Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Owen Woods]

So I've been learning a few bits and pieces about box acoustics over the last few months. Today I finally got round to reading some stuff which had been sitting in my favourites for months after someone linked it to me on this forum. They are a few papers on box acoustics, especially on reed blocks.

First is a paper on resonance in chambers by Tom Tonon (of pitch bending accordion fame). Here it is: http://www.concertina.org/pica/pica_2005_2/html/reed_cavity_design_resonance.htm. The main conclusion of it is, as far as I see it, that Helmholtz resonance of the chamber is a Bad Thing, but that quarter wave or full wave could boost volume of the reed and amplify certain harmonics, if a full or partial partition was placed in the chamber. This is interesting, but if you look at the table at the bottom it becomes clear that both quarter wave and full wave resonance is exponential (as expected), meaning that it's application to reed blocks in it's simplest form is impractical. One could "tune" blocks to exhibit quarter or full wave resonance at different partials, meaning that every chamber would cause some kind of resonance, but since different partials would be excited for every reed it is quite likely that the resulting sound would be haphazard and disjointed.

It would however be interesting to see if (say) a single octave instrument could be constructed with a block large enough at the low end and small enough at the top end to consistently excite the quarter wave resonance. I think it would be possible, but you'd have to tabulate all of the lengths so as to match to plates as closely as possible. And of course there would be little practical application.

The conclusions are backed up by this paper, which concludes that "For a given reed, general characteristics of the spectrum seem independent of cell [i.e. chamber] length" (http://www.public.coe.edu/~jcotting/acoustics.html). However it does note that increasing the chamber length can amplify the 1k-3k frequencies, which is potentially useful. However, since doubling the length of the reed chamber results in an increase in response of around 1.5 times, the effect isn't hugely significant over the range of practical chamber sizes. What is more interesting is that the presence of a chamber does change the sound of a reed, no matter what the size of the chamber is. This sounds similar to putting a loudspeaker in a baffle, which radically and immediately changes the sound. In fact the bellows act as a baffle - the sound of a reed inside the bellows is very different from the sound of a reed mounted outside the bellows [That might be wrong, this is based on a lecture that I was mostly asleep in].

So it appears that the function of reed block tapering is purely to ensure that there is enough air going into the reed so that it doesn't choke. Resonance is likely to be a bad thing, either because it interferes with the reed (which is bad) or because it results in uneven amplification and timbre across the instrument. Changing the dimensions of a chamber will affect the sound of the reed, but in a fairly negligible way.

What about mounting? Flat mounted reeds produce a very different sound from block mounted reeds. Part of this is due to a different transient response (i.e. attack), but there is more to it I think. I can think of two different mechanisms. One is that by distributing the force on the reed in a different way, the motion of the reed changes somehow and affects the sound. As the dominant frequency in the reed's vibration is the fundamental and the sound is based on pressure fluctuations caused by the reed blocking off and unblocking off the air supply, I'm doubtful of how much this would change the sound, especially since most traces I've seen show the movement of the reed as pretty much sinosoidal. Although modes of vibration of the reed do appear in the spectrum when excited by air (http://www.public.coe.edu/~jcotting/ICA2007%20Free%20Reed.pdf, it seems unlikely that they would necessarily translate into sound. The other explanation is that by altering the air intake the flow pattern changes, meaning that the air coming out of the reed has a different direction and velocity. Again though, the frequency is dependent on the period of blocking and unblocking - what effect would different flow actually have?

My problem now is that I have lost sight of how reeds differ, except by their fundamental mode of vibration. Nobody that I can find has ever studied different kinds of reed for the same instrument of the same pitch. So if the sound of boxes is independent of soundboard and is independent of reedblock, what makes good reeds sound good and bad reeds sound bad?

That might be unintelligible, but it is nearly 4am and I have a project meeting about stringed instrument acoustics in 5 hours.

[Theo]

So if the sound of boxes is independent of soundboard and is independent of reedblock, what makes good reeds sound good and bad reeds sound bad?

This statement does not follow from your previous discussions, its false logic.  Your arguments clearly describe why some possible effects on sound production are not going to happen, but there may be others.   One could test the statement of yours that I have quoted with some fairly straightforward experiments in a test rig that simulates a real instrument but where reedblocks and/or soundboards of different designs and materials can be substituted. 

My own personal hunch is that the materials and construction of reedblocks is critical, but I'm afraid I don't understand the factors involved well enough to design the ideal reedblock.  One small clue is in the design of the Castagnari Mory.  I'm not alone in believing that the models made prior to the late 1990s are superior players to the ones made around 2000 and onwards (though recent examples are better).  The early Mory (identifiable by the semicircle design on the grill) have reedblocks carved out of a single piece of timber with the grain running lengthwise, with a cap and base glued on.   Later ones are built up from a mid rib with chamber walls glued on.   The most highly regarded Paolo Sopranis also have their blocks made in the same manner as the early Mory. I don't know what this means, but there could be something there.   What I suspect is important is the ability of the reedblock to hold the reed rigidly and to resist any vibration of the reed plate, thus ensuring that as much energy as possible is retained in the vibrating reed.

I have one other hunch, nothing stronger than that, about the difference in sound from a flat mounted reed as opposed to one on a block is, at lease in part, simply a result of the site where the sound is produced is closer to the outside world, therefore the emerging sound has a shorter path with less opportunity for being modified by interaction with the surfaces it passes.

[pete /acorn]

Hi,
I think Theo is correct,I've been doing a lot of looking and comparing recently but more with piano accordions,the reed block material effects the sounds greatly,also with piano boxes a lot of the well respected older makes use leather under the reeds and pin the reeds on.
I think with the reed  flat onto the soundboard,as Theo said makes the sound sharper.A good example of how sounds change,get hold of a cassotto box,especially one with a cedar cassotto chamber,the resonance makes the bassoon reed in a PA sound richer very like a Bassoon and if the clarinet reed is in the cassotto chamber this then produces a sound more akin to a wooden clarinet,obviously the resonance of the chamber has a great influence on the sound of the instrument.

Pete

[Theo]

A good example of how sounds change,get hold of a cassotto box,especially one with a cedar cassotto chamber,the resonance makes the bassoon reed in a PA sound richer very like a Bassoon and if the clarinet reed is in the cassotto chamber this then produces a sound more akin to a wooden clarinet,obviously the resonance of the chamber has a great influence on the sound of the instrument.

I fully agree that the cassotto chamber makes a difference to the sound, but I don't think it is correct to describe it as being produced by "resonance of the chamber".  Ukebert has given a good summary of why resonance is almost certainly not a beneficial thing in reed block acoustics.  What does happen in the longer and larger chamber such as a cassotto is that there is a more complex pattern of reflection of sound waves within the chamber which can cause a variety of interactions between partials, and a much greater opportunity for absorption of sound by the chamber walls.    I strongly suspect that it is this increased internal reflection of sound that is responsible for the different sound character.

[mory]

Theo Quote [I have one other hunch, nothing stronger than that, about the difference in sound from a flat mounted reed as opposed to one on a block is, at lease in part, simply a result of the site where the sound is produced is closer to the outside world, therefore the emerging sound has a shorter path with less opportunity for being modified by interaction with the surfaces it passes.]

When I place a reed flat on my tuning bellows either on a flat/half block or straight on the surface they are always much louder than any on upright blocks, I always thought of it as the angle of attack if you like with the air coming from behind the tongue on the flat mounted and end on on the upright racks, maybe to do with turbulence I suspect, same loud volume in a Connie for same reason regardless of other contributary factors AtB mory

same happens with a mouthie comb

[Theo]

Good points Mory.

As a scientist by training I know it is important to distinguish between things we can actually observe and the separate step of explaining why the differences are there.  For example we can easily observe that flat mounted reeds sound different from reeds on an upright block, but explaining why is a much more complex business. Between us we have  come up with some ideas of what we guess might be going on, now we need someone like ukebert to do some investigation and careful experiments to try to get a working theory that we can use to predict which design factors determine the kind of sound we want.

[blafleur]

It is a mind numbing but fascinating thing.  There are quite a few variables other than just the reed block that also come into play.  Sound hole size, chambers varnished or not, bellows size, and of course the reeds themselves and how they're set up.  I've wondered how much the taper matters.  And maybe you can explain why blocks are tapered larger at the bottom of the reed to smaller at the top.  And, I wonder, what is optimum taper?  I use 2 1/2 degrees, not for any particular reason, I just ended up there.  Seems every bit of testing I do is contradictory.

Randy Falcon has done some testing of different size/shape reeds of same pitch and has some definite opinions on which he prefers, but like Theo pointed out, explaining why is another matter.  And deciding which one you like best can be completely nullified by not being able to buy them in that configuration.

I was convinced reed block material made a big difference in the sound(and still hold on to that some), even though a highly respected builder told me it didn't, until I made one with balsa blocks, which has just as rich and loud a sound as the ones I make out of hard maple.  I did make some with a wood harder than hard maple, and I didn't like the sound. It seemed more shrill and less rich.

I really think accordions laugh at people trying to apply human logic to them.  They're baffling creatures.

[mory]

I think the tapered block (bottom to top) is an exercise in getting the air flow behind the tongue as in the flat mounted reed and is a real necessity in an upright block to get a quick response to the reed even if it lacks a little in volume by comparison AtB mory

[pgroff]

I think the tapered block (bottom to top) is an exercise in getting the air flow behind the tongue as in the flat mounted reed and is a real necessity in an upright block to get a quick response to the reed even if it lacks a little in volume by comparison AtB mory

This is something I was also considering, after our recent discussion of the Baldoni and V. Soprani reedblocks shown here:

http://forum.melodeon.net/index.php/topic,8225.msg109766.html#new

PG

[diatonix]

As a builder I do not worry too much about what to make my reed blocks from. My personal choice is alder: it works easily, it is stable, resistent and not too heavy, glues well. It is also relatively inexpensive. I also used abachi in two or three batches, but continuous exposure to its dust may lead to health issues...
My priorities: Sound improves noticeably when the blocks are glued straight on the board (no bottom, no gasket, no screws). Large sound holes. Following this simple recipe you should get more or less the same result as with flat mounted reeds.

[pgroff]

As a builder I do not worry too much about what to make my reed blocks from. My personal choice is alder: it works easily, it is stable, resistent and not too heavy, glues well. It is also relatively inexpensive. I also used abachi in two or three batches, but continuous exposure to its dust may lead to health issues...
My priorities: Sound improves noticeably when the blocks are glued straight on the board (no bottom, no gasket, no screws). Large sound holes. Following this simple recipe you should get more or less the same result as with flat mounted reeds.

I know some electric guitar enthusiasts who claim to hear a difference between alder and ash for solid-body instruments (and prefer alder).  

Regarding the second point, you can sometimes hear the difference cited by diatonix in the same instrument -- that is, in accordions and concertinas that have one or more reedblock glued to the soundboard, and another one or more reedblock removable, with gasket and screws.  As an example, in the photo below, the box on the left has two removable reedblocks and the notes from the inside row are a little more muffled than from the outside row of buttons, probably due to the position of pallets down close to the keyboard.  The box in the middle has the same arrangement of pallets, but the notes from the inside row seem brighter than from the outside row -- probably because the reedblock for the reeds sounded from the inside row of buttons is glued in place with no gasket.  Some modern german-made 20 key concertinas* also have a mixture of glued-in and removable reedblocks and the same phenomenon can be noted.

PG

*for example, the model for sale at the top of this list:


http://www.elderly.com/vintage/cats/160U.htm

[Owen Woods]

This statement does not follow from your previous discussions, its false logic.  Your arguments clearly describe why some possible effects on sound production are not going to happen, but there may be others.   One could test the statement of yours that I have quoted with some fairly straightforward experiments in a test rig that simulates a real instrument but where reedblocks and/or soundboards of different designs and materials can be substituted.

You are of course, quite right! My statement was a rather unscientific piece of frustrated rhetoric. My problem is that I feel that many of the more conventional (in musical instrument acoustics terms) ways of defining the sound have been disproved and I can't think of adequate mechanisms to explain the rest.

My own personal hunch is that the materials and construction of reedblocks is critical, but I'm afraid I don't understand the factors involved well enough to design the ideal reedblock.  One small clue is in the design of the Castagnari Mory.  I'm not alone in believing that the models made prior to the late 1990s are superior players to the ones made around 2000 and onwards (though recent examples are better).  The early Mory (identifiable by the semicircle design on the grill) have reedblocks carved out of a single piece of timber with the grain running lengthwise, with a cap and base glued on.   Later ones are built up from a mid rib with chamber walls glued on.   The most highly regarded Paolo Sopranis also have their blocks made in the same manner as the early Mory. I don't know what this means, but there could be something there.   What I suspect is important is the ability of the reedblock to hold the reed rigidly and to resist any vibration of the reed plate, thus ensuring that as much energy as possible is retained in the vibrating reed.

I don't think that the material of the reedblock makes much difference, but I think that the construction of the reedblock may well have an effect. Certainly, as you say, the vibration of the reedplate (or rather the lack of it) is very important, as is the lack of vibration of the block. As for multiple pieces, I'm not sure how that would affect the vibration of the block - probably decrease some modes and increase others, so it might well be that the construction that you state gives a more rigid block. With some thought a multiple piece reedblock could possibly be designed such that vibration was selected against, as the glue join probably would constitute a block to vibration. Talking out of my hat here you will have guessed, it's a complex shape and I've got little experience with how it would behave. The material might have some effect in how the block vibrates, but I think that the construction dominates.

I have one other hunch, nothing stronger than that, about the difference in sound from a flat mounted reed as opposed to one on a block is, at lease in part, simply a result of the site where the sound is produced is closer to the outside world, therefore the emerging sound has a shorter path with less opportunity for being modified by interaction with the surfaces it passes.

I think that this is probably true. But I don't think that it's the whole story. I think that the difference is about air flow and I think (at the moment) that it is air flow that makes a reed good or bad. The fundamental is dependent on reed geometry, but my hunch is that the shape of the resultant waveform and perhaps more importantly the way that the reed behaves as it is being excited is dependent on the flow out of the block, which would primarily be based on the reed shape and the path of the air.

I fully agree that the cassotto chamber makes a difference to the sound, but I don't think it is correct to describe it as being produced by "resonance of the chamber".  Ukebert has given a good summary of why resonance is almost certainly not a beneficial thing in reed block acoustics.  What does happen in the longer and larger chamber such as a cassotto is that there is a more complex pattern of reflection of sound waves within the chamber which can cause a variety of interactions between partials, and a much greater opportunity for absorption of sound by the chamber walls.    I strongly suspect that it is this increased internal reflection of sound that is responsible for the different sound character.

This is possible, but I dunno... I do think that the finish of a reedblock makes a difference, but something doesn't seem quite right, that mechanism doesn't quite sit right with what's in my head. I suppose because I'm still getting my head around the radically different form of sound generation that a reed affords. I think that it's more helpful to think about flow than sound, i.e. a cassotto makes a different sound because the flow does a dogleg. Ready to be called out on that though.

When I place a reed flat on my tuning bellows either on a flat/half block or straight on the surface they are always much louder than any on upright blocks, I always thought of it as the angle of attack if you like with the air coming from behind the tongue on the flat mounted and end on on the upright racks, maybe to do with turbulence I suspect, same loud volume in a Connie for same reason regardless of other contributary factors AtB mory

I think that this is probably the right answer (see above about air flow). What would be interesting would be to see how much of the difference in sound is transient and how much isn't. This could be quite simple, given enough time to set up a repeatable experiment. i.e. set up a air chamber with a compressed air supply and release a valve which triggers a microphone to start recording (preferably in an anechoic chamber) and which will allow air to pass through the reed. The microphone will record the sound coming out until it stabilises (i.e. after a pre-defined period of time) at which point the flow would be swiftly ramped down and once the reed has stopped vibrating the microphone would be turned off. Put the resulting data through a FFT and you could get a sonogram, which would show how the frequency spectrum and amplitude changes over time. The trick would be comparing it to a block, but identical chamber dimensions with reeds from the same manufacturer but the only difference being the location of the hole... yes, I'd say that it would be possible. And that would give you a direct comparison, albeit only for one note and you'd have to design the chamber well such that you wouldn't have any problems with resonance.

I think the tapered block (bottom to top) is an exercise in getting the air flow behind the tongue as in the flat mounted reed and is a real necessity in an upright block to get a quick response to the reed even if it lacks a little in volume by comparison AtB mory

I think that this is an excellent explanation. It would not surprise me in the slightest if this is the reason and the explanation would make it much easier to design good blocks.

As a builder I do not worry too much about what to make my reed blocks from. My personal choice is alder: it works easily, it is stable, resistent and not too heavy, glues well. It is also relatively inexpensive. I also used abachi in two or three batches, but continuous exposure to its dust may lead to health issues...
My priorities: Sound improves noticeably when the blocks are glued straight on the board (no bottom, no gasket, no screws). Large sound holes. Following this simple recipe you should get more or less the same result as with flat mounted reeds.

That observation doesn't surprise me actually and probably the difference between mounting on blocks and flat is partly due to the reedblocks being properly secured. The securing method for reedblocks is pretty suboptimal and if it wasn't for ease of tuning I would think that every block would be glued on, it makes much more sense from a vibration point of view. There will be give in the gasket, which means that there are a whole selection of block modes that just wouldn't be available if they were glued down. Even if they were glued down though there would potentially be more modes available for the blocks than there would be for the pan. The experiment that I detail above wouldn't be affected by that, since both block and pan would only hold one reed and would be clamped, so if that would not yield any data then the clamps would be a potential explanation.

What is frustrating is that I know exactly where in my department there is a lab where such an experiment could be carried out. There is an anechoic chamber and everything. I even know that there is a rig that could be adapted for this purpose. I know all this because it's the basis of the fourth year project of a close friend of mine. For this reason though it is impractical to do anything, as the rig is most definitely in use and will be for the foreseeable future! Plus of course I'm busy enough with my own project (on stringed instrument acoustics, which are a lot more straightforward). Would be an interesting thing to do though and it wouldn't be completely impractical to do it as an amateur, although you'd have to be very careful with noise interference when measuring the sound output from the reed - particularly if you are using a compressor. And you'd need a copy of MATLAB, which is a bit pricey.  

[Anahata]

You may be on to something with the mention of starting transients. It's well known that much of the recognisable character of the sound of a musical instrument is in the way the notes start. Snip those off recordings of a flute and a trumpet playing the same note, for example, and they sound surprisingly similar. The speed of starting also affects the feel of the instrument for the player, and is a major difference between good reeds and bad reeds.

[blafleur]

I do not strongly doubt the argument that it doesn't too much matter the material of the block, and my recent balsa block project helps confirm that, but if so, why does it make such a drastic difference if the reed is on an aluminum or brass base?  Why shouldn't the material that the plate is attached to also make a difference?  I also think there is a distinction between what make no difference and what can make a little difference. A little difference can be the difference between preferring this accordion or that one.

With my last 2 projects, I also helped confirm Diatonix's report that a bigger sound hole is better, which also kills what I had previously thought.  I assumed that a hole needed to be big enough, but too big would use too much air.  So I made square holes, which added quite a bit more surface area than my round holes.  I can tell no difference in air consumption, but a noticeable difference in volume and richness of sound. Still can't help but think there is a point of diminishing returns.

Back to blocks. Also as an experiment, I made my blocks on another accordion (with my conventional maple blocks) with dual level chambers.  The thought was since the push/pull reeds are set the space of the plate apart, I would try to take up that space under the push reed. First impression was inconsistent.  Some reeds seemed louder, and some choked.  So I tried shaving down the top of the raised portion under the tip of the push reed, problem solved.  My next 3 will also have these stepped chambers, so hopefully I'll be able to see if it is consistently better than what I had been doing.

My experiments have very little scientific thought, but a whole lot of "what if..",but is definitely the fun part for me.

[Theo]

IBack to blocks. Also as an experiment, I made my blocks on another accordion (with my conventional maple blocks) with dual level chambers.  The thought was since the push/pull reeds are set the space of the plate apart, I would try to take up that space under the push reed. First impression was inconsistent.  Some reeds seemed louder, and some choked.  So I tried shaving down the top of the raised portion under the tip of the push reed, problem solved.  My next 3 will also have these stepped chambers, so hopefully I'll be able to see if it is consistently better than what I had been doing.

You are in good company there.  Castagnari have made their reed blocks like that for many years.

[blafleur]

You are in good company there.  Castagnari have made their reed blocks like that for many years.

Ha, I assume that everything I try has probably already been tried.  Do they mill theirs out or do they glue in strips for the step up?

[Theo]

They are milled.

[IanD]

I do not strongly doubt the argument that it doesn't too much matter the material of the block, and my recent balsa block project helps confirm that, but if so, why does it make such a drastic difference if the reed is on an aluminum or brass base?  Why shouldn't the material that the plate is attached to also make a difference?  I also think there is a distinction between what make no difference and what can make a little difference. A little difference can be the difference between preferring this accordion or that one.

With my last 2 projects, I also helped confirm Digitonix's report that a bigger sound hole is better, which also kills what I had previously thought.  I assumed that a hole needed to be big enough, but too big would use too much air.  So I made square holes, which added quite a bit more surface area than my round holes.  I can tell no difference in air consumption, but a noticeable difference in volume and richness of sound. Still can't help but think there is a point of diminishing returns.

Back to blocks. Also as an experiment, I made my blocks on another accordion (with my conventional maple blocks) with dual level chambers.  The thought was since the push/pull reeds are set the space of the plate apart, I would try to take up that space under the push reed. First impression was inconsistent.  Some reeds seemed louder, and some choked.  So I tried shaving down the top of the raised portion under the tip of the push reed, problem solved.  My next 3 will also have these stepped chambers, so hopefully I'll be able to see if it is consistently better than what I had been doing.

My experiments have very little scientific thought, but a whole lot of "what if..",but is definitely the fun part for me.

It seems generally that the heavier and more rigid the reed plate mounting is the better; brass reed frames are a lot heavier than aluminium ones (more mass), glued-in blocks are better than ones mounted on gaskets (more rigidity). Also big area pallet holes in a thin pallet board are better (my Oakwood used big rectangular holes), as is a clear open path for air flow to the reed. Bigger holes don't use more air, by far the smallest cross-section in the air flow path is the reed in its frame.

So it'll be interesting to see what flat-mounted reeds (short air path) Allen-bolted down to a milled aluminium "reed pan" with no gaskets sound like, the reed should think it's mounted directly on a big rigid metal block and you can't much get more massive and rigid than that... 

Ian

[pgroff]

Hi Ian,

I have wondered about integral, milled soundboard/reedplates, but thought of the plates waxed on. Probably your idea of screwed-on reedplates would be more secure in case of temperature changes.

I have played a highly customized Hohner Amatona (a 4 voice D/D#) years ago, owned by a very creative and handy guy.  The reedblocks had been modified to have metal bases (aluminum or stainless, I can't remember which), machined to exactly mate surface-to-surface on a metal soundboard, and locked in place to yield an airtight metal-to-metal fit with no gasket.  This box had a really unique, rich, bright and present sound, but it had many other modifications also (for example the original grille was off and an open, curved metal "collar shape" installed) and I don't know what it sounded like before the reedblocks were modified.

PG

[pgroff]

You may be on to something with the mention of starting transients. It's well known that much of the recognisable character of the sound of a musical instrument is in the way the notes start. Snip those off recordings of a flute and a trumpet playing the same note, for example, and they sound surprisingly similar. The speed of starting also affects the feel of the instrument for the player, and is a major difference between good reeds and bad reeds.

Anahata, this observation (and ukebert's) is crucial -- thank you! It might deserve another topic.  But in my opinion, this factor also is highly influential in the differences among playing styles developed by good musicians who play dance music.  Some musicians seem to have developed a style using very high bellows pressure to get quick response and good articulation of rhythms at a micro time scale, and this can work with reeds that are less than optimally efficient and/or somewhat leaky instruments. You can achieve the accuracy without harshly overblowing the reeds on the attack.  Other musicians have a lower-pressure style that depends on extremely efficient reeds and very tight instruments to achieve those great dance rhythms.  Each approach can sound accurate and tonally beautiful, on the appropriate instruments for each style.

PG