An Interesting Dulcimer Experiment

Just share stories or offer advice

Postby rtroughear » Sat Dec 20, 2008 12:13 pm

Hello all,

I'm back home now, and online again.

Dave: The software I use is Praat V 4204 for offline analysis. It's quite a powerful signal analysis program, but also fairly idiosynchratic. It's an academic program mainly used in speech research. It's free at www.praat.org For realtime acquisition and spectral capture I use Visual Analyser 8.10 It's set up more like a cathode ray oscilloscope and is available free at www.SillanumSoft.com But it doesn't have the signal manipulation capabilities of Praat. I use Audacity for audio recordings. The recordings are just done in my office, with all the reflections that entails. But measured reverberations are low, and mics are close. It's not a research lab, but it doesn't really need to be - I'm looking for general conclusions, not fine details.

Ken: I noted quite awhile ago your comment about the back of concert zithers being the main active board, but I can't find any information about the acoustics of them. If you have any references I'd be pleased to get them. The fretted zithers may have some parallels with mountain dulcimers, and they have been around longer, so it seems reasonable to think there might be actual knowledge of their acoustics. But as it stands, I'm not convinced about the back of a mountain dulcimer being the principal soundboard. I do agree that the traditional MD with full length fret board, and ones with the bridge on the top plate are acoustically different instruments. But how different; who knows? - maybe it's not an either/or situation, but a continuum where the physical acoustics change from predominantly one (full fretboard), to predominantly the other (pin bridge, x-braces, the works) depending on the nature of the construction. All this discussion is only about full fretboard instruments.


I'll briefly describe what was done in this test.

All recordings were made under the same conditions and on the one day - background noise floor was within one dB over all recordings.

A standardised string striker (plectrum on a low friction pendulum) was used to excite the strings. Repeatability trials showed it was OK as a string strike method.

The top was removed in sections, half the upper bout, all the upper bout, half the lower bout, then all the lower bout. ie full top, 3/4 top, 1/2 top, 1/4 top, no top. Then filled with sand, then sand only on the bottom (ie free sides). Then (no sand) light card board top, then newspaper.

For each configuration three strikes of each open string were recorded, and three strikes of the three strings (ie a chord). All results were averaged over the three string strikes. Average and peak SPL and sustain were measured. Tuning was DAdd.

A 20 sec tune (first 4 bars of Wildwood Flower) was recorded for real life average sound pressure measurements and long term average frequency spectra.

The end result of it all was that, within experimental error, there was no change in loudness from full top, to no top, to paper top. There was about 10dB SPL reduction filled with sand - this is a very significant change, and the loss was mainly below 2000Hz, where it seems to matter. With free sides, but back constrained with sand, it was essentially the same as filled with sand - maybe it sounded marginally better.

As a constructional aside, the instrument has been in the no-top configuration for a couple of months now, tuned up, and the string action has not changed. The braces are keyed into the side linings, though.

In summary: Taking the top off a mountain dulcimer, but leaving the fretboard in place, hardly changes the sound level output.

BUT: It did sound better with a top, than without one, so I for one will continue to put them on. I just won't worry as much as I used to about it's parameters (thickness, wood type etc). And bracing will be for strength - ladder bracing only.

Then why doesn't the top matter much to the overall loudness of the instrument? That'll have to wait until another thread on a general sound production model for the mountain dulcimer (which will be largely speculative, so you can all have a go at it).


What might the top do then, if it doesn't control loudness?

I can think of four things:

1. It might add colour to the sound by independantly vibrating weakly at one or more of its bar modes. From measurments of the top plates I have in stock, these might be in the ranges of 20 to 40Hz for the first bar frequency, 50 - 100Hz for the second, 100 - 200Hz for the third and 170 - 350 for the fourth. The first and second are probably too low to be heard. This is pure speculation - I have no evidence.

2. The top encloses the air. With no top all the air resonance vibration modes are lost, and don't contribute to the sound. With a paper top, even though it was fairly floppy, the air resonances between 500Hz and 2000Hz were restored (although shifted in frequency), but not the two strong lower resonances of the full top at 234Hz and 350Hz. These two probably contribute significantly to the sound (and are likely to be what Mike1952 thinks he heard that was missing). A top of moderate stiffness (cardboard) might be enough to enable the 1st and 2nd air resonances to fully develop.

3. The air enclosed by the top also resonates at some frequencies that don't radiate directly from the sound holes. Some of these might couple their energy with the box, maybe the back, maybe weakly, but might add some colour to the overall sound. This is not entirely speculation, but I don't have any evidence for higher air modes coupling to the wood. In my dulcimers the lowest (Helmholz) air resonance does couple with the wood as well as radiate from the sound holes, and that coupling lowers its frequency by about a semitone or a tone.

4. The top physically constrains the sides so the overall box stiffness might rise, with a general energy shift upwards in frequency. This is speculation. The dulcimer in question here doesn't seem to have moved dimensionally without it's top - maybe it over-engineered, or maybe the top doesn't generally contribute in this way.

So, that about does it for tonight. Thanks for everyone looking and commenting.

Richard T
rtroughear
Super Mbr (501-2000 posts)
 
Posts: 601
Joined: Tue Jun 05, 2007 8:18 am
Location: Brogo near Bega, NSW, Australia

Postby Ken Bloom » Sat Dec 20, 2008 1:13 pm

Hi Richard,

Most of the very little that has been done in terms of studying how a concert zither works and reacts is in German. A good place to start is the rather epic and landmark work Handbuch fur Zither by Brandlmeier which was published many years ago. I found a copy in the Northwestern Music Library in Chicago. Zithers.com used have copies for sale. The web site is no longer there. The book covers so many aspects of the zither and its ins and outs. I'm sure there is better stuff available. My congrats on having the intestinal fortitude to do the actual science involved in this. I will be following your results with great anticipation.

Ken Bloom
http://www.boweddulcimer.com
Ken Bloom
Super Mbr (501-2000 posts)
 
Posts: 1817
Joined: Thu Dec 08, 2005 4:55 pm
Location: Pilot Mountain, North Car

Postby mrchips » Sat Dec 20, 2008 4:31 pm

Verry Intesting....... and thanks for the links to that software... They may be better than what I used in the past for something simular about vibration modes and tuning with/for a frind who is slowly turning his whole house into one massive pipe organ!
mrchips
Dulcified! (>2000 posts)
 
Posts: 4586
Joined: Mon Nov 07, 2005 2:23 pm
Location: Fort Worth, Tx

Postby KenH » Sun Dec 21, 2008 12:19 am

Veeeeerrrrryyyyy interesting! A builder could use some beautiful (but thin) veneer of very expensive (in thicker sizes) wood for tops, without losing anything in terms of sound quality. Make a lattice of ribs and stringers and attach the veneer to a parchment backing before gluing to the top frame.
User avatar
KenH
Dulcified! (>2000 posts)
 
Posts: 10147
Joined: Fri Mar 15, 2002 7:16 pm
Location: Afloat in Fort Myers, FL

Postby rtroughear » Mon Dec 22, 2008 4:37 am

Ken H
You could make such a top and I suspect it would sound just as good as a thicker solid top. But could you lean your elbow on it? So I'm not advocating changes to current methods one way or the other. But it would be interesting if someone decided to try it out.

Nothing in the testing of this topless dulcimer, or my comments, addresses the matter of the quality of sound produced. Also, nothing so far is worth considering as a guide to making "good" dulcimers.

As far as my impressions of the quality of the sound goes, which is not something that can really be measured, I'd say this:

The full spruce top sounded OK to me. The general trend as the top came off in stages was to lose some lower frequency warmth, which I attribute to the shifting upwards, then loss, of the low air resonances. Trebles seemed unaffected. However my personal sound preference was for the light cardboard top. The paper top sounded better than no top at all, but not very much better. Filled with sand, either free or constrained sides, it was not worth playing. Overall, if I was stuck on an island, with this dulcimer, without it's top, I'd be happy with it. It could be better, but it IS better, without its top, than some factory dulcimers I've played. I keep getting it down to make sure that I'm not fooling myself, and every time I do I end up thinking "That's not too bad at all, really".

I note that twice as many people have downloaded the video as have downloaded the higher quality audio file. The sound in the video version is very poor and is not as good as the instrument actually sounds on the knee. So it's worth listening to the separate audio file if you haven't already.


HighNoonHunter: Your mishap with the Dremel adds some confirmation to my conclusions - you could have shocked the mountain dulcimer world with the information at the time. I would have believed you.

Folk Fan: Moving the string saddle off the end block does generally increase loudness. In this topless instrument putting a bridge about 4" from the end instead of 1.5" makes it a LOT louder (admittedly in the strum hollow, and subjectively, I didn't measure it). So at least in this one dulcimer the increase in loudness is not to do with moving the bridge nearer to a larger vibrating area of the top. I have an explanation for this loudness increase which I'll include in a more general sound production model later.

Paul C: The braces you use seem to be more like partitions across the box than top and back braces in the traditional sense. Because of their simultaneous connection to the top, back and sides, it doesn't surprise me that they affect the sound - but what is the mechanism? Doesn't really matter if you, and your customers like the sound.

Richard T
Last edited by rtroughear on Mon Dec 22, 2008 5:06 am, edited 1 time in total.
rtroughear
Super Mbr (501-2000 posts)
 
Posts: 601
Joined: Tue Jun 05, 2007 8:18 am
Location: Brogo near Bega, NSW, Australia

Postby harpmaker » Mon Dec 22, 2008 4:51 am

A quick question.

I notice in the video there are what appear to be some sort of cross bars of stock, kind of like ribs, the run from side to side at right angles to the fretboard.

Are those pieces of the original top left in place? Or are they braces that run across and under the fret board? Or both, that is, top material with brace under it?

Ok...that's three questions... :lol: :lol:
harpmaker
Dulcified! (>2000 posts)
 
Posts: 4687
Joined: Mon Feb 10, 2003 12:33 am
Location: Pershing, IN

Postby rtroughear » Mon Dec 22, 2008 5:17 am

Dave

They are top braces, with the remains of the top that they were glued to. They are also therefore glued to the underside of the fretboard where they cross. It's an arched fretboard in this case. I always put 4 or 5 top braces, and in the early ones I keyed them into the side linings, as is the case here. I don't do that any more.

When I get time I intend to study the effect of cutting them off. If I'm right, I expect the instrument to get louder when I do.

Richard T
rtroughear
Super Mbr (501-2000 posts)
 
Posts: 601
Joined: Tue Jun 05, 2007 8:18 am
Location: Brogo near Bega, NSW, Australia

Postby harpmaker » Mon Dec 22, 2008 9:15 am

Thanks, and I think your reply has answered some of my own questions about this experiment.

When I get a chance, I'll take some photos of how I build top sets and the total dulcimer bracing for comparison to yours and maybe (hopefully) build a topless one for side by side comparison.

I think part of the key to the mystery is in the top bracing you use....
harpmaker
Dulcified! (>2000 posts)
 
Posts: 4687
Joined: Mon Feb 10, 2003 12:33 am
Location: Pershing, IN

Postby rtroughear » Tue Dec 23, 2008 12:03 am

It's not a mystery Dave, there's an explanation for it - which might even be true, as KenH would say.

I would be very surprised if the nature of the braces had much to do with the conclusion that the top doesn't contribute much to the loudness of mountain dulcimers. That conclusion is also independently shared by one of Australia's leading steel-string guitar luthiers (who started out making MD's in the 1970's, bridge on soundboard), and also by a consulting mechanical engineer/outstanding guitar luthier that I've discussed this with. It would be clutching at straws to think the braces negated the general conclusion, and would give bracing an acoustic prominence that I for one haven't found to be the case (Paul C's bracing method is in a different class, like bridges on soundboards). I intend to test it anyway, but not today. It will take a couple of days to set up the test and make the measurements, and another day or so to do some anaylsis and draw conclusions. But now I have to make about 25 instruments (mostly ukes) for a folk festival in April, and I haven't started yet, so the chances are I won't get to do any dulcimer testing till after then. At which point I can also start on the sheaf of other dulcimer experiments I've written up that might help prove or disprove a general dulcimer sound production model that I'll present in due course.

I'd actively encourage you, and others, to repeat this and any other experiments, or devise your own. Only by multiple confirmations will any new findings become the new orthodoxy, and maybe help make better instruments.

Meanwhile, the general guidelines you present to new makers from time to time remain as valid, because they are based on actual experience. Any model that can't explain actual experience must either be rejected if it's too divergent from reality, or, more likely, be modified to accommodate those experiences.

I hope you all have a good Christmas, it's surprisingly cold here in Brogo.

Richard T
rtroughear
Super Mbr (501-2000 posts)
 
Posts: 601
Joined: Tue Jun 05, 2007 8:18 am
Location: Brogo near Bega, NSW, Australia

Postby harpmaker » Tue Dec 23, 2008 1:30 am

Actually, what I was refering to was the minimal differences you are experiencing, and my own experiences with various top set ups.

I do agree that the top can vary quite a bit, such as in thickness, without much change in the overall voice. It has been my experience that the top is just one factor in the voice, and some of those other factors will have a much bigger influence. IOW, I agree with an earlier statement in which you said (paraphrased) that the top is not as critical as some might think.

However, there is one huge difference in the way you build compared to the way I do it, and that is in the top bracing...or more accurately, a lack of bracing on mine.

I noticed in the pictures on your website that you use quite an extensive bracing on the tops. IIRC, in some cases there were 4 or 5 braces that ran from one side to the other, and they seemed, at least to me, rather large. Please understand, I am not being critical of this method, but I can't help but wonder how much affect that has on the overall flexibility of the fret board/ sound board assembly. Kind of like joists under a floor....the more points at which it is reinforced, the stiffer it will be.

On my instruments, the sound board is glued to the fret board and then the area under the fret board is hollowed out....and that is it. There are no cross braces that come in contact with the FB/SB assembly. At best, if the sound hole is especially intricate I will reinforce the area around it, but even then I do not use any cross bracing. The reinforcing is done by gluing a couple of layers of non-directional fabric over the underside of the hole and then it is trimmed back to the shape of the hole.

My goal is to keep the FB/SB assembly as light and as flexible as possible. And I think that our basic differences in building is why my experiences in building top assemblies has had different results. For instance, on my dulcimers, an overly large sound hole results in a brighter and appreciably quieter instrument.

One other difference is I tend to set the bridge a lot farther out onto the sound board than how it appears to be on yours. On mine it is at least 2" away from the end, and never set directly over the tailblock, unless the customer has specifically requested a brighter tone. Then it will be moved closer to the end...one of those design details which has a huge affect on the voice.

Don't get me wrong....I am not by any means rejecting your results, (OTH, I don't think I am clutching at straws either. ;) ) I do think that if I were to duplicate your experiment using my construction methods the results would be different.

I too am a bit back logged, but as soon as I get caught up I intend to make up a few experimental instruments to see what happens. I will post my results as they happen.

I do want to stress that I really appreciate all the work you have done on this. It has given me a lot to think about and some new ways of looking at things. After the first 1000 instruments or so, one can kind of get in a rut, so when someone comes along and shakes things up, it makes the work so much more interesting.

Looking forward to continuing this, and I hope it warms up for you down there.
harpmaker
Dulcified! (>2000 posts)
 
Posts: 4687
Joined: Mon Feb 10, 2003 12:33 am
Location: Pershing, IN

Postby paul c » Tue Dec 23, 2008 8:11 am

Hi Richard--

I much appreciate your further brainstorming and am really intrigued by some of the possibilities you've mentioned. And I really appreciate your effort and willingness to put concrete measurements to things. Good luck with your building and I'll look forward to the time when you get a chance to return to the testing.

I suspect that you're entirely correct when you suggest that the effects I'm observing are due to other mechanisms. (The role of the top as coupling mechanism (the physical transfer of vibration) hasn't really been mentioned.)

And when we approach the nuances of sound and what is a "more pleasing" sound, we enter a really murky world of subjective evaluation with questions of whether or not mechanical devices are capable of measuring what the human ear can hear and how the brain processes that.

It will be interesting to see what happens as your sample size increases--and as you research expands to include instruments (and perhaps those other than your own) that are considered pleasing and of high quality.

I have been brainstorming about possible "tests" that would relate to top function--things aren't in order yet, but I do hope to post something in the near future.

Thanks again,

paul c
paul c
Senior Mbr (101-500 posts)
 
Posts: 264
Joined: Tue Aug 22, 2006 12:45 pm
Location: Holmes County Oh

Postby rtroughear » Tue Dec 23, 2008 8:50 am

G'day all

Prior to cutting off the braces of this instrument I wanted to redo the sound analysis to make sure nothing has changed, so I quickly did that today. The average sound level output is unchanged from the first recording (less than 0.5 dB difference), even though today is more humid. So that's good.

For those who expressed an interest in seeing and hearing the same thing without the top, I also recorded the longer tune again today.

The good quality audio file is at:

http://www.4shared.com/file/77242136/4f ... 23Dec.html

and the Youtube video, with the same, but poorer audio is at:

http://au.youtube.com/watch?v=_oBZBSSHwrQ


Dave: The bridge on this dulcimer is 1.5" from the end; the edge of the endblock is about 1/2" back beneath it, it's cut back on the top, but not the bottom. These days all my dulcimers are longer and have the bridge about 3.5" to 5.5" from the end, depending on the scale length.

Sometimes the bracing I have put on the tops has been stiff, and sometimes light, in an attempt to match the cross-grain and long-grain stiffness. I've tried the method that Dwain Wilder uses, and the method that Terry Hennessy uses. But I don't think any variation I've tried has had a clear effect, so my conclusion is that the top bracing is mainly for strength, and doesn't do much one way or the other to the sound (I'm only talking about cross braces here). In the guitar world the top bracing contributes to making the cross-grain stiffness of the plate more similar to the long-grain stiffness. That also used to be my aim with dulcimer tops, but I no longer think it's a valid analogy. So I think no braces is good too, as far as sound goes, just not as strong. The sound holes, along with the box volume, set the two lowest air resonances (in 4-hole dulcimers anyway), which I do think are quite important to the low end warmth of the sound, such as it is in mountain dulcimers.

Paul C: I won't be even attempting to discover what makes a good dulcimer or not - I doubt I'd manage that in my lifetime - it's far too subjective. I'll be happy if I can find out, in the crudest way, how mountain dulcimers make sound. Right now I don't know. That might then be a starting point for making them sound good. There'll be a sample of one - one test, one instrument, or jig, to demonstrate some specific aim. I hope the conclusions should be broad enough to generalise to other dulcimers, but confirming tests by others would be a good thing.

Richard T
rtroughear
Super Mbr (501-2000 posts)
 
Posts: 601
Joined: Tue Jun 05, 2007 8:18 am
Location: Brogo near Bega, NSW, Australia

PreviousNext

Return to Making Dulcimers

Who is online

Users browsing this forum: No registered users and 4 guests