Official Luthiers Forum!

Fylde Guitars were mentioned a couple of days a go. Here is a blind test of various back and sides wood built by them:

https://www.sciencedaily.com/releases/2019/01/190122115013.htm

Seems to be a trend here

Ed[/url]

Yeah that video has been making the rounds. I tend to like the domestic woods myself though I am just about to build my first BRW guitar. I'll let ya know in a few months

"Overall our results suggest that the back wood has a negligible effect on the sound quality and playability of an acoustic guitar, "

I'm not sure what that says about Fylde instruments or the people testing them. And what do they consider "negligible"?
I do agree many woods can be substituted for endangered species with good results, as long as those woods share similar qualities.
I do think, however, that woods with dissimilar qualities do change the sound of the finished product.

Ironically I was reading an older issue of AL last night and they had a similar article on classical guitars. Fifteen more or less similar guitars, five out of tropical hardwoods (rosewood) and 10 out of sustainable non-tropical woods. They were played by blind folded players to a critical audience who evaluated the guitars. Interestingly when the listeners could not see the guitars (played behind a screen) the tropical and non-tropical were almost 50/50 as far as best sounding. When the audience could see the guitars they preferred the tropical ones by a significant number. Mmmmmmm.....

http://www.luth.org/back_issue/al121-124/al124.html

Tropical vs. Nontropical Woods for Classical Guitars
The Leonardo Guitar Research Project
by Brian Garston and Jacky Walraet
Everybody loves rosewood. Sounds great, looks great. In fact, it looks so great that we might be “hearing” its lovely, familiar deep color. Would non-tropical woods also make good-sounding guitars? And would they “sound” different if we took off the blindfold? To answer these questions, the Leonardo Guitar Research Project built a matched set of fifteen guitars and compared them under controlled conditions.

Blind tests don't matter unless you always wear a blindfold when playing or going to a gig.

That's like saying placebo controlled clinical trials only matter if no one reads the prescription labels after a drug is on the market.

_________________
Bryan Bear PMoMC

Take care of your feet, and your feet will take care of you.

Aha! Brain scan research which I have obviously only vaguely read somewhere, show that compared with identical generic drugs, branded drugs do work better for pain relief. They actually reduce pain more than identical drugs if the subject knows they are branded. Another I really like, again with brain scans, subjects enjoy identical wines more if they are told they are expensive. The exact same wine actually tastes different according to what subjects are told. Senses don't work in isolation so why pretend they do? I'm just being awkward really!

I'm just just being a butt-head, I knew what you were saying My point was just that controlled tests are important in answering a questions like is there a detectable difference. Of course the other influences still exist but it is nice to know if it is largely due to bias or if people really can hear a difference.

Let's pretend that we were able to do enough reliable testing to demonstrate that traditional woods don't offer any real advantage to the potential of a guitar (again, we are pretending this is also true). Your point is still valid because the buying public are still influenced by the biases that exist. But if we demonstrated that there really was no difference, over time that would be accepted and much of that bias wold go away.

_________________
Bryan Bear PMoMC

Take care of your feet, and your feet will take care of you.

It's not even just biases as in 'oh rosewood sounds best'. The sound of dark coloured guitars may be processed differently in the brain so they genuinely can ' hear a difference' when able to see the guitar despite there being no measurable, material difference in the real and measurable world!

I would call something being processed in the brain differently based on the look of the wood a bias. Most of our biases we don't even know we have.

_________________
Bryan Bear PMoMC

Take care of your feet, and your feet will take care of you.

I would say that hearing happens in the brain and the way it processes the information could depend on the colour of the guitar, what you ate for lunch, the personal inner musings of the subject at that moment!

I think I should probably leave the house and speak to some actual real adults tomorrow. I spend the majority of my time with my own or with my four year old son.

I've been involved in a large scale long term clinical test of something. It was double blind, participants took two different substances, one or both or neither a placebo. Test went for 10 years, I faithfully took my pills (even to the point where I would take a supply on trips so I wouldn't screw up my part of the test). Each year I had a physical and a blood sample that was sent to the testers. Twice a year I would fill out a fairly long form on my medical history and any changes. It would start with a question like "during the past 6 months have you died? If so, which one of these caused you to die?......

The results were published just last month in a rather prestigious medical journal. Surprised a lot of people.

I've also been involved in a couple of blind listening tests, including Alan Carruths work on sound ports. I think a lot of people have been surprised by these also.

I agree that as the brain processes sound in a unique fashion in every individual and as scientific research has shown that the brain does not actually process every bite but actually fills in the gaps with what it thinks should happen, factors other than the actual sound may play a huge role in what we actually hear.

Guitars have definitely sounded different to me on different days, morning or evening, or with more or less sleep. Same guitar.

On top of that every guitar is going to be unique in some way which brings up the controversial question of -is commissioning a guitar actually a lousy way to buy an instrument?

Tell us more about your study Freeman! What was the title and journal? Did it turn out exercise and eating your fruits and vegetables is actually bad for you?

_________________
It's not what you don't know that hurts you, it's what you do know that's wrong.

Can anyone on here please tell me if the backs of the guitars in this study is active by the graph from the paper. It seems like the Yamaha might be.

Does blind drunk count?


Sent from my iPhone using Tapatalk Pro

A similar study was done in Europe with the players and the makers blindfolded. Not only did the players get things mixed-up, but
the builders could not even pick out their own instruments. For myself, I only see (and hear ) two categories really. Softer materials
like walnut, mahogany etc and harder woods like rosewoods etc.
Sergei De Jonge is one of the best in the business and when he was asked what his favorite kind of wood to work was, his rely was
'the kind that comes from trees' .

Brent

It's nice to see some actual discussion of blind testing, rather than the usual simple bandying of opinion. I often wonder if the folks who complain about these studies have ever participated in one, let along tried to set up a 'good' scientific study of something as complicated as guitar acoustics.

That is a very interesting test. I would have have guessed there would have been a distinct difference between oily, heavy woods (rosewood ebonies and such) and the lighter stuff.

It demonstrates once again how none of our five senses work on their own.

_________________
Joe Beaver
Maker of Sawdust

I have several guitars and they are all pretty decent sounding. However, my favorite is a plain birch back and sides in terms of sound. Not nearly as nice as my rosewood or wenge or padauk, but sounds so much better. For a player, the look of a guitar seems to be important as well as the sound. I like nice guitars. i.e rosewood, etc.

_________________
Under Compensated Nut!

I would venture the players and builders got muddled because they weren't used to not being able to see their guitars. I imagine their brains were processing sensory information very differently to normal as they weren't used to being deprived of a major sensory input. I'd say a blind guitar player would do a lot better under those conditions.

So reading the results: 52 "blinded" guitarists rated the sound of the guitars (good, fair, poor?) and gave similar ratings to them (similar but not the same). 31 of those same guitarists couldn't distinguish the guitars by their sound or feel (21 could?). Acoustically the differences between the guitars were minimal (they all sounded like guitars?). From this they draw the conclusion that the back and side woods have a "negligible effect" on the sound of the guitar.
Although I do agree that cheaper sustainable woods can be substituted for expensive endangered woods without loss of sound quality, I think those woods need to be selected to have similar qualities to those they replace. And although I don't think the back and side woods determine the sound of the guitar totally, I think they affect it to some degree.
I think studies that have a predetermined point of view publish the results in a way that makes them fit their message.

I think I posted this earlier - it is an extension of the Leonardo Project that Freeman posted above. 16 guitars by 8 makers built to the same design, 1/2 in tropical woods, 1/2 in non-tropical. I know your computer speakers are small, but can you say which one sounds "better"?

http://www.leonardo-guitar-research.com

In Freeman's post, in blind tests the listener preference was 50/50, but in "unblind" tests, the tropicals won 75/25.

And I don't see why I would have to to be blindfolded whenever I listen to get the same results. Only if I were "trained" to hear that lovely Rosewood sounded better would I think it sounded better. I personally have not been so trained.

Ed

Note that when you get a 50/50 split in a blind test it normally means that people are guessing, they can't really tell the difference. The safest choice of responses to solicit is something binary, such as 'same' or 'different', or 'rosewood' or 'not rosewood'. Too much nuance complicates analysis and requires an even bigger sample. 'Small sample size' is already a problem in this sort of thing.

Why would listening 'blind' alter your perception? Or, if it did, why would it always seem to do so in a way that causes you to confuse rosewood and mahogany guitars, or whatever, rather than reinforcing any difference? 'Blind' does not always mean 'blindfolded'; it means that you can't see the guitars. I've done blind tests where the player sat in the back of the room so that the listeners could not see. Another common method is to put the player behind a thin cloth screen, which does not block the sound but does prevent seeing the guitars well enough to pick them out. You can argue that the listening experience has been changed in some way by this, and it's likely enough to be true to some limited extent, but what counts it whether this introduces a bias. I can't see why it should.

Master violin maker Joe Curtin made the point that putting a violin in the hands of a good player is not a good way to find any differences, because part of the qualification of a good player is that they can get the sound they want from any reasonable instrument. Bad payers are no better; they can't get a good sound out of anything. Playing tests are of only limited use. OTOH, it's hard to find anything that everybody agrees is better.

Joe Beaver wrote:
"I would have have guessed there would have been a distinct difference between oily, heavy woods (rosewood ebonies and such) and the lighter stuff."

This brings up the question of what you need to control. There are several things you could think of that might cause different back woods to have varying effects on the sound. A heavier back would tend to move less than a lighter one, and thus ought to be more 'reflective' and less, 'active', all else equal. Leaving the wood thicker in a less dense back to get the weight up would tend to move the pitches of the back tap tones upward, though, as it would normally be stiffer. Getting both the mass and the resonant pitches matched would be tricky, and changing either could well affect the tone. There are a number of other variables that could plausibly alter tone that would be even more difficult to get under control, and that's just the stuff that we've already thought of.

printer2:
It's possible that the dip in the spectrum up around 300 Hz in the Fyldes is linked to the 'main back' resonance. That's about a fifth higher in pitch than the apparent top resonance, which is not all that uncommon. In the Yamaha the notch in the big peak at about 200 could well be an indicator of coupling between the 'main top'and 'main back' resonances. It would be a lot easier to tell what's what with some more data, such as Chladni patterns. We're unlikely to get that at this point, though.

Given that making 'identical' guitars from 'the same' wood that sound the same is, at best, very difficult (you should try it some time!) the fact that people only hear 'small' differences between guitars made with different B&S woods suggests that that doesn't account for much of the change. The question becomes 'how much' along with, of course, 'how do you know'? I'd love to see a test with some minimal number of variables, say two types of B&S wood (soft mahogany and hard rosewood) with some other parameter, such as back tap tones, fixed, and tops matched as well as possible. You'd need some reasonable number of instruments, but pretty stringent controls, which works against a large instrument sample. I think you'd be lucky to get six of each. The choice would be 'mahogany' or 'rosewood' in a blind test. If people could actually hear that reliably we'd be in a position to start asking what it was they were hearing.

Personally I think the methodology they used was seriously flawed and thus the conclusion is a load of bull. The methodology is very likely to produce the result they got, and not reliably answer the question they want answered. Human hearing memory is short, only a few minutes. After about 3 minutes reliability goes down rapidly. You can't possibly reliably evaluate and rate 6 guitars in 3 minutes! Also, rating guitars the same does not mean they sound the same. The randomised double bind tests are designed by scientists to eliminate bias, but often the evaluation process they design becomes unreliable because of our own human limitations. The only really reliable way to evaluate music instrument sound is a one to one comparison that is within that 3 minute interval. That is how I do it, and hopefully all the rest of us do it. It is very easy to fool yourself if you go outside of that 3 minutes. I use a known reference instrument as my reference to do the initial comparisons and then move on to one on one comparisons of the other instruments if I have more than one, swapping back to the reference to make sure I have got it correct. You always need a known reference point to refresh your short term auditory memory. I think Alan has it dead right, the number of variables needs to be reduced to just one. A much better method would be to reduce the comparison to 2 different back woods and in a blind test see if people can reliably hear a difference between a reasonable number of guitars. A known reference guitar would still be necessary as a memory refresh. I suspect the conclusions would be somewhat different. My ears and experience tell me it would be different. Getting reliable subjective evaluations free of bias is not easy, and has a practical limitation of what instruments and how many are actually available to evaluate. Bit of a can of worms.

peter.coombe wrote:
"After about 3 minutes reliability goes down rapidly."

I've been told it's more like 15 seconds, and you can't have any sort of sound in between. Whatever: you're absolutely right about the need to be quick, and quiet too.

The tests I've done have all been pairwise. a couple of examples:

For a test of string height off the top vs break angle I used very carefully made recordings of the same guitar with different setups, since it was impossible to change the setup quickly and silently. There were three cases, A, B, and C, so I had three 'synthetic strums' compiled from well-controlled individual string plucks. Every possible pairing of these was tested: A-A, A-B, A-C, B-B, B-C, and C-C, and the reverse of each, by playing the recordings back through headphones. Listeners were asked 'same' or 'different'. With about 100 full sets of 12 pairs the result was that virtually everybody could hear a difference in string height off the top, but nobody could tell when the break angle was changed with no difference in string height.

More recently I tested two 'identical' mahogany/Red spruce OMS. A 'ringer' guitar was included to see if it would be any easier to tell from the test instruments, so, again, we had A, B and C, and the same sort of pairing was used, in random order, of course. In this case tests were live, with a player in the back of the room. After every pair the audience was asked 'same' or 'different'. The rooms were small and 'live', with about 10-15 people each time. We tried to get quiet rooms, but one set of tests was conducted at a folk festival, and for one session there were people practicing bagpipes in the next room (really!). Again, with a hundred or so listeners, the results were quite clear: everybody could tell the 'identical' guitars apart just as easily as they could tell either of those from the 'ringer'. Back to the drawing board...

It has been said that one reason science is so difficult and time consuming is that you're always doing something you're not good at. You're addressing a question that nobody knows the answer to, so you have to do an experiment that's never been done in exactly that way before. By the time you get good at running the experiment, you've got all the data you need, and it's time to think of another question. This is especially true when you're talking about something like the guitar, where there is no agreed performance metric. With drag racers you can go down to the track and see which one covers the quarter mile faster, but everybody has a different idea about what makes a 'better' guitar, and they're all correct.