This June, Sotheby’s will be auctioning off the Macdonald viola, created in 1719 by the legendary Italian violin-maker Antonio Stradivari. They’re asking for a minimum bid of $45 million, which would make this viola the most expensive instrument ever sold.

This price tag may break records but Stradivari’s instruments often fetch millions of dollars, due to both the cachet of the name and their reputed quality. Many people genuinely believe that they are superior to newly made violins and many scientists have tried to work out why.

But to Claudia Fritz from Sorbonne University, the search for Stradivari’s secrets is a “perennially fruitless one”… because they don’t exist. In two studies, she has shown that professional violinists can’t tell the difference between the so-called “Old Italian” violins and newly made ones.

“Strads are amazing instruments. They have survived 300 years and are beautifully made,” says Fritz. “I don’t want to destroy the Strads but I want to show that their amazing properties aren’t unique. You can find them in new violins as well. The new makers are doing a great job and are making amazing violins. They should be able to sell them with pride and recognition.”

Here’s a 28-minute documentary about the experiment.

In her first study, Fritz recruited 21 professional violinists from the International Violin Competition of Indianapolis, ushered them into a dimly lit hotel room, and asked them to play six violins. Three were new; one had been made just days before. The others had been crafted by either Stradivari or Guaneri “del Gesu” in the 17th and 18th centuries.

The violinists couldn’t tell. When they tested pairs of the violins, they were just as likely to prefer the new or old ones. When they played all six instruments together, and had to choose which they would most like to take home, 62 percent picked a new violin. And the oldest Stradivarius—an instrument that is held in an institution and loaned to only the most gifted players—was the most frequently rejected one.

“Many people were convinced that as soon as you play an old violin, you can feel that it’s old, it has been played a lot, and it has a special sound quality,” Fritz told me at the time. “People who took part in the experiment said it was the experience of a lifetime when we told them the results.”

John Soloninka was one of them, and he talked about his experiences in a comment on my write-up. “I expected to be able to tell the difference, but could not,” he wrote. “Claudia sent me my comments about the instruments that I made while I was playing them, and it was hilarious how wrong my impressions were at the time!”

The study was published in 2012, and Fritz says that the reactions ranged from delight to anger. Critics were quick to point out the experiment’s limitations—see the comments here for a sampling. The violinists only tested six instruments, and they played them for just 20 minutes in a dry hotel room. That wouldn’t do. To get the most out of the violins, the players needed hours—maybe weeks—of testing, and they needed to play in a concert hall. One distinguished violinist reportedly said, “You don’t test a Ferrari in a parking lot.”

Fair enough, thought Fritz. Let’s go to a concert hall.

“We couldn’t address all the issues in one study anyway,” she says. “We needed the first one to attract attention, so we could do a better one. This time people were really happy to loan me some instruments.”

This time, she worked with six Old Italian violins and six new ones. She recruited players of the highest calibre—10 soloists who were either internationally renowned or had won major international competitions. (Seven of them play old violins themselves.) And she assembled a seven-person team that included several scientists, a violin-maker, a soloist, an instrument dealer, and a string engineer. (Three of them own and play an Old Italian.)

The team laid all 12 violins out on a table and told the soloists to pick one that could hypothetically replace their own instrument for an upcoming tour. They had 50 minutes to play the violins as they wish, using their own bows, and they picked their top four. (The team polled several soloists before the experiment about the time they’d need to comfortably evaluate a dozen violins—the average estimate was 50 minutes.)

They ran this experiment in two Parisian venues—the home of a family of professional string players, and a 300-seat concert hall that’s well-known for its acoustics. In the hall, the soloists could ask for a piano accompaniment or for feedback from a listener of their choice. They could even ask one of the team to play the violins so they could check the sound from elsewhere in the hall.

The team gave the violins four points every time they were chosen as a top pick, three points for second place, and so on; they deducted a point every time an instrument was rejected.

Violin rankings. The six new instruments are on the top; the six old ones are on the bottom. The left bars represent the scores in the house; hte right bars are the scores in the concert hall.

Violin rankings. The six new instruments are on the top; the six old ones are on the bottom. The left bars represent the scores in the house; hte right bars are the scores in the concert hall.

Although the soloists varied in their tastes, two new violins consistently scored the most points, with an old Stradivarius tailing in third place. Overall, the new violins collectively scored 35 points and the old ones scored 4—a six-fold difference.  Fritz writes, “We can find no plausible scoring system by which the old fare any better.”

The concert hall made little difference either. Some violins rose in the ranks and others fell, but the new ones maintained their lead over the old. “There is certainly no evidence here to support the belief that Old Italian violins come into their own in concert halls whereas new ones fall behind,” wrote Fritz.

In the final 12 minutes of the experiment, the team provided the soloists with three violins—their own, their top pick, and the best-rated instrument from the opposite age category. Their job was to rate each instrument according to six qualities: overall quality, articulation, timbre, playability, projection, and loudness. They rated the old and new violins similarly in terms of overall quality, but gave the new ones higher scores in all the other categories. The concert hall closed the gap slightly, but not enough to turn the tide in favour of the old.

Violin_rankingsFinally, the team asked the soloists to guess whether they were playing an old or new instrument. They were wrong half the time.

The results are very clear: Stradivarius violins, despite their reputation, inordinate price tags, and indisputable craftsmanship, are no better than the best modern ones.

Critics will undoubtedly cling to their beliefs, and reiterate that it takes ages to master an old instrument. But let us reiterate: ten violinists—among the best in the world—couldn’t tell the difference between the old and new violins, and were more likely to prefer the new violins when allowed to play the instruments in a concert hall. When they were debriefed later, they said that the experiment was more than realistic enough for choosing a violin for a tour (although not for buying one).

Fritz has also done two more studies: one assessing the listeners’ experience as the violins were played in an concert hall, and another analysing the physical properties of the instruments. Both sets of results will be published soon but—spoiler alert!—neither will support the superiority of the Old Italians.

“We should stop mentioning on the programme what soloists play because who cares?” she says. “That would allow young soloists to make a career without struggling to have a Strad on loan. If we judge players on how they play rather than what they play, that would be better.”

UPDATE: I asked John Soloninka for his thoughts on the new experiment and he kindly sent a very detailed response. Over to him: 

I was very privileged to be one of the original participants in the so-called “Indianapolis Experiment”. Here is a detailed account of my experience; I won’t repeat those details here.

During the summer of 2013, I was one of a privileged group of violin makers and acousticians who were invited to the Oberlin Violin Acoustics workshop, in Oberlin Ohio. Claudia Fritz, Joseph Curtin, and Fan Tao shared the results of the new experiments to get confidential feedback from this group.

At one point during the week, I had the great privilege to play a Strad and a Guaneri for about an hour.    These instruments together are worth about $15M at today’s prices. I was so impressed with the sound that on seeing Claudia later in the afternoon I said, “I totally accept the double blind study… but when I pick up that Strad, it is so amazing!  What is it about the playing experience when you “know” what the instrument is, that makes it sound so great?”

She said, “You think you could tell that instrument from amongst 4 new and old other ones?”

My response: “Of course…but that is not fair…I just played it for an hour, I would be able to pick it out.”

She proceeded to arrange another test. We did the same procedure later that day and, again, under blinded conditions, I felt the “magic” but not just on the old ones but on the new ones as well. In fact, I preferred a new one!

Paraphrasing Sam Zygmuntowicz , one of the world’s leading modern luthiers, if you know an instrument is from the Golden Period, you will assume any poor sounds are attributed to your playing, rather than quirks of the instrument, and you would apply all subtlety in trying to coax nuances of sound of the instrument. If is it new instrument, you may attribute any poor sound to the instrument, and be less willing to search for the subtleties.

The new experiments will again bring up the criticisms of how this was not a perfect test (for example, optimizing bow, set up, time with the instrument etc.) and these are all true.  Laurie Niles of violinist.com (who also participated in the Indianapolis Experiment) has posted the first such analysis here.

But I chose to draw different conclusions.   These factors mentioned cannot all be biased in favour of new instruments and against old.   The bow, for example would be suboptimal in a random way…not just benefiting the new ones.

What I conclude is the following:

- Great old instruments and great new instruments are great!

- The McDonald Strad Viola on sale for $45M is a great instrument.  But NO ONE used its tone or playability when determining the price.  It is the rarity, provenance, physical state of preservation of the antique, and most importantly, what the market will bear, that led to the price.

- There is no “secret” that makes old instruments magically better, and their sound and playing qualities are certainly reproducible to the extent that soloists or audiences can notice.

- That confirmation bias in decision making is very powerful… and you can fool yourself into believing almost anything, and seeing or hearing what you want to hear.   So if it is violin, or wine, or venture capital investing or anyone one of many complex, preference-based, subjective decisions we make, we should ensure we are being objective and control for our biases.

- I think Laurie Niles was right when she said:  The Golden Period was a peak in of violin making, and we are in a second one right now… and that is a great thing for professional musicians!

 

Reference: Fritz, Curtin, Poitevineau, Borsarello, Wollman, Tao & Ghasarossian. 2014. Soloist evaluations of six Old Italian and six new violins. PNAS http://dx.doi.org/10.1073/pnas.1323367111