Jump to content


Photo

Timbre


  • Please log in to reply
58 replies to this topic

#37 David Barnert

David Barnert

    Ineluctable Opinionmaker

  • Members
  • PipPipPipPipPipPip
  • 3051 posts
  • Gender:Male
  • Location:Albany, NY, USA

Posted 11 May 2005 - 01:10 PM

For what it’s worth I carried out a small experiment and these are my findings.

On removing both ends of my 46 key Lachenel Maccan I determined that the reed shoes for the G above middle C on the treble and bass sides of the instrument were the same size.  I swapped them over and could detect no difference in how well they fitted.  I reassembled the concertina and measured the pitch on my Korg CA-30 meter.  Having measured the pitch prior to dismantling I found no measurable change in pitch.  The timbre did not transfer with the reeds.

If you're looking for a change in timbre, pitch is not what you want to measure, as they are unrelated. See previous posts.

#38 JimLucas

JimLucas

    Ineluctable Opinionmaker

  • Members
  • PipPipPipPipPipPip
  • 10127 posts
  • Gender:Male
  • Location:Denmark

Posted 11 May 2005 - 02:04 PM

Having measured the pitch prior to dismantling I found no measurable change in pitch.

If you're looking for a change in timbre, pitch is not what you want to measure, as they are unrelated.

<{POST_SNAPBACK}>

I'm just guessing, David, but tony may have said that as a response to my comments in the Moving Reeds Into Different Slots topic. I don't think he meant it as a comment on timbre, but on moving reeds. He moved his reeds with a different purpose, but reported that he did not see the other effect that I warned about... that the tuning would likely be slightly off.

#39 Greg Jowaisas

Greg Jowaisas

    Heavyweight Boxer

  • Members
  • PipPipPipPipPip
  • 1519 posts
  • Gender:Male
  • Location:Kentucky, USA just south of Cincinnati and the Ohio River

Posted 11 May 2005 - 03:01 PM

The way I read Tony's explanation of his reed swapping experiment I thought he was saying that the timbre did not change with the reed. (the pitch incidentally remained constant)

Which "might" warrent an exploration of the relationship between chamber size and timbre.

Tony?

#40 JimLucas

JimLucas

    Ineluctable Opinionmaker

  • Members
  • PipPipPipPipPipPip
  • 10127 posts
  • Gender:Male
  • Location:Denmark

Posted 12 May 2005 - 08:24 AM

For what it’s worth I carried out a small experiment and these are my findings.

<{POST_SNAPBACK}>

This has become a somewhat muddled Topic. Tony appears to have responded here to comments made in the Reed Voicing Topic, where I believe the context was quite different. I'm going to try to sort out what I think are the relevant points before I present the results of my own "experiment".

.. 1. Tony reported an observed change in his instrument from before to after it was retuned, and he wondered if such a change could have been the result of something done in the retuning, in particular by an accordion specialist who didn't fully understand the differences between accordions and concertinas.
.. 2. I said yes, I thought that was possible. I based that in part on my own observations of improperly done "repair" work, especially that done on reeds. But a necessary assumption was that such a significant change in tone quality had actually been introduced.
.. 3. Others have suggested that the tonal difference between the ends was there all along, but that tony didn't notice it before the retuning, perhaps because other factors dominated. If that's the case, then my comment is irrelevant, because it would be speculating on the cause of something that didn't happen. There is another possibility, but I'll get to that after the results of the experiments. The important point is that...
.. 4. Neither I nor anyone else who responded to the query claimed that inexpert retuning would necessarily alter the timbre, much less in any particular way, and we most particularly did not claim that the profile of the reed is the only factor affecting timbre, even though it is the one most susceptible to alteration by a retuning job.

Now to the experiments:

On removing both ends of my 46 key Lachenel Maccan I determined that the reed shoes for the G above middle C on the treble and bass sides of the instrument were the same size.  I swapped them over and could detect no difference in how well they fitted.

I thought I would try to match this on my 55-button Lachenal New Model Maccann. To my surprise, I couldn't, as the shoes and reeds for that G on the right hand were larger than those on the left hand and wouldn't fit into the left-hand slots. The RH reed tongues are 26 mm long, and the LH ones are only 22 mm.

I reassembled the concertina and measured the pitch on my Korg CA-30 meter.  Having measured the pitch prior to dismantling I found no measurable change in pitch.

Though I was unable to switch the reeds between the ends, I was able to swap them between the two sides of the reed pan in a single end. I don't have a fancy meter, but I could hear a slight difference in pitch between the two reeds when they were in the swapped positions, yet no audible difference when on their proper sides of the reed pan. Was the fact that tony's didn't shift slightly a lucky coincidence? I don't know.

The timbre did not transfer with the reeds.

I would not have expected it to. If there was a difference in timbre due to different profiling, I might expect some effect, but I would expect other differences -- particularly the chamber dimensions -- to be significant or even dominant.

I then measured the dimensions of the chambers.  These are my findings:

Treble side (in millimeters): Length (two sides): 38 and 34.  Width (two sides): 17 and 7.  Depth (two sides): 10 and 10.  From this I calculate the volume to be approximately 6120 cubic millimeters.

Bass Side (in millimeters):  Length (two sides): 37 and 34.  Width (two sides): 15 and 6.  Depth (two sides): 9 and 8.  From this I calculate the volume to be approximately 4526 cubic millimeters.  This means that the volume of the chamber on the treble side is something like 35% larger than the volume of the chamber on the bass side.

I'm not sure what formula you're using to compute volume. I get much lower figures. But since I get nearly the same ratio (treble side 35% "larger"), we can work from that.

My own comparable measurements are:
Right Hand
.. 2 sides both 37 mm long
.. width 9mm at inner end, 16 mm at outer end
.. depth 8 mm (one measurement, in a spot with neither slot nor reed)
Left Hand
.. sides 33 & 38 mm long
.. width 7 mm at inner end, 15-6 mm at outer ends of each side
.. depth 7 mm
I figure that to give the treble side about 32% more volume.

What else? Well, my instrument has tilted reed pans, shallower at the high end than at the low end. The chamber depth at the high end of the right hand is 5 mm, while that at the low end of the left hand is 11 mm. I've been told by several experienced sources that this helps balance the tone (timbre) and volume between the lower and higher reeds, and that deeper chambers give "richer, mellower" sound. My own experience agrees with that. The hand rest is cut out underneath, and there are slots cut into the edge of each side (a feature mentioned in another recent topic), features which might be intended to counter the effect suggested by Paul Groff, that reeds located below the hand/handrest are relatively muffled. One more thing: The RH reeds seem to have a flatter and thinner profile than the corresponding LH reeds.

What about the timbre? If I strain, I think I can detect a slight difference between the ends in the range of overlap, with the left end being slightly "brighter". That would be consistent in direction with what tony reported, but not in magnitude. It's so slight that I wouldn't notice if I weren't looking for it. I would say that the both loudness and timbre are well balanced between the ends.

What do I conclude? If, as has been suggested by more than one experienced repairer, the difference in timbre between the ends is a characteristic common in duets, then whatever causes the difference must be somehow counteracted in my instrument. The main differences I've noted are 1) the tilted reed pans, 2) the additional openness -- under the hand bar and around the edges of each end, 3) the larger size of the treble-side reeds, 4) the slightly different chamber sizes, and 5) the profiles of the reeds are different.

What effects are these likely to have?
.. 1) The purpose of tilted reed pans is to vary the depth of the reed chambers, with shallower chambers for smaller, higher-pitched reeds and deeper chambers for the larger, lower reeds. In my instrument, it allows deep chambers for the lowest reeds and shallow chambers for the highest reeds, yet essentially equal depths for reeds of the same pitch on either end. One might expect this to help balance the timbre. A question is whether the chambers on each side of tony's instrument are of uniform depth. If so, I'm surprised that he reports the depth of the left end to be less -- albeit only slightly less -- than that of the right end. As noted above, I would expect that to give a "brighter" or "harsher" (take your pick of adjective) timbre to the entire left end than what would result from deeper chambers. Could that have been deliberate?
.. 2) If Paul Groff's suggestion of the lower right-hand reeds being muffled by their placement under the hand and handrest has merit, then these openings might be expected to result in a brighter sound for those reeds, and thus more balanced timbre between the sides.
.. 3) I'm not sure what effect this should have. It would change the relative dimensions of chamber and reed in a direction which I would expect to give a brighter sound to the reed than if it were in a larger chamber, but I don't know what effect -- if any -- the reed size alone should have on timbre. I can only say that if this difference is deliberate, the experienced craftsmen at Lachenal probably knew what effect it would have.
.. 4) I'll address chamber dimensions in more detail below. For now I'll simply note that tony's instrument and mine seem to exhibit similar differences between the chambers of the two ends, which might suggest that this is not a significant factor in the difference between our instruments.
.. 5) There is a difference in profile, but to my eye not a huge difference. Is it enough to be significant? Is it simply a consequence of tuning reeds of different sizes to the same pitch? I don't know. But if the difference in reed size was for a purpose -- almost a certainty in a high-end instrument with other unusual features, -- then I would expect that the different profiling would at least not have opposed that purpose.

...my suspicions, which are as follows:
1. Changing the profile of a reed tongue in order to alter its timbre is probably a fruitless exercise.

If you mean that it has no effect, I believe the experts have claimed experience to the contrary. On the other hand, they and I have never claimed that it's possible to create any arbitrary timbre at will simply through profiling. In particular, once a reed has been given a particular thickness profile, it becomes severely restricted in the changes that can be made. It's not possible to reverse the removal of metal.

With your own instrument, whether or not the reed profiles were changed by the recent retuning, I would agree that to try now to change the timbre of the one end to match the other in the overlapping range is probably futile.

2. The chamber volume, rather than its shape, has the greatest affect on timbre and, for me, changing this is not an option on a vintage instrument.

Probably not for anyone, but I believe you're wrong about the relative effects of shape and volume. It's a general principle of acoustics (and of lasers and other resonant chambers) that the shape is of extreme importance. A bell shape has a very different timbre from a rectangular bar containing the same amount of metal. A rectangular resonant chamber with dimensions 10 cm x 20 cm x 5m has a fundamental frequency 24% lower than a cubic chamber 10 mm on a side, though both have the same volume. At least as significant is the ratios of their first 8 resonances (each to its own fundamental).
.. the rectangular prism: 1.00 1.07 1.25 1.31 1.81 1.85 1.96 2.00
.. the cube (same volume): 1.00 1.41 1.41 1.41 1.73 1.73 1.73 2.0
(The repeated numbers on the cube are due to its symmetry.)
And the prism's frequencies expressed as ratios to the cubes fundamental are:
.. 0.76 0.81 0.95 0.99 1.37 1.40 1.48 1.51
This is only the tip of the iceberg, since it ignores relative strengths of the resonances, differential absorption by the walls of the chamber, and the fact that the chamber is not a simple rectangular prism. But it does demonstrate that it's the shape of the chamber, and not simply its volume, that affects the frequency spectrum, which is what timbre consists of.

Now returning to a couple of my initial points:
.. 4. Reed profile is certainly not the only factor influencing timbre. Chamber dimensions are probably the most significant, and chamber depth is the dimension I've seen mentioned most often. The materials of the ends and possibly of the reed pan also matter. But of these, only the reed profile is likely to be altered after an instrument is completed.
.. 1-3. Which returns us to tony's original question of whether that could have happened to his own instrument. While others have suggested that the difference in timbre was there but unnoticed before the retuning, I'd like to examine the other possibility, to see if there's anything in the evidence which absolutely precludes it.

If we accept that a reed's profile can significantly affect the timbre -- something which I believe has been affirmed by some makers and restorers and so far denied by none, -- then it seems possible that when an instrument is built, reeds might be deliberately profiled to counteract other factors that would result in inconsistent timbre. Specifically, the two sets of reeds for the overlapping range of the two ends of a duet might be profiled differently in order to give them similar timbres, in spite of other factors. Now what happens if during a full-scale retuning, the profiles of both sets of reeds are flattened, i.e., made more uniform, erasing their differences? Well, the compensation would disappear, and any difference in timbre caused by the other factors would become more evident.

So far, that seems consistent with tony's reports, including the results of his experiment. And it doesn't require us to postulate that there was a difference before the retuning, but that he didn't notice it. His description of the difference after the retuning is quite vivid, and it seems unlikely that such a difference could be overlooked in a playable instrument.

But his measurements leave some questions. The chamber dimensions between the ends are different, but not hugely so, only about 10%. Should that be enough to cause the observed difference in timbre? I don't know. Maybe in combination with the hand and handrest covering the reeds on the one end? All I can say is, "Maybe".

Also, Dave Elliott said, "When the instrument was tuned, I would 'assume' - that dangerous word- that new valves were fitted, and the reeds were properly cleaned and their gapping re-set." So far, tony hasn't told us whether this was the case. If so, then the potential effects of the other changes should be taken into consideration But if the only change to the instrument was treatment of the reeds themselves, then assuming that tony's perception of the timbre was affected by other factors seems inappropriate.

I haven't personally experienced tony's instrument either before or after the retuning. I have to depend on him for infomation about it. So far, I can say that a change in reed profiles during tuning is consistent with the facts as reported, but so are other interpretations. On the separate question of whether new profiling could restore the timbre balance that tony heard before the retuning, I doubt it very much. That's not because I think the profile has no effect on timbre, but because I expect that to get the necessary change in profile would require making the reeds thicker in some places, and I don't know of any way to do that besides making new reeds.

#41 Stephen Chambers

Stephen Chambers

    Ineluctable Opinionmaker

  • Members
  • PipPipPipPipPipPip
  • 4402 posts
  • Gender:Male

Posted 12 May 2005 - 08:24 AM

The timbre of a free reed is a complex issue, it is affected by the composition and thickness of the metal the tongue is made from, the way that tongue is profiled/voiced, and the composition/thickness of the metal in the reed frame.

But, as Tony seems to be discovering for himself (he might even start to agree with me yet ;) ), that is only one element in what makes up the timbre of the instrument. Even given the exact same reeds being used, different chambering would produce a different tone, as does the timber used, the way the reeds are attached to the pan, different materials or designs in the ends of the instrument, the height of the action, and even what the valves are made of !

And as for "tonal balance", or what makes a reed on one side sound different to the same note on the other ... well I would still say that the chambering is the main influence on that, and Tony's conclusions from his "experiment" would seem to confirm it, though I'm a little surprised by the measurements he has provided. With that difference I would certainly expect the left hand G reeds to have a brighter sound than those on the right, as he initially stated, but it is the reverse of the problem I would more commonly encounter.

I've now dug out my own 46-key Maccann (he was a Brummie too, so let's get his name right !), which is a very early example numbered 31 (anybody got an older one ?). It has slightly deeper pans on the left (but not as pronounced as most of the Anglos I work on), but this is compensated for by the chambering extending right up to the inside circle of the pan with no internal division, making the volume much more similar to the other end. I would therefore expect it to have a rather different tonal balance to Tony's, when it is tuned.

#42 Stephen Chambers

Stephen Chambers

    Ineluctable Opinionmaker

  • Members
  • PipPipPipPipPipPip
  • 4402 posts
  • Gender:Male

Posted 12 May 2005 - 11:09 AM

I meant (and thought I wrote) that some high notes on the LH side have pads that open away from the player (hence brighter in timbre), but the low notes on the RH side (which ARE those same pitches, making up the "overlap") may be more likely to have pads that open under the right hand (hence duller in timbre).

<{POST_SNAPBACK}>

Hi Paul,

Thanks for the clarification, now I see what you were getting at, and I've been able to observe it in an instrument too. However, I wonder if it isn't more a question of changing the loudness of those notes, the timbre remaining the same ? I wonder what the wave forms might reveal ?

Mind you, putting in a baffle for those notes probably would change their timbre, as it would kill some of the overtones.

#43 JimLucas

JimLucas

    Ineluctable Opinionmaker

  • Members
  • PipPipPipPipPipPip
  • 10127 posts
  • Gender:Male
  • Location:Denmark

Posted 12 May 2005 - 11:14 AM

...putting in a baffle for those notes probably would change their timbre, as it would kill some of the overtones.

<{POST_SNAPBACK}>

I'm finding this baffling very discussion. :wacko:

#44 Stephen Chambers

Stephen Chambers

    Ineluctable Opinionmaker

  • Members
  • PipPipPipPipPipPip
  • 4402 posts
  • Gender:Male

Posted 12 May 2005 - 11:45 AM

On removing both ends of my 46 key Lachenel Maccan I determined that the reed shoes for the G above middle C on the treble and bass sides of the instrument were the same size.   I swapped them over and could detect no difference in how well they fitted.

<{POST_SNAPBACK}>

As both the slots and the reed shoes are of the same size and were machine-made, this is not surprising.


I reassembled the concertina and measured the pitch on my Korg CA-30 meter.   Having measured the pitch prior to dismantling I found no measurable change in pitch.

Though I was unable to switch the reeds between the ends, I was able to swap them between the two sides of the reed pan in a single end. I don't have a fancy meter, but I could hear a slight difference in pitch between the two reeds when they were in the swapped positions, yet no audible difference when on their proper sides of the reed pan. Was the fact that tony's didn't shift slightly a lucky coincidence? I don't know.

A Korg CA-30 meter is a handy gadget, but hardly an "instrument". You could probably achieve a reasonable "rough tuning" with it, but it would not be capable of providing the kind of exact readings required for "fine tuning".


I then measured the dimensions of the chambers.   These are my findings:

Treble side (in millimeters): Length (two sides): 38 and 34.   Width (two sides): 17 and 7.   Depth (two sides): 10 and 10.  From this I calculate the volume to be approximately 6120 cubic millimeters.

Bass Side (in millimeters):  Length (two sides): 37 and 34.   Width (two sides): 15 and 6.   Depth (two sides): 9 and 8.   From this I calculate the volume to be approximately 4526 cubic millimeters.  This means that the volume of the chamber on the treble side is something like 35% larger than the volume of the chamber on the bass side.

I'm not sure what formula you're using to compute volume. I get much lower figures.

Unfortunately tony hasn't supplied the diagonal measurements, or the angles, without either of which a proper computation cannot be done.

Edited by Stephen Chambers, 12 May 2005 - 12:38 PM.


#45 JimLucas

JimLucas

    Ineluctable Opinionmaker

  • Members
  • PipPipPipPipPipPip
  • 10127 posts
  • Gender:Male
  • Location:Denmark

Posted 12 May 2005 - 11:50 AM

Anyone who states that something does not exist, such as a formulae, belongs to the Flat World Society.

<{POST_SNAPBACK}>

A formula is a description -- usually in some sort of symbolic notation -- of a natural process. It is not the process itself.

Saying that there is no formula is quite different from saying that any formula is impossible.

I said the first, as it is my understanding that no such formula has yet been formulated by humans. I did not say the second. In fact, I said that I believe it would be possible to develop such a formula, and I volunteered to do so, if someone would fund the effort.

If they are saying they know of no such formulae then that's fine...

Thanks. That's what I did.

...but they, at least, could leave the Forum open for someone who may know.

Tony, you seem to be implying that if one person posts, it prevents others from posting. That's not even remotely true. Quite often it has the opposite effect, stimulating others to respond.

#46 ttonon

ttonon

    Chatty concertinist

  • Members
  • PipPipPipPip
  • 182 posts
  • Gender:Male
  • Location:New Jersey, USA

Posted 15 June 2005 - 07:28 PM

2. The chamber volume, rather than its shape, has the greatest affect on timbre and, for me, changing this is not an option on a vintage instrument.

Probably not for anyone, but I believe you're wrong about the relative effects of shape and volume. It's a general principle of acoustics (and of lasers and other resonant chambers) that the shape is of extreme importance. A bell shape has a very different timbre from a rectangular bar containing the same amount of metal. A rectangular resonant chamber with dimensions 10 cm x 20 cm x 5m has a fundamental frequency 24% lower than a cubic chamber 10 mm on a side, though both have the same volume....(snip)


Hi Jim, you have many interesting comments, though I must warn here that one cannot assess the importance of geometric parameters unless one first specifies the mode of acoustic vibration. For instance, Tony's conclusion may indeed be correct if in fact the acoustic cavity mode that is predominate when the reed sounds is when the cavity behaves as a Helmholtz resonator. This is rudimentary acoustic theory at work. In the Helmholtz case, the resonant frequency is given by f = (c/2/pi)*(AL/V)^0.5, where c is the speed of sound, pi the ratio of circle circumference to diameter, A the area of the pallet hole, L the effective length of the air mass vibrating in the pallet hole, and V is the cavity volume. Thus keeping the same pallet hole geometry and changing volume will uniquely change resonant frequency.

In all your examples, the cavity dimensions are significant with respect to the wavelength of oscillation, and you are correct in pointing out that cavity shape is very important for those modes that contain boundaries defined by these shapes. For Helmholtz resonance, all cavity dimensions must be much smaller than this wavelength, in which case, cavity shape is immaterial, and only A, L, and V are important.

One can study the resonance properties of concertina cavities and conclude that the modes that are most likely to occur over the musical range of the entire concertina family are Helmholtz and quarter wave, with Helmholtz predominating. Thus, it's very possible, if not probable, that many of the cavities Tony has studied exhibit Helmholtz resonance, in which the cavity volume as a whole is the significant feature of the cavity itself. For those cavities that exhibit quarter wave resonance, the cavity length is critical (along with A and L), and cavity width is not as critical.

Best regards,
Tom
www.bluesbox.biz

#47 Stephen Chambers

Stephen Chambers

    Ineluctable Opinionmaker

  • Members
  • PipPipPipPipPipPip
  • 4402 posts
  • Gender:Male

Posted 16 June 2005 - 01:24 AM

Tom,

Whilst the volume of the chamber (typically modified by varying its depth, as the length/width is generally more related, for practical constructional reasons, to the dimensions of the reed frame) is undeniably important, it has long been considered by concertina makers that the typical Jeffries/Crabb linear reed pan (with parallel chamber walls) produces a sharper sound, sometimes described as more "honk", than the typical Wheatstone/Lachenal radial one (in which the walls are not parallel) which is considered sweeter. Indeed, those Lachenal models that have linear reed pans do seem to sound brighter too, whilst a Jeffries (English) with a radial pan sounds duller.

On the C.net Home Pages you will find various references to this :

Dana Johnson : "Things like chamber size, shape, method of mounting the reeds ... are only a few of many factors that can play a major role in determining the final sound character."

Steve Dickenson : "... reckons that Wheatstone's original calculations show that it is better to avoid parallel sides."

Richard Morse : "All the Jeffries we've seen have the parallel layout. A lot of the honkiness has to do with the fact that the chambers HAVE parallel sides ..."

Do you have any thoughts on that ?

(Edited to add quotations.)

Edited by Stephen Chambers, 17 June 2005 - 05:59 PM.


#48 ttonon

ttonon

    Chatty concertinist

  • Members
  • PipPipPipPip
  • 182 posts
  • Gender:Male
  • Location:New Jersey, USA

Posted 16 June 2005 - 09:54 PM

Do you have any thoughts on that ?



Hi Stephen,
Your interesting question has many aspects to it, and I’ll try to provide a straightforward response, with more elaboration added on.

I see no reason why, everything else being the same, parallel-sided instruments would produce, throughout the musical range of a given instrument, characteristically different timbre than non-parallel side instruments. I can see specific instances where such differences can exist, involving specific reeds, pitches and sizes, but even here, and from a theoretical point of view, any one geometry can produce brighter tones, for instance, than the other.

When we speak of the effect chamber shapes, sizes, etc. have on musical tones, we are interested the interactions with sound waves over a wide frequency range, and all acoustic systems behave very differently at different frequencies. As I indicated in my previous post, a change of angle between facing walls of a cavity, without changing cavity volume, and everything else being the same, would have minimal acoustic effect on the sound for those frequencies whose wavelength is much larger than all cavity dimensions. In other words, the wave doesn’t see the details of the cavity; it only senses its capacitive effect, in conjunction with the inertial effect of the pallet hole. Such a change in angle would alter the frequency that excites the cavity as a quarter wave tube, which in turn could either cause resonance with some partial of the musical tone, or destroy resonance with some partial of the musical tone. In addition, the walls of the cavity are part of an acoustical system, involving also the other parts of the cavity, the reeds, the pallet hole, the pallet, the box, with baffling and tone chamber effects, etc. It’s difficult for me to see how changing one aspect of this system can always have the same effect on musical tone.

If it is indeed true that parallel-sided instruments produce brighter tones, I suggest that we need to investigate further what other changes in the instrument accompany such parallel sides. For instance, are the pallet holes the same area, do the pallets move away from the hole the same amount, is there greater influence from baffling (or “cassoto”) effects within the body of the instrument, etc., and of course, are there significant differences in the way the reeds themselves are made, scaled, sized, etc.?

Perhaps it’s true that brighter or softer sounds occur with parallel sides for certain pitch regions in a certain instrument, or for certain isolated pitches. Even here, I suggest we would again have to look at all the important acoustic features of the instrument and perhaps measure frequency spectrums at different locations within and outside the instrument in order to understand why. Even then, our study may be inconclusive, simply because of the complexity of the acoustics and the fact that we are dealing with subjective judgments on tone quality.

I read through the comments by the Authors you cited. The statement (from Rich), “A lot of the honkiness has to do with the fact that the chambers HAVE parallel sides AND that this enables the reeds to be size-optimized for response.” is interesting, though I’m not sure what it means. I don’t think Rich intended this statement to explain his full understanding of the topic. But here, I can only ask, Has someone proven that the parallel sides are indeed the reason? But then, what does “size-optimized” mean? Perhaps the parallel sides enable another construction feature that also influences tone, and thus, the effect is only secondarily through the fact that the cavity walls are parallel?

The complete statement (from Steve), “My reed chambers are all parallel, but there are differing views on that too. Steve Dickenson told me that it was simply a matter of construction method, depending on how you secured the reed pan while cutting the rebates for the chamber divisions. I think it improves the tone, but he reckons that Wheatstone's original calculations show that it is better to avoid parallel sides. Some larger Wheatstone English models have a mixture so maybe even Wheatstone himself was prepared to use parallel sides when it suited the layout of the instrument better," indicates that there is some disagreement on the effects of parallel/non-parallel layout. The statement, “Wheatstone’s original calculations show that it is better to avoid parallel sides,” doesn’t mean much to me, since I don’t know what kind of calculations these are and I don’t know what is meant by “better” in “better to avoid.”

I think the main thrust of Dana’s article is to indicate how complicated sound production is in a concertina and that there are no simple blanket statements that can be made on what feature causes what sound. I tend to agree with this. To say that parallel/non-parallel sides have such-and-such effect on tone to me seems like a blanket statement. Perhaps it’s true, and if it is, I don’t understand the physics behind it. Thanks for asking such an interesting question, and although I don’t think I answered it very well, it has allowed me to think more on the topic.

Best regards,
Tom
http://www.bluesbox.biz

#49 Stephen Chambers

Stephen Chambers

    Ineluctable Opinionmaker

  • Members
  • PipPipPipPipPipPip
  • 4402 posts
  • Gender:Male

Posted 17 June 2005 - 01:18 AM

If it is indeed true that parallel-sided instruments produce brighter tones, I suggest that we need to investigate further what other changes in the instrument accompany such parallel sides.  For instance, are the pallet holes the same area, do the pallets move away from the hole the same amount, is there greater influence from baffling (or “cassoto”) effects within the body of the instrument, etc., and of course, are there significant differences in the way the reeds themselves are made, scaled, sized, etc.?

<{POST_SNAPBACK}>

Thanks Tom,

Indeed I have already cited many of those variables myself, earlier in this topic, and there are certainly plenty of other differences than the shape of the chambers between say, Wheatstone (radial) and Jeffries (linear) construction.

Well if, as Jim suggested, somebody was prepared to finance an experimental project ... ? :unsure:


The complete statement (from Steve), “My reed chambers are all parallel, but there are differing views on that too. Steve Dickenson told me that it was simply a matter of construction method, depending on how you secured the reed pan while cutting the rebates for the chamber divisions. I think it improves the tone, but he reckons that Wheatstone's original calculations show that it is better to avoid parallel sides. Some larger Wheatstone English models have a mixture so maybe even Wheatstone himself was prepared to use parallel sides when it suited the layout of the instrument better," indicates that there is some disagreement on the effects of parallel/non-parallel layout.  The statement, “Wheatstone’s original calculations show that it is better to avoid parallel sides,” doesn’t mean much to me, since I don’t know what kind of calculations these are and I don’t know what is meant by “better” in “better to avoid.”

To be fair to Steve, much of what you have quoted seems to be more the opinion of the author, rather than that of Steve Dickinson, and the rest is paraphrased. I only used that small excerpt to give some indication of his thinking (and indeed, the same for the other makers), but if I remember rightly, what he has said to myself is that he feels that the Jeffries-style parallel chambering encourages the development of partials that he regards as undesireable in his (Wheatstone-style) instruments. But I don't think there were ever any formal "calculations" as such.


I think the main thrust of Dana’s article is to indicate how complicated sound production is in a concertina and that there are no simple blanket statements that can be made on what feature causes what sound.  I tend to agree with this.

Likewise, and I heartily concur with his conclusion that :

The short of it is that people who are making instruments that sound like they want them to are doing a lot of things right. Probably more than they realize! It is very difficult to pin down the one thing that makes an accordion or concertina sound like our subjective stereotype of that family. No matter what any of us think the reasons behind it all, What really counts is the quality of the resulting instrument, by accident or design. Tradition plays an important role in passing on the things that worked, whether they were really understood or not. No one, least of all me is the last word on such a subject.

Like I said previously, (at least so far) "it's more like alchemy than physics".

Edited by Stephen Chambers, 20 May 2006 - 10:52 PM.


#50 JimLucas

JimLucas

    Ineluctable Opinionmaker

  • Members
  • PipPipPipPipPipPip
  • 10127 posts
  • Gender:Male
  • Location:Denmark

Posted 17 June 2005 - 04:30 AM

A couple of quick comments. Maybe I'll have time for more after the weekend.

I wasn't familiar with Helmholtz resonance. I've googled, and I think I now understand it, more or less. It makes sense that it could/should be relevant to timbre in concertinas, but I don't believe it's the only relevant effect. Some reasons:
.. 1. The Helmholtz references I've located so far seem to deal only with fundamental vibrational frequencies of the cavities, not with harmonics. But it's the frequencies and relative strengths of the harmonics that comprise "timbre". The fundamental doesn't come from the cavity at all, but from the reed's vibration.
.. 2. Helmholtz resonance seems to be relevant only for frequencies where the wavelength is much longer than any of the cavity dimensions. For the fundamental frequencies of most of the notes on a concertina, that may be the case, but not for all the harmonics. The full wavelength of the 3rd octave above A440 is 47 mm, which is very close to the length (about 45 mm) of that reed's chamber in two different instruments I've just measured, one with parallel chambers and one with radial chambers. In fact, the 2nd octave above A440 is a half-wave (closed chamber) resonance for that length, and the 1st octave above is a quarter-wave (open-ended tube) resonance for the same length.
.. 3. The reed chamber with its open hole is not the same as an open tube, but something more complex. But that's OK, since my point is that this whole thing is more complex that any single simple formula.
.. 4. Acoustic chambers (Stephen, are you acoustic? :) ) also have 2- and 3-dimensional resonances, which are not simple combinations of the linear resonances of individual dimensions, and which I believe can be significantly lower than any of the individual linear resonances.

All of which indicates that there are many different factors involved, and no single simple formula can encompass them all. Empirical measurements are still the only way to develop any reasonable formulation of the relationships among shape, dimension, and timbre. So far, I'm not aware of this having been done in any rigorous and comprehensive manner. Until it is, we will have to depend on the qualitative observations of those persons with considerable experience. What I believe they have universally reported is that the shape of the chambers does affect the timbre, and that particular aspects of shape (e.g., chamber depth) appear to have qualitatively predictable effects on timbre.

Edited to add: The above treats only resonances of an acoustic chamber. There might also be effects due to impedance coupling of the reed pan and the rest of the instrument, differential absorption of the chamber walls and ends, and maybe other things I haven't thought of. I really don't know. But I don't think anyone else knows for sure, either... at least not yet.

Edited by JimLucas, 17 June 2005 - 04:40 AM.


#51 ttonon

ttonon

    Chatty concertinist

  • Members
  • PipPipPipPip
  • 182 posts
  • Gender:Male
  • Location:New Jersey, USA

Posted 17 June 2005 - 05:36 PM

Hi Jim,
Concerning Stephen’s original question, and having given my previous post, I think I can succinctly state my response as follows.

Changing only the angles of cavity walls (parallel vs. non-parallel), without changing volume or other construction parameters, would have negligible effect on cavity response, when such response is that of a Helmholtz resonator. For quarter wave response, or any other response involving standing wave oscillation in the cavity, such a change at most alters the resonant frequency of the response, and I see no way that this change could always produce the same result on tone (e.g., a more brighter tone). These statements apply to any frequency of excitation, including reed fundamentals as well as reed overtones.

In my response to Stephen’s question, I thus see no reason to bring in the various issues you refer to, which to me, would unnecessarily complicate my explanation and perhaps confuse people. These issues are of course interesting and I’m glad you mention them, but before I myself move off the main question, I’d like to be sure I made my simple point.

I have done a detailed study of these broader issues and put them in a paper that is scheduled to appear in the PICA, Volume 2, which is organized by Allan Atlas. In this article, Allan has allowed me to present a detailed quantitative analysis of these issues, and I hope you and other concertina enthusiasts find it interesting.

Best regards,
Tom
http://www.bluesbox.biz

#52 Stephen Chambers

Stephen Chambers

    Ineluctable Opinionmaker

  • Members
  • PipPipPipPipPipPip
  • 4402 posts
  • Gender:Male

Posted 17 June 2005 - 07:56 PM

I have done a detailed study of these broader issues and put them in a paper that is scheduled to appear in the PICA, Volume 2, which is organized by Allan Atlas.  In this article, Allan has allowed me to present a detailed quantitative analysis of these issues, and I hope you and other concertina enthusiasts find it interesting.

<{POST_SNAPBACK}>

Tom,

An excellent journal, I have been known to write for it myself !

I will be very interested to read your paper when it comes out.

In the meantime, I can't help wondering what would happen if you built a Wheatstone using Jeffries-style woodwork, and a Jeffries using Wheatstone-style woodwork, hmmm ...

Edited by Stephen Chambers, 17 June 2005 - 08:29 PM.


#53 Clive Thorne

Clive Thorne

    Chatty concertinist

  • Members
  • PipPipPipPip
  • 271 posts
  • Location:Northamptonshire, UK

Posted 20 June 2005 - 11:07 AM

Changing only the angles of cavity walls (parallel vs. non-parallel), without changing volume or other construction parameters, would have negligible effect on cavity response, when such response is that of a Helmholtz resonator.  For quarter wave response, or any other response involving standing wave oscillation in the cavity, such a change at most alters the resonant frequency of the response, and I see no way that this change could always produce the same result on tone (e.g., a more brighter tone).  These statements apply to any frequency of excitation, including reed fundamentals as well as reed overtones. 

http://www.bluesbox.biz

<{POST_SNAPBACK}>


Tom,

In an earlier post you said that only the length of the chamber is relevent for 1/4 wave resonance., but a cubiod chamber has three dimensions any one of which could cause resonation surely, in addition to the Helmholtz resonance (blowing over a bottle top type resonance for those who havn't understood the above technical descriptions)?. Then, as soon as you get more than one resonance your into the whole feild of difference and summation tones.


Surely making the chamber chambers sides non parallel immediately removes one of these tightly tuned (hi Q) resonators and, at best replaces it with a broad response (Low Q) resonation.
Further, ff the whole reed plate is angled such that the chamber depth varies this removes another highly tuned resonator.

I'm not trying to predict what affect this would be on the timbre, but if the 1/4 wave resonance has any affect on timbre then removing one or two of the resonators will surely have an affect?.

If Helmholtz resonance if the principal audible resonance then, yes, its hard to see that the sides would have much affect.



Another thought:

Adding weight to the end of a reed is bound to make the fundamental note more dominant because it makes the reed closer to the theoretic 'Mass on a weightless spring' model, which would result in a pure Simple Harmonic Motion (ie a sine wave) movement of the end of the reed.
Conversely a purely parallel beam fastened at one end, has a wealth of vibration modes that can can be excited simultaneously.

I don't know which model is the closest to a real concertina reed, and of course neither takes account of the fact that a concertina reed is a damped, forced vibration, and is chopping an air flow. However adding weight at the end will certainly push it towards the former model.



Clive

#54 Clive Thorne

Clive Thorne

    Chatty concertinist

  • Members
  • PipPipPipPip
  • 271 posts
  • Location:Northamptonshire, UK

Posted 20 June 2005 - 11:07 AM

Changing only the angles of cavity walls (parallel vs. non-parallel), without changing volume or other construction parameters, would have negligible effect on cavity response, when such response is that of a Helmholtz resonator.  For quarter wave response, or any other response involving standing wave oscillation in the cavity, such a change at most alters the resonant frequency of the response, and I see no way that this change could always produce the same result on tone (e.g., a more brighter tone).  These statements apply to any frequency of excitation, including reed fundamentals as well as reed overtones. 

http://www.bluesbox.biz

<{POST_SNAPBACK}>


Tom,

In an earlier post you said that only the length of the chamber is relevent for 1/4 wave resonance., but a cubiod chamber has three dimensions any one of which could cause resonation surely, in addition to the Helmholtz resonance (blowing over a bottle top type resonance for those who havn't understood the above technical descriptions)?. Then, as soon as you get more than one resonance your into the whole feild of difference and summation tones.


Surely making the chamber chambers sides non parallel immediately removes one of these tightly tuned (hi Q) resonators and, at best replaces it with a broad response (Low Q) resonation.
Further, ff the whole reed plate is angled such that the chamber depth varies this removes another highly tuned resonator.

I'm not trying to predict what affect this would be on the timbre, but if the 1/4 wave resonance has any affect on timbre then removing one or two of the resonators will surely have an affect?.

If Helmholtz resonance if the principal audible resonance then, yes, its hard to see that the sides would have much affect.



Another thought:

Adding weight to the end of a reed is bound to make the fundamental note more dominant because it makes the reed closer to the theoretic 'Mass on a weightless spring' model, which would result in a pure Simple Harmonic Motion (ie a sine wave) movement of the end of the reed.
Conversely a purely parallel beam fastened at one end, has a wealth of vibration modes that can can be excited simultaneously.

I don't know which model is the closest to a real concertina reed, and of course neither takes account of the fact that a concertina reed is a damped, forced vibration, and is chopping an air flow. However adding weight at the end will certainly push it towards the former model.



Clive




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users