Measuring the acoustic behaviour of the closed soundbox- a better method to predict the sound?

[David Burgess]

19 minutes ago, Salve Håkedal said:

There is 3 stages in the graph: first the unvarnished fiddle body alone, with the soundpost in place, then the varnished violin with neck and pegs and, of course, sound post in place, then the fully set up instrument.

Interesting!
Regarding the instrument being fully set up:
Are the strings damped, or are they free to vibrate?
Thanks.

[Salve Håkedal]

Yes, damped:

And the fiddle is held at the neck by the saddle, approximately. The recorded sound file is 0.2 second long.

[Davide Sora]

7 hours ago, LCF said:

I think that the best place to hide secrets is in plain sight. Nobody suspects that they're real or true then. 

In the areas of instrument making where I have good skills I have freely taught what I know but with human nature being what it is most people are driven to reinvent things for themselves. I think nobody learns much except by reinventing and internalising the things you teach them so it becomes their own experience. 

Tell people all you know, spill all the beans. They'll never be able to keep up!

I can't help but agree with you.

[Andreas Preuss]

8 hours ago, Salve Håkedal said:

There is 3 stages in the graph: first the unvarnished fiddle body alone, with the soundpost in place, then the varnished violin with neck and pegs and, of course, sound post in place, then the fully set up instrument.

(If you are interested in more stages, like fluting in only the back or top of the body without neck but sound post inserted, I have it. Or the change after the last stage: adding the sympathetic strings, I have that too. But as I state above: I can not really say that I have had much practical use for it.)

Thanks for the clarification.

One thing which I find noteworthy in your sound spectra comparison is that high frequencies in the overtone range  come out only after stringing the instrument up.

My rough guess about the other stages is that there are not so recognizable and consistent shifts in the entire spectrum. I’d rather be interested in a second spectrum which is taken half year after set up. (If you have something like this)

[Salve Håkedal]

10 hours ago, Andreas Preuss said:

Thanks for the clarification.

One thing which I find noteworthy in your sound spectra comparison is that high frequencies in the overtone range  come out only after stringing the instrument up.

The knock to the instrument is in different places with no strings and then after set up. It's no longer possible to knock against the middle of the bass bar with the bridge and strings in place. Knocking to the bridge activates the high frequencies a lot.

10 hours ago, Andreas Preuss said:

I’d rather be interested in a second spectrum which is taken half year after set up. (If you have something like this)

I don't think I have that for an instrument that was not altered in between. But here is one that had an accident a few months after delivery and came back for repair (gluing a crack from end to end of the table).  Though it probably shows my inconsistency in measuring method better than it shows the actual changes to the instrument, you can all have it:

 

[Salve Håkedal]

Relating to the original posting, I recorded a knock against the left bridge foot position first after carving the channel and then after finishing the fluting (concave part of arching) on a closed sound box with the sound post in place.

As I don't control the impact of my little hammer very well, the amplitudes can not be relied upon. I don't control the hammered spot very precisely either, but the lower resonances should not be affected by this, I think.

Removing the wood from the concave part of the arching between the channel and  the rest of the arching is actually a rather drastic thing when talking about thicknessing, isn't it? Grams is removed from both back and top. But most resonances sinks just a little bit.

My tactile impression is that it takes less impact from the hammer to excite the body. Which is of course only logical and to be expected. Sooo.. : what did I get from this hammering activity? Answer: a break from more productive work :-D

(The instrument is a Hardanger d'amore, which in this context can be considered a very small viola (38.5cm))

[Don Noon]

9 minutes ago, Salve Håkedal said:

Sooo.. : what did I get from this hammering activity? Answer: a break from more productive work :-D

That's the normal result.  Readings at this stage of assembly are so removed from the final result, so any "correction" you might think is a good idea will have an unknown actual effect in the end... unless you build up a huge database of cause/effect of these corrections.  Then you'll KNOW how pointless it all is .

[Davide Sora]

3 hours ago, Salve Håkedal said:

 Sooo.. : what did I get from this hammering activity? Answer: a break from more productive work :-D

Well, but it's fun and after the break you will return to work more concentrated and relaxed, so in a certain sense it will improve the final result.

[Greg Sigworth]

"The Secret Violin Mystery" was a PBS documentary in the 80's by Fry (I think) and I was hooked and then morphed into Hutchens' ideas about plate eigen mode frequencies. I communicated with Peter Tourin who was either in Vermont or New Hampshire about the Fultin Varnish I was making and told him about the TV show I saw and its so called secret. His reply was this," There is no secret violin mystery, it's just hard to make a violin." Thanks to all those who do this hard work and are willing to share their insights, experiences and work. It is my opportunity to fulfill the need to learn from others. Can't wait to finish this violin and make another.

[Andreas Preuss]

5 hours ago, Don Noon said:

Readings at this stage of assembly are so removed from the final result

If you say so….

…. but on the other hand isn’t it a bit strange? We know (or should know) the basic parameters to finish the instrument, combined string force neck weight, string angle etc etc. What I am trying to say is that we don’t put another complicated ‘mechanism’ on the box. 
 

So what makes in the end everything so unpredictable? Simply the fact that it is somehow too complicated to put the shape of a violin into mathematical formulas with the additional obstacle that wood is not a homogene material like steel or plastic? 

—————

I see on the bottom line only the practical solution not to think about how to predict the final result, but rather figuring out procedures to trigger the tonal result towards a self-set goal. 
 

 

[Andreas Preuss]

6 hours ago, Salve Håkedal said:

Relating to the original posting, I recorded a knock against the left bridge foot position first after carving the channel and then after finishing the fluting (concave part of arching) on a closed sound box with the sound post in place.

As I don't control the impact of my little hammer very well, the amplitudes can not be relied upon. I don't control the hammered spot very precisely either, but the lower resonances should not be affected by this, I think.

Removing the wood from the concave part of the arching between the channel and  the rest of the arching is actually a rather drastic thing when talking about thicknessing, isn't it? Grams is removed from both back and top. But most resonances sinks just a little bit.

My tactile impression is that it takes less impact from the hammer to excite the body. Which is of course only logical and to be expected. Sooo.. : what did I get from this hammering activity? Answer: a break from more productive work :-D

(The instrument is a Hardanger d'amore, which in this context can be considered a very small viola (38.5cm))

Thanks for posting it this. 
 

Hard to say what changed. I don’t think that those minor frequency shifts for the peaks below 1khz do much. if anything there are rather changes at the upper end of the spectrum. 

 Probably this shows exactly what happens when you do the same experiment on a strung up instrument: not too much for the sound or sound colour itself but rather something how it feels under the bow. 

[Don Noon]

1 hour ago, Andreas Preuss said:

So what makes in the end everything so unpredictable? Simply the fact that it is somehow too complicated to put the shape of a violin into mathematical formulas with the additional obstacle that wood is not a homogene material like steel or plastic? 

I would not say that "everything" is unpredictable, at least for my own work.  I can predict that my signature modes will end up in a perfectly normal, useful place without modifications.  Maybe not within 10 Hz (except A0), but close enough for any reasonable player.  A normal violin construction just comes out that way (after a bit of experience).

The higher modes are a lot like a herd of cats; they do what they do, and trying to coax them into a desired result -even after full assembly where you know exactly what you have- results in frustration.  I've tried.  Bridge and soundpost adjustments are something else, but useful to some extent.

At this point, I don't really make modifications any more, unless I'm doing an extreme experiment.  It goes together, and it is what it is.  I will experiment to some degree with arching shapes, and try to infer from the result if the direction is what I want.  If it's not good enough, I'll make another plate.  Or make it thinner and keep it for one of my personal fiddles.

What do other pro makers do?

 

 

 

[Andreas Preuss]

9 hours ago, Don Noon said:

What do other pro makers do?

Well, it seems that each successful maker has his/her own approach and philosophy. 
 

There are acoustic tinkerers (meant in a positive sense)

Acoustic lab approach people

Historic informed rebuilding approach with or without antiquing.

Or based on a simple strategic idea like one acclaimed maker building classic models who said ‘I want to be known as the master of asymmetry’ 

There are also makers who complete the instrument with a top plate too thick and only after the varnish has settled they tune it to their acoustic ideas.

But that’s presumably not new to you. 
——————

What I often think about acoustic research and the entailing experiments with listeners is that we are getting spiraled into a logic where only causality counts. So if we simply assume that this particular causality can’t explain everything (what is important to musician), we know also that there is a problem. A sort of insufficient explanation of phenomena.
 

9 hours ago, Don Noon said:

The higher modes are a lot like a herd of cats; they do what they do, and trying to coax them into a desired result -even after full assembly where you know exactly what you have- results in frustration.  I've tried.

I found it also extremely tricky to influence the high frequencies on traditional built instruments. And somehow this looks to me like the ‘real problem’ or somehow more important than signature modes. In fact I think that ‘signature’ is in this context the signature of the classic construction concept. So related to a particular concept those modes must fall in certain frequency ranges with certain amplitudes in order to function. 

I started to adopt for the ‘herd of cats’ a ‘blurred vision approach’. We don’t need to see details but rather the total. this became for me the reason rather to look on low resolution spectra for high frequencies (over 2khz). 
 
For constructional methods to trigger this particular spectrum band it’s much more difficult. I believe that Cremonese makers simply figured out methods of material selection and arching shapes to get the ‘female soprano voice’ formant. This also partly explains why we struggle so much to imitate it especially if we start with the wrong material. And if we work with different material for each violin, well, it would be almost a miracle to hit the target each time. 
 

 

 

 

[David Burgess]

14 minutes ago, Andreas Preuss said:

I found it also extremely tricky to influence the high frequencies on traditional built instruments. And somehow this looks to me like the ‘real problem’ or somehow more important than signature modes.

Somewhat agreed. I think the early emphasis on the lower modes came about because they were the easiest to measure and control.

[Don Noon]

18 minutes ago, David Burgess said:

Somewhat agreed. I think the early emphasis on the lower modes came about because they were the easiest to measure and control.

and the higher modes were more numerous, harder to measure, and essentially impossible to analyze and control... even though they are extremely important.

[Anders Buen]

When we work with the wood, like scraping or using other tools, we hear the response of the body and plates. So unconciously or consciously we monitor the sound. Measuring the steps in narrow band FFT may give some detailed clues or none.

The higher frequencies are more controlled with bridge properties and what it rest on, maybe wooden properties and effects that are rather difficult to asess in detail, like the radiation efficiency maximum. We may hear it as we work on both the plates or the bodies.

I have recently gotten access to Graham Caldersmiths data collection of free plate frequencies, weights, arch heights and assebled violin signature mode data. I got it from a genrous research and academic in Mexico who have written about some of the materail.  It is about 60 violins and some violas. And I can reveal that there are signifcant correlations between free plate resonance frequencies and the higher signature modes. Archeight also is significant, although the dataset is not full. I will retur to this with correlation matrixes and plots. 

[Andreas Preuss]

5 hours ago, Don Noon said:

and the higher modes were more numerous, harder to measure, and essentially impossible to analyze and control... even though they are extremely important.

My general impression is that if musicians talk about certain (outstanding) qualities of instruments they talk about high frequencies in one or another way. Especially when it comes to modulation of sound colors it must have something to do how the player can influence the ‘overtone behaviour’ of the instrument. For this, the body must be sensitive enough to react to alterations of the bow weight, bow speed and the contact point on the string. Otherwise what some players describe as ‘resistance’ seems to be the ability of the instrument to swallow any energy input without collapsing. I be a as m guessing that the sound collapse goes from top (high frequencies) to bottom because often the fundamental frequency 
 

The fact that high frequencies look in a spectrum like the jungle shouldn’t discourage us to figure out methods to cope with them in a creative way. There were other things in the history of science where the right person finally figured out how to solve the ‘impossible’. 

[Andreas Preuss]

29 minutes ago, Anders Buen said:

And I can reveal that there are signifcant correlations between free plate resonance frequencies and the higher signature modes

Sorry to ask. Which are the higher signature modes?

[Anders Buen]

2 hours ago, Andreas Preuss said:

Sorry to ask. Which are the higher signature modes?

Jansson description: T1, C3, C4

Bissinger: B1-, B1+, "ring mode in back plate" (Last one is not Bissinger)

[Don Noon]

I'm not sure of the names, but there are two strongly radiating modes that I consistently see:  one around 850 Hz and one around 950 Hz.  The 850 Hz mode is strongly affected by how close the soundpost is to the bridge foot; if it is too strong (as determined by playing), I move the post closer.

On the Strad3D Titian animation, these modes are at 839 and 990 Hz if you want to look at the mode shapes.