Stradivarius vs. Guarneri—Test your audio system, test your critical listening skills

[ARQuint]

This thread is prompted by Part 2 of bluesman's series of feature articles titled

Realism vs Accuracy for Audiophiles. Be sure to read it carefully, as well as part 1. Great stuff.

 

At the beginning of his piece, the author listed ten "facts and assumptions" that form the basis for his treatise. He included these three:

 

·  It doesn’t matter if you know that you’re hearing a Selmer Mk 6 or a Yanagisawa tenor saxophone.  What matters is that you understand how a player’s choice of instrument can affect both  his / her playing and what you hear. The combination of player and instrument shapes the unique, recognizable tone you identify with a player or group.  You don’t need to be able to identify Stan Getz or Sonny Rollins or Ben Webster by name – but if these three sound the same (or even similar) to you on decent recordings, you’re missing a major part of the experience.

 

·  With guided experience over time, you should come to know and recognize a reasonable approximation of the real sound of your favorite artists and music.

 

·  With this knowledge, you will be better able to judge the accuracy of your equipment and the quality of the recordings you play.

 

 

In support of bluesman's premises, I'm presenting an opportunity to consider in detail the sonic differences among word-class musical instruments as played by a world-class musician. I'll repeat the background described previously in comments to bluesman's article.

 

In December of 1994, the Chicago rare instrument dealer Bein & Fushi brought together 30 of finest violins in existence, 15 made by the Cremonese violin-maker Antonio Stradivari and 15 by Giuseppe Guarneri del Gesù. They were transported under tight security to the Performing Arts Center at Purchase College, outside of New York City, which has a recital hall with excellent acoustics. For two days, they were photographed and, more importantly for our purposes, played by one violinist, the American virtuoso Elmar Oliveira. He surveyed a range of repertoire but recorded the opening minute of the Sibelius concerto, unaccompanied, on all 30 instruments. He was recorded by Mark Levinson on custom equipment, and the resulting coffee table book (The Miracle Makers) comes with three compact discs worth of music, the third of which is entirely devoted to the Sibelius excerpt.

 

To potentially test your ears and your stereo system's capacity to differentiate among some of those instruments, I've provided three zip folders (below) containing music files. The first [Strad vs Guarneri 2_Unblinded] holds six examples of the Sibelius excerpt, recordings of three Strads and three Guarneris, identified as such. None of these instruments are included among the "unknowns" contained in the other two folders. Listen to them as much as want to get a fix on what differentiates a Stradivarius from a Guarneri. You are welcome to read up as well, and here are a few distinctions between the two studio's violins that Geoffrey Fushi noted in The Miracle Makers.

 

                       "Guarneris tend to have a brilliance that is edgier or more biting

                        in the upper register and more expansive and alt-basso in the lower

                        register. Stradivaris are said to have a bright, clear and bell-like high

                        range and a compact, concentrated low range, still with edge or bite."

 

Maybe that's helpful, maybe it's not.

 

The other two folders contain recordings of 12 different instruments. Not necessarily equal numbers of Strads and Guarneris, but I promise they are not all by the same maker. One folder [FLAC Dis-identified] has twelve 16-bit/44.1 kHz rips from the CDs that came with The Music Makers. The other [Miracle Makers_MP3_dis-identified] has twelve files that have been down-sampled to 145 kbps mp3 quality. It is not the same 12 tracks, by the way, as on in the FLAC folder.

 

If don't feel like dealing with the mp3s, don't sweat it. It's the CD-quality files that are of greater interest. But it would be fascinating to see whose "scores" change when listening to the lower resolution source.

 

Listen as many times as you like and use the sheet attached here [Stradivarius vs Guarneri violins identification sheet] to record your identifications. Then send your completed sheet to me via PM. I'll collect the responses for two weeks, until midnight on 4/26. Then I'll report the results, with some statistical help from my mathematician brother, and of course will provide the identities of all the instruments. I won't, of course, be revealing any individual's score, so please join in without fear of embarrassment, and encourage others in your orbit, especially musicians, to participate

 

One last thing. Please don't suggest ways to improve the methodology of this investigation. This isn't for publication in a scholarly journal. I'm not trying to emulate the admirable techniques of Drs. Toole and Olive but rather to explore further, with a touch of quantification, the set of considerations that bluesman has opened for discussion.

 

Thank you!

Andy Quint

 

 

 

Strad vs Guarneri 2 Unblinded

 

FLAC_Dis-identified

 

Miracle Makers MP3 dis-Identified

 

Stradivarius vs Guarneri violins identification sheet

 

 

 

[ARQuint]

On 4/15/2021 at 6:31 PM, ARQuint said:

 

Zero takers so far.😪

 

As the original post went up on 4/15, I'll withhold the correct responses until 6:30 PM on 4/29.

(If anybody wants 'em, that is!)

 

AQ

[ARQuint]

Well, I'd argue that this isn't really "listening to music" but, rather, using musical material to judge recordings and playback gear. I'm fairly sure that, as an audiophile. you've done hours of critical listening to decide on purchases and set-up parameters. Once the work is done, we all kick back and listen for pleasure. Though, in truth—whether they admit it nor not—very few audiophiles ever get completely out of the critical listening mindset....

[ARQuint]

22 minutes ago, fas42 said:

Yes, it's very easy to pick the difference between the styles of the two makers, for the marked samples - just on laptop speakers. Okay, I now know the signatures of the makers - so, what's the point of identifying which variant was used for a particular recording ... it's like going back to school, 😉.

 

With all due respect, Frank, could you put your money where your mouth is? My numbers are small to date (which is why I'd like so much to get AS community members involved) but most of the people that have tried to differentiate between the two makers' violins haven't done better than chance.  You're saying it's "very easy," so why not give it a whirl? It would take 15 minutes or less for you to listen to the FLAC version of the 12 unknowns (each sample is about a minute long) and, when the correct identities of the instruments are given on 4/29, if you do as well as you anticipate, I'd love to have you discuss in detail your approach. But you gotta take the trial...

[ARQuint]

Crickets.

It's been two weeks, and no replies via PM.

 

I've been thinking about why no one is choosing to participate in this exercise—or did, but hasn't reported his or her answers. Here are some possibilities as to what people are telling themselves.

 

• It wouldn't be any fun.

 

AudioDoctor put it this way: "You guys make listening to music a chore. Who wants to do that?" My response above was that, although a recording was being employed as the test material, this wasn't "listening to music." Honestly, this complaint surprised me the most. The obsessive nerdiness of what we sometimes do as audiophiles is a big part of this hobby. Is there anyone out there who hasn't had an audio buddy spend an afternoon with them moving loudspeakers around in 2 cm increments to find the best imaging and soundstage presentation? I think that audiophiles, God love 'em, are always listening analytically to some extent, even when the listening is allegedly purely recreational and not focused on achieving better sonic results.

 

 

• It's a poorly designed experiment so don't waste my time.

 

Fair enough. I did ask, nicely, that AS members accept the conditions of the trial as presented. But since when has that stopped any respectable AS forum member from complaining about anything they thought was wrong, intellectually or otherwise? I've not heard many objections to the design of the exercise. Bluesman's superb articles explored some important concepts and received 90 comments. This experiment was, potentially, a way to continue a vital discussion, potentially relevant to our experiences as individual critical listeners.

 

 

• I don't need no stinking test. Of course I can tell a Strad from a Guarneri.

 

This is what we heard from fas42, a prolific contributor to AS, and one who enjoys a good dustup without losing his cool. When I asked that he take the 15 minutes necessary to prove his claim, he said it would, But he hasn't—at least so far. I gather this is kind of what Frank does, wriggling out of providing specifics after he's made some pretty broad assertions. It's the Internet, and it's kind of hard to hold folks' feet to the fire, so to speak. Online, you can make your own rules, when it comes to proving an argument.

 

 

• I might embarrass myself.

 

I suspect this is a major factor and maybe the most important—even though I have  promised not to disclose any individual's score, or even the identity of any subject. At some level, some potential participants might imagine that the next time AQ is reviled as a tin-eared hack suckling at the teat of Big Audio, I'll come back with something along the lines of "Yeah? Remember the violin test? Well, stereogod99 only got 4 of 12 right!" I wouldn't. You'll have to take my word my on that.

 

•          •          •          •          •          •          •          •          •          •          •          •          •          •          •          •          •    

 

So here's what I'll be doing, assuming Chris approves. I'll leave the files up indefinitely. If I get between 1 to 9 responses, I'll send you—privately—your score, 10 out of 12 correct, 6 out of 12, whatever. If I get completed trials from ten individuals, I'll publish the right answers, identifying the specific violins and reveal the 10+ scores without naming names.

 

Why these stipulations? I worry that if I just put the answers up, those who did participate and did well would let us know about it and those that didn't do well would not. That's just human nature. My admittedly selfish motivation is that I'd like to get some more data to add to that I already have, examining the potential for people—audiophiles, musicians, and regular folk—to hear subtle differences of musical consequence with recorded music played back through a carefully configured stereo system.

 

So please (re)consider. Take a half-hour to listen to the identified instruments and then the unknowns. Rope in some interested non-audiophiles and run a group listening session. And send the data!

 

 

Andy

 

 

 

 

 

 

[ARQuint]

2 hours ago, yamamoto2002 said:

Maybe young violinists such as TwoSet redditers can detect the difference of the luthier with their $1 earbuds...

I guess it needs training of ears, and it does not need expensive audio equipment

 

I provide two sets of unknowns: One at MP3 quality and one at CD-quality FLAC. My numbers are small (which is why I'm pleading for audiophile volunteers) but my more successful subjects to date have done better with the higher resolution files. Those that didn't do well with FLAC didn't seem to do any worse with the lossy format.

 

Why not give it a go? I'll send you your results within a few hours. Instant gratification.

 

AQ

[ARQuint]

47 minutes ago, danadam said:

I'm missing an option: I don't need no stinking test. Of course I cannot tell a Strad from a Guarneri. 🙂

 

Semi-related (in case someone wasn't familiar with those already):

• Player preferences among new and old violins (2012): https://www.pnas.org/content/109/3/760
• Soloist evaluations of six Old Italian and six new violins (2014): https://www.pnas.org/content/111/20/7224
• Listener evaluations of new and Old Italian violins (2017): https://www.pnas.org/content/114/21/5395

 

Fascinating stuff. Thanks.

There's a slight difference between these studies and my exercise. The participants in the experiments you brought to our attention were really asking the question "which violins are better?" whereas I'm asking "can you tell the difference between the two makers' instruments?" I do think the latter question is sort of implied by the former in the studies you referenced: the test subjects surely all believe that, overall, the old violins are superior and wanted to show they could identify them reliably by proclaiming their superiority.

 

Anyway, thanks again for showing me these papers.

[ARQuint]

Excellent! Promise to get the results back to you within a few hours.