could not interfere with the scavenging after the power stroke. This is a terribly simple to the point of nonsensical explanation. What fascinated me when I first heard this is that an open end on a pipe could send a return signal. Still don't understand it, one of these days I will look it up and it will probably be easy to find somewhere in google land. It was the moment two strokes became the flavour of the moment. Anyway, there are plenty of opportunities for return signals in concertinas, and perhaps this might account for some of the wiggles and for some of the other "noises".

 

 

from me

… and should some odd shivering vibration passing down the reed not prevent the reed from all passing through the frame at the same time then there will be no effect on the sound of the reed. Should the vibration make the reed pass through the frame in a strange pattern then there may be partials produced but they will not be as loud because the reed will not produce the same amplitude and therefore the fastest speed.

 

from you

Because I think your basic contention here is not supported by experiment, I won’t comment much on this last statement, which I do think has problems, and I again emphasize that conclusions on what the vibrating tongue is doing based on static measurements can be misleading.

 

Fair enough..!

 

Cheers

 

Chris

[ttonon]

Hi Chris,

 

The problem is every "fast" reed can be played softly, at which point it would move as a "slow" reed and if speed is the major creator of difference in sound then that difference should be apparent when a steel reed is played quietly. Is it?

 

I think your point is valid, and it indicates some of the complexities here; however, if we find that two reeds of different materials but of the SAME length, pitch, and bellows pressure have different speeds (presumably because of different amplitudes) there may be differences in sound spectrum as a consequence. Thus, speed by itself is meaningless, but speed in a context where other parameters are held constant may provide clues.

 

The ones that I have here have a approx 30 degree twist to the reed about halfway along. This would certainly interfere with partial creation.

 

I’m not sure what you mean by “interfere with partial creation,” but I’ll guess that you conclude that the harmonium reed sounds less bright because of the bend in its tongue. Is this according to your experience? If so, it might be an interesting clue on how the beast works. Some simple experimentation could answer this question. If on the other hand, your statement results from a belief that such a shape simply cannot produce many partials, then I must ask for the evidence for it.

 

Having seen the Cottingham graph and the other pressure changes you mention, the "wiggles", very interesting, and what could produce those? My first thought…

 

In this graph, I have trouble making sense out of the phase difference between the tongue displacement trace and the two pressure traces. The graph seems to be telling us that the sound pressure just above a vibrating tongue gets a positive pulse when the tongue just exits the slot, traveling upward. This is counter intuitive, since, for the reed to work with the microphone placed as stated, positive air pressure must be imposed on the same side of the tongue that the microphone is placed. If so, why would the pressure pulse be positive and not negative? I can see that it would be positive, only if the cavity were pressurized, but the reed would then not work. The author doesn’t specifically state that the phase relationships of this graph are correct, as occurring during the experiment. Without such a statement, we really don’t know if the tongue displacement/pressure phase, as displayed, is correct. But that indicates a sloppy presentation, and I’m a bit doubtful that a researcher would do that; especially since modern data acquisition equipment can indeed record correct phases.

 

I think your interpretation of the wiggles of pressure pulses inside the cavity is reasonable. The phenomenon you describe associated with motorcycle exhaust pipes is the same phenomenon by which most any musical instrument with a resonator works, when operating in sustained mode of vibration. A very close example is the beating reed organ pipe, or say, a clarinet. A positive pressure pulse that moves along the length of the air column is timed to arrive at and strike the reed, opening it at the precise moment energy input, by means of a pulse from the musician’s breath, is required for periodic motion. The original positive pressure pulse is thus intensified as it rebounds and travels back towards the open end, with some of its energy passing into the room, but also much of its energy reflected from the open end as a rarefaction (low pressure) pulse, which in turn travels back along the tube and arrives at the reed when the reed is closed, helping to keep it closed, and rebounding again as a rarefaction, which travels back to the open end and reflects as a positive pulse, completing the cycle. This action is inherent to what we call tuned resonators. The situation inside the free reed cavity is similar in that there are pressure pulses bouncing off walls, but there’s a fundamental difference because the air of the cavity is not tuned, or resonating as a coherent mass in a way (except by happenstance) that is tuned to the motion of the tongue, thus the more complicated shape in the pressure wave form. The effect is more like the sound you hear bouncing off the walls in an empty room.

 

Best regards,

Tom

[Chris Ghent]

Hi

 

The ones that I have here have a approx 30 degree twist to the reed about halfway along. This would certainly interfere with partial creation.

 

I’m not sure what you mean by “interfere with partial creation,” but I’ll guess that you conclude that the harmonium reed sounds less bright because of the bend in its tongue. Is this according to your experience? If so, it might be an interesting clue on how the beast works. Some simple experimentation could answer this question. If on the other hand, your statement results from a belief that such a shape simply cannot produce many partials, then I must ask for the evidence for it.

 

It is true I think of a reed which travels through the compression point all at the same time as a "perfect" reed and anything else is not good enough; this is consistent with my singleminded pursuit of my current model of thinking on what makes a good reed, and that assumption leaked all through my statement above about the harmonium reed! Asking for evidence is fair enough from a scientist and I don't have any.

 

Cheers

 

Chris