Listening tests of experimental carbon composite violins

[Anders Buen]

Have anyone seen and discussed this new JASA article? https://asa.scitation.org/doi/10.1121/10.0001159

Abstract:

For centuries, wood, and more specifically spruce, has been the material of choice for violin top plates. Lately, carbon fiber instruments have entered the market. Some studies show that composite materials have potential advantages for making instruments [Damodaran, Lessard, and Babu, Acoust. Aust. 43, 117–122 (2015)]. However, no studies exist that evaluate violins made of different composite materials as judged by listeners. For this study, six prototype violins, differing only by the material of the top plate, were manufactured in a controlled laboratory setting. The six prototype violins were judged by experienced listeners in two double-blind experiments. In contrast to popular opinion that violins made from carbon have or lack a specific sound quality, the study provides insights in the diverse sounds and timbres violins from fiber-reinforced polymers can create. It allows an investigation of the links between the perception and the variations in material properties of the soundboards. Additionally, as neither players nor listeners are acquainted with these instruments, these results provide an interesting view on what type of qualities of violin-like sounds are preferred by listeners.

[David Burgess]

1 hour ago, Anders Buen said:

Have anyone seen and discussed this new JASA article? https://asa.scitation.org/doi/10.1121/10.0001159

 

Thanks, Anders. This looks to have been a very well-run test.

[Jul]

Very entertaining article. The experience seems convincing as a test bench for violin top plate material.

however, a violin is not only a sound maker. It can also be an exquisite object made by a human being. 

[Pate Bliss]

there was a comparison of violins that were actually 3d printed.  mostly they sounded bad, but one in particular i wouldn't have minded having!  it makes me think there's nothing magical about material, and it shows a lot more goes into it than just material, just as you face with wood

[Fossil Ledges]

I love these materials experiements.

[RyanJ]

https://www.alibaba.com/product-detail/Hot-Sale-Carbon-Fiber-Violin_60739706620.html?spm=a2700.galleryofferlist.0.0.332e6a01KAjdVs

Ever see a man tap dance to a carbon fiber violin?

[matesic]

Oh boy. I used to regularly get asked to review papers like this submitted to various journals, so I tried to brush off my cobwebs and approach this one in the same way as I would back then.

On the plus side it's a sincere and transparent attempt to objectify what we all know to be a notoriously subjective business. They tested 5 prototype cornerless instruments built mostly in resin and as near as possible identical in all but the material of the front, and recorded the subjective evaluations of 37 experienced listeners on 8-point scales described as: warm v. cold, clear v. dull, loud v. quiet, soft v. harsh, open v. closed, good v. bad, nasal v. clear, round v. sharp, powerful v. weak, rich v. poor, bright v. dim. This is exactly the kind of thing I was advocating on this site a couple of weeks ago! Their double-blind design seems to control pretty effectively for subconscious bias on the part of the listeners and the player.

Unfortunately then the variables start to proliferate. The thickness of the front is deliberately varied from instrument to instrument in order to try to equate the longitudinal bending stiffness, but this of course entails significant variations in weight and a huge variation in transverse/radial stiffness that I don't understand and the authors don't comment on. A second violin gets built in one of the materials in order to check the consistency in the listeners' evaluations for supposedly identical instruments (poor) and one of these two also gets evaluated twice to gauge the effect of order of listening (large).

The most encouraging result, I think, is that there was some measure of consistency between listeners for all of the violins on most of the evaluation scales; the people tended to agree that almost all sounded more “warm” than “cold” while three sounded “loud” and one “soft”. “Clear”, “bright”, “harsh”, “nasal” and “round” were consistent judgements in one or two violins each. Unfortunately the TwillCA violin that got evaluated twice was judged significantly “dim” first time but “loud” “sharp” and “powerful” second time! Although the UDflax violin scores were the most tilted and the Sandwich violin the least tilted in favour of “warm”, that doesn't actually prove that one violin sounded warmer than another. They really should have done some kind of ANOVA to see if the warmness scores really were different across violins, followed by paired t-tests to compare specific instruments. The violin that clearly stands out as most often judged to be “loudest” was the Sandwich, but every single violin was selected by at least one listener as either “richest” or “warmest” (and probably both, although they didn't look to see how closely the various scales correlated with one another).

Then unfortunately they go into “preference” on whatever grounds each listener feels like prioritizing. I don't think that tells us anything useful. The overall conclusions are cautious enough but it's hard to accept that “Composite materials allow the creation of violins (that) offer possibilities to change the sound to the criteria of the player”.

 

[Don Noon]

Drat... it looks like the article was available for free until today, and now it's $30.  I skimmed thru it yesterday, and wanted to read more carefully today.  Bummer.

From what I remember:

One of the main problems I have with this is that all of the violins are basically extremely abnormal to start with, which can skew any preferences.  A cornerless, composite molded body with abnormal F-holes seems a strange place to start.  A second (minor) complaint is that the impact spectra are too low resolution to distinguish the B modes... if there are even separate B- and B+ modes with this abnormal body.

As for interesting results:

1) Not a big surprise, but at least the heaviest, low stiffness/weight top generated the least sound, and the lightest, highest stiffness/weight top generated the most.

2) I think the most interesting result was that the heavy, low-dB top was the most preferred tonally by the LISTENERS... but the PLAYER hated it.  And the light, high-dB top was hated by the LISTENERS, but not so much by the player.

In the response spectra, the listener most preferred top had the least output in the "transition hill", and the least preferred top had the highest output in the "transition hill".  That goes along with the "Old Italian Sound" concept, where the most striking feature (to me) is the relatively low output in the midrange.

As far as the difference of opinion between listener and player, the listeners just hear the sound, but the player is very much aware of how much work he has to put in for what sound comes out.  My takeaway  is to continue trying to make the most powerful instruments that have a preferred tone, but I don't see much in the article that is immediately useful for reaching that goal.

 

 

[Fossil Ledges]

This reminds me of the early 3D printed stuff at Oberlin, the carbon fiber is very interesting since the properties are so different than wood. Somewhat dark and menacing, like Darth Vader's helmet, or my ex-wife's heart!

[Fossil Ledges]

Someone mentioned copying on a molecular level and it makes one wonder how far away we are from being able to scan an instrument and do that. It seems a lot depends on the printer technology, but NASA can print pizzas and medical scientists are working on printing animal organs and such, why not violins? The big carpenter bumblebees have been doing it in reverse for years as they slowly disassemble my horse barn and make it into cellulose!

[matesic]

1 hour ago, Don Noon said:

Drat... it looks like the article was available for free until today, and now it's $30.  I skimmed thru it yesterday, and wanted to read more carefully today.  Bummer.

From what I remember:

One of the main problems I have with this is that all of the violins are basically extremely abnormal to start with, which can skew any preferences.  A cornerless, composite molded body with abnormal F-holes seems a strange place to start.  A second (minor) complaint is that the impact spectra are too low resolution to distinguish the B modes... if there are even separate B- and B+ modes with this abnormal body.

As for interesting results:

1) Not a big surprise, but at least the heaviest, low stiffness/weight top generated the least sound, and the lightest, highest stiffness/weight top generated the most.

2) I think the most interesting result was that the heavy, low-dB top was the most preferred tonally by the LISTENERS... but the PLAYER hated it.  And the light, high-dB top was hated by the LISTENERS, but not so much by the player.

In the response spectra, the listener most preferred top had the least output in the "transition hill", and the least preferred top had the highest output in the "transition hill".  That goes along with the "Old Italian Sound" concept, where the most striking feature (to me) is the relatively low output in the midrange.

As far as the difference of opinion between listener and player, the listeners just hear the sound, but the player is very much aware of how much work he has to put in for what sound comes out.  My takeaway  is to continue trying to make the most powerful instruments that have a preferred tone, but I don't see much in the article that is immediately useful for reaching that goal.

 

 

I made a copy of the text. I guess it would be illegal to paste it here but anyone interested could PM me

 

 

[jezzupe]

I gotta be honest about my opinions, this seems like a really large expenditure of money to find out something that doesn't really matter and that .01%of the population might care about.

Why would anyone be surprised that in a close minded non experimental {comparative speaking to say guitars} field of making musical instruments , that when someone decides to make a couple "experiments" that they sound ok. It's all most like they funded an expedition to the top of the hill only to be surprised that there was a top of the hill at the top of the hill. 

One thing I will add about these types of materials. They refer to the fabric , such as flax. IMO what we are hearing with any of these composites is the epoxy shell,be it carbon fiber,denim,flax or what ever,to me that is just a medium to absorb the epoxy , which is what is determining the over all thickness, and is what the mass of the body is, so I think the fabric medium has very little to do with any of this.

I do think it's strange how liberal they were with their "Goldastini" designs and also strange that they would even compare these instruments to "regular" violins which are not only a different material, but a different design. Like comparing apples to watermelons.

I guess we've learned what I've always known. That is if you put strings an an apple and a watermelon that you'd be surprised that they don't sound that different from each other.

[Marty Kasprzyk]

I had a chance to hear these composite violins being played at the 2019 Oberlin Acoustics Workshop and I was very impressed with them and their maker Tim Duerinck.  He gave a talk regarding the building of them and if you google his name you will find many references about his work.

Regarding the JASA paper I would have preferred a different testing of the various instruments-many players giving their opinions rather than lots of listeners because I think the player's opinions are more important.  I would expect quite a bit of scatter in their preferences which is what happens with conventional wood instruments too.

The sandwich type of top (~43g) gave the best projection because of its light weight but its sound wasn't well liked.  It was made with  a carved polyurethane foam covered with carbon  fiber cloth coverings on both the upper and lower surfaces so that the resultant composite had a longitudinal stiffness/cross stiffness ratio of 1.0 so it was quite different from wood or the other composites that used unidirectional carbon fiber.  So it is not surprising that it sounded brighter and not as warm as some of the others that were better liked.

I would recommend that other sandwich plates be made with the same foam core but with one or both surfaces covered with unidirectional carbon fiber in the longitudinal direction to better duplicate wood.

Besides the light weight advantage the use of a foam core might give more damping than a solid plate of epoxy&graphite fiber because it could give constrained layer viscous damping effect.  The damping amount can be varied by using different foam densities and/or thicknesses and an optimum amount of damping might be discovered.

Claudia Fritz made a comment that there is no agreement of choices between the instruments but that "rich" and "powerful" are wanted.  She also said there is no agreement that there was a trade-off between quality and power.