Sign in to follow this  
DMartin

Fritz-blind test

Recommended Posts

Just heard on National Public Radio (east-coast USA) an interview with Claudia Fritz on her recent publication describing results of a blindfold playing comparison of new and old instruments. For the sake of accuracy,I won't try to recall details at this moment.

Linkman, we need you.

Claudia is well known to participants in the Oberlin workshops.

Doug Martin

Share this post


Link to post
Share on other sites

Thanks for the link, John.

Unfortunately, my phone wouldn't play the two samples (Strad and contemporary) and I saw the answer at the bottom of the article.

Share this post


Link to post
Share on other sites

This is a double-blind test: Neither players nor listeners know which instrument is being played. So the results are especially interesting.

I couldn't have told you which of the two fiddles was which, and have seen the results of enough of such tests and have played contemporary instruments which were as fine as the best Strad (the Betts) I've ever played to know that making that decision is a fool's errand. My preference was number one, the new fiddle.

"When are we going to make fiddles as good as the finest classic Italians?"

Maybe that's a stupid question for the following reasons:

Which are the finest classic Italians? Do they all sound the same?

Some contemporary instruments sound better than some classic Italians already.

Does anybody really believe you're going to make something better than X by simply copying X?

Are the notions of great violin sound today the same as they were in middle of the 18th century? If you can soup up a 1930s Chevrolet to run as fast as a 2012 Corvette, does that mean that generally 1930s Chevrolets are faster cars than 2012 Corvettes?

Share this post


Link to post
Share on other sites

well besides the fact that the strad seems to be playing an octave lower than the modern(because it has so much more fundamental), the modern sounds kinda dissonant, obnoxious by comparison to the strad, not even close in my book

i dont know if mr curtin knows anything about the rigermaroll behind double blind testing, but devising a double blind test of strads vs i presume the tester, mr curtins modern violins and overseeing the whole test, well thats not exactly double blind, is it, in a true double blind the tester is not supposed to know whos playing what

im sure a double blind test of bmws vs mercedes designed and operated by bmw mechanics and engineers is going to favour bmw, just has to doesnt it!!

Share this post


Link to post
Share on other sites

You got lucky. The STrad was the better violin...

Me too...I thought it was pretty obvious...

Really? I thought violin 2, the Strad, sounded pretty thin on top, generally uneven over the various strings. But we're making judgments based on recordings played back on our individual speakers.

Share this post


Link to post
Share on other sites

Really? I thought violin 2, the Strad, sounded pretty thin on top, generally uneven over the various strings. But we're making judgments based on recordings played back on our individual speakers.

I would not've recognised it as a Strad if I would not've known one of the two must be a Strad

Share this post


Link to post
Share on other sites

"...the modern sounds kinda dissonant, obnoxious by comparison to the strad, not even close in my book..."

You know you're setting yourself up for a test where the answer isn't at the bottom of the page?

Share this post


Link to post
Share on other sites

Very exciting article, this can only be positive for new makers. More experiments like this please.

More experiments will just show the same thing. What's needed more of is makers building violins based their own cultivated sense of the sound a violin needs to have for today's standards. But that implies that makers come to building with a goal of achieving a concept of sound in their heads. But that implies developing a concept of sound, and not making violins by the numbers, whether the numbers are plate dimensions or some kind of numbers derived from physics/acoustics.

Share this post


Link to post
Share on other sites

well i predict in a double blind test conducted at the louvre in paris, no one could tell the difference between the mona lisa by da vinci and a cartoon of bugs bunny, just goes to show you with a double blind test you can prove just about anything

Share this post


Link to post
Share on other sites

The bottom line for me is that, based on the two samples, the player had captured my willingness to listen to the player play the rest of the concerto on either violin, whichever one the player wants to use.

But I go to concerts to hear players, not instruments.

Share this post


Link to post
Share on other sites

"...the modern sounds kinda dissonant, obnoxious by comparison to the strad, not even close in my book..."

You know you're setting yourself up for a test where the answer isn't at the bottom of the page?

I agree, it's pretty easy when the answer is right in front of you (in black and white). I personally liked the first one better, it sounded clearer and not as choked as the second. My question is who made the first one. This ought to be exciting for us as makers, the gap is getting narrower.

This is what I was saying in the thread "Perseption Is Everything".

Berl

Share this post


Link to post
Share on other sites

well i predict in a double blind test conducted at the louvre in paris, no one could tell the difference between the mona lisa by da vinci and a cartoon of bugs bunny, just goes to show you with a double blind test you can prove just about anything

Geez, Lyndon, this was a test deliberately trying to remove the visual component, and also the prior knowledge component, so how does your painting metaphor fit in? Shall we put a bow to the painting? :lol:

And it's even more interesting because so many published tests have been based on audience evaluations, and when they didn't turn out as expected, some went on to criticize the listening ability of the audience. As I read this one, it was all about player impressions.

It isn't anything which hasn't been done many times before, but some such tests are private (so I can't always talk about everything I've been exposed to), or are not published.

Edit:

I'll also say that the differences are obvious, but at the same time, not necessarily sufficient to tell more than the difference between two Strads.

Share this post


Link to post
Share on other sites

Yup, I got it too. It was pretty easy to tell the difference. The first one had no, for lack of a better word, "magic", in it's tone. That's a really awful word to describe it, just don't have a better one.

Share this post


Link to post
Share on other sites

Just heard on National Public Radio (east-coast USA) an interview with Claudia Fritz on her recent publication describing results of a blindfold playing comparison of new and old instruments. For the sake of accuracy,I won't try to recall details at this moment.

Linkman, we need you.

Claudia is well known to participants in the Oberlin workshops.

Doug Martin

I thought the first was the strad, but thought they were very close. This was a well designed double masked (we don't like "double blind" in my field of diabetes because blindness is the outcome of interest). BRAVO on this test by NPR. I note that 7 experts couldn't tell, 7 got it wrong, and only 3 were correct. Didn't see a statistical analysis except a trend that the strad was the worst, but doubt this was more than a chance result. However, I truly wish I could have picked out the Strad. I note (ref previous suggestions) that it was double-masked, the players were blindfolded, there were multiple expert players. Smell was not masked (not hard to do, a little aromatic petrolatum in each nostril like the pathologists do). It would have been a more objective test if each observer had rated each instrument on multiple factors (don't know what these might be but perhaps tone, timbre, projection, sustain??the experts could pick out the outcomes of interest). For these each observer would rate each characteristic on a Lichert scale of 1-10 (better than 1-5), like ice-skating judges. Then the results would be compiled and an appropriate statistical test done. If this was done it wasn't apparent from what I could see. This test could better be done with each player playing the instrument in an acoustical environment (recording studio, concert hall etc), and the recordings sent to multiple experts to judge based on pre-selected criteria. Doubt a room is the best venue for this experiment. This would be cheaper than getting everyone in the same room. I am not clear if it was a blind vote without discussion (better) or if they talked it out. Having everyone in the same room makes the claim of double-masked a little dubious. Too many opportunities for subtle clues to be perceived by others. Great test nevertheless!! Not conclusive in IMHO, but a good start.

Share this post


Link to post
Share on other sites

Here are the long term average narrowband FFT spectra. They are extracted from mp3 files cutting at some 12kHz ish.

[Edit] Included a version of graph where it might be easier to compare the two curves in the highs.

post-25136-0-92494800-1325545169_thumb.jpg

post-25136-0-50430300-1325545927_thumb.jpg

Edited by Anders Buen

Share this post


Link to post
Share on other sites

Here are the long term average narrowband FFT spectra. They are extracted from mp3 files cutting at some 12kHz ish.

Looks very hard for the human ear to tell the difference1

Share this post


Link to post
Share on other sites
Guest
This topic is now closed to further replies.
Sign in to follow this  

  • Recently Browsing   0 members

    No registered users viewing this page.