Tonewood Differences Measured: Interesting Study

[-Asdfgh-]

I've seen people deny that there is an impact made by wood that can be heard. So a demonstration of people hearing such a difference strikes me as useful there. As for how MUCH difference it makes ... depends ... .

It then comes down to whether this experiment was sufficient to demonstrate it. Without knowing the number of participants, 74% doesn't mean much.


EDIT: I read the paper - ignore the above. New comments coming.

 

[Thumpin6string]

We all know that the type of wood, electronics, hardware, pedals and amps have absolutely nothing to do with tone. It's only the fingers.

 

[-Asdfgh-]

Both sapele and rosewood are staple tonewoods in guitar building but are used in different parts of the instrument.

Plywood is used for quite a few bodies too!

 

[-Asdfgh-]

It would be better to have a body and neck, neck the same for all experiments as it would be closer to an actual guitar. The system here is much too stiff overall - it's more like a lap steel than anything else. Also, is only one string is used? It's not much like a guitar overall. It may not matter.

More controls (e.g. the two benches, no body method, cement, resin/cardboard lattice) would have been good to try as extreme options.

The volume of the output should have been normalised - all bets are off without this. It's not clear it was normalised for anything other than the overall volume of the fundamental. It might be worth rerunning it with a different normalisation method. If they recorded the sounds, then if they open sourced the WAV files it could be conducted quite widely with a variety of 'last mile' protocols and normalisations.

The spectral analysis is a nice touch. Interestingly, the frequencies are all within the human range of hearing. All samples are relatively similar apart from the pine.

I'd like to see plots of density and frequency with more points (see the cement option, etc.) but I can do that myself.

I'd like to see differential plots, e.g. sapele to plywood, as on the face of it they look incredibly similar.

I see that the participants could tell a difference at 91.7% with 67 listeners. That's statistically significant.

I may have to modify my view of tone wood if follow up studies agree. See caveats above, though.

 

[Deleted member 365158]

In my experience, when i play different basses unpluged, i can hear a difference. I play my 71 p bass, which is alder, sounds different compared to a bass made from ash. Or maple, as the Pedullas.

Yup!
Such a simple experiment too.
Not only can different body wood sound different, it can feel different.
I noticed this when i was putting my glarry back together (Paulownia wood body)
I strung it up before I messed with electronics to make sure it would still set up ok.
Huge difference in unplugged sound compared to my mahogany body’s or basswood body’s.
They all sound great plugged in But at that point they are drastically different because they have drastically different electronics too.

 

[lomo]

I agree that larger sample sizes are usually good, however it's still necessary to account for the hearing idiosyncracies that the listeners may have and with a large sample size that could be extremely difficult.

If each listener has enough data points, both inter and intra-observer variability can be accurately measured. That would contribute a lot towards the "some people can hear it and others cannot" part of the equation.

 

[GrapeBass]

I think the funniest part of this debate is that people are so stuck in their opinion that they don't want to consider changing their stance. "Show me a study where tone is perceivably different between two woods", show them that study and then they complain about how the study was run or that it wasn't published in a respectful journal. I'm not saying that those people are wrong or picking on them, I'm just pointing out how entrenched these beliefs seem to be implanted in some peoples mind. This debate never ends well but please carry on.

True, and not only for their opinions on 'tone wood,' the current global state of affairs as well. Tone wood philosophy.

 

[brianrost]

Irrespective of measurements, it is most generally not possible to distinguish materials of construction of bass guitar based on purely audible listening, either on tracks or from the audience live. A bass generally sounds like a bass.

If you mean can someone tell that it's an ash body vs. alder, maple, mahogany, etc. probably not. Can they tell the difference between two otherwise identical instruments differing only in the woods? That's what this study was investigating and their conclusion implies people can indeed hear the differences. It's already been well established in many studies that it can be heard in non-electric string instruments. Whether this study was rigorous enough is of course open for discussion.

 

[-Asdfgh-]

If you mean can someone tell that it's an ash body vs. alder, maple, mahogany, etc. probably not. Can they tell the difference between two otherwise identical instruments differing only in the woods? That's what this study was investigating and their conclusion implies people can indeed hear the differences. It's already been well established in many studies that it can be heard in non-electric string instruments. Whether this study was rigorous enough is of course open for discussion.

As I noted above, the log used was wide and thick and homogeneous, so it's more a study of whether you can tell the difference in wood for a lap steel than anything else. It's a step towards better testing, and if the recordings are available, some additional work can be done without the need to reproduce the physical elements.

 

[bigdaddybass12]

I have always said the finish can affect the tone. Oiled finish or 1/8" of paint. Paint has to reduce the vibrations or resonance. Also, how hard the bass presses against your body. Again, reducing resonance.

 

[bigdaddybass12]

If you mean can someone tell that it's an ash body vs. alder, maple, mahogany, etc. probably not.

Some studies show people cant tell the difference between a Fender & a VM Squier!

 

[40Hz]

I think it’s an honest attempt by the researchers, and a step in the right direction. But as the authors themselves said, it’s still inconclusive. And they also acknowledged what I felt was the most serious potential flaw in the design of the experiment itself. That took a remarkable amount of candor IMO and definitely boosted my respect for their scientific integrity.

But like all legitimate and carefully conducted studies so far (and this is probably the best I’ve seen to date) they all end up saying “inconclusive” and “further study is needed.”

And I think we can all agree on that much at least.

So, it appears the jury is still out. But at least researchers keep trying to find a definitive answer to this puzzle. And that can only be a good thing.

Props to the researchers for such a concise and readable paper. It was actually an enjoyable and thought provoking read. I’d have liked to see more of the actual test sample data and the maths. But it’s a summary research paper, not a lab notebook, so it’s not unreasonable they left those details out for the most part.

A+ paper all the way!

 

[James Collins]

Sapele, plywood and pine - not generally reputed as tonewoods, and rosewood is rarely used to construct a body, unless you are George Harrison. So odd choices.

Plywood is used in double basses as a tone wood. Sapele and rosewood are used in guitar necks and bodies respectively. Pine i guess could be, but generally I agree on that one even though I've seen electric guitar bodies in pine.

They might be better off creating one of those cardboard and resin bodies for a bolt-on neck, and then a concrete guitar, and seeing if people can distinguish those as those are probably the two extremes. And ensure that the recordings are normalised relative to each other. And also do things like not just do AB and ask the difference, but multiple combinations of A and B in random orders, different for every listener.

How do you know cardboard and concrete are the extremes? I think you have to make some assumptions to get to that point.

What does recordings are normalized to each other mean? If I build two identical acoustic violins but one with a mahogany body and one with a applewood body, and the mahogany is louder, why do I have to normalize the volume? I get that if the volume of one is louder it is easier to hear a difference, but if you normalize the differences away that seems like a taking a measurable difference and making it unmeasurable.

Interesting but incomplete, as they're missing a crucial aspect in my opinion: they should have used multiple examples of each wood to get a sense for the variation among a single 'species'.

What they show for each wood is the variation between measurements, or the technical reproducibility. That's cool, but it doesn't say anything about the variation between different examples fo the 'same' wood. That metric would be far more informative.
I imagine some woods will exhibit far more variation than others, and for some the overlaps will be pretty big while for others the overlaps will be small, giving far more predictive power to their usage...

Otherwise, we're still talking anecdotal evidence. I'm pretty certain there are real differences (whether big enough to matter to individuals, that's another matter, it clearly matters to some and not so much to others) but when you get some ash that seems porous and light and other much tighter in grain and heavier, you can't just take one single sample and extrapolate to the whole species. I mean, you can but if I were a reviewer I'd ask for some further work or to qualify their conclusions accordingly.

Agreed. But that would be asking a different question: what is the variability between samples of the same species of wood? In which case, how many samples do you need to do? How many species do you need to do? You couldn't for example just choose one species of tree because maybe Bolivian Rosewood is very uniform and if you choose that then you make it look like a species is uniform. Or if you choose woods like "maple" you have to question which maple tree or where was it grown, etc. Or if you happen to choose a wood with wide variability, like eastern long leaf pine, then maybe you make the results look like variability within a species are so high as to make it a useless endeavor to care about the wood.

Also, this study is probably easier to do because if you measure no difference, then I think you can probably just move on. Of there is no difference between any of the samples, what is the point of going back and checking same species variability. Whereas this answers one question: what the string is attached to affects string vibration. And leaves more questions to answer.

 

[LBS-bass]

I really have no questions in my mind about this, as it seems intuitively obvious if you think about extremes. Make a bass out of a pillow and see how it performs. Make another one out of steel and do the same. It seems intuitive that diffusing the string vibration by reducing the density of the material used would create a sonic difference. That really is the crux of the question. So I have no horse in this race, because it seems clear enough to me.

The problem is that we don't think about pillow vs steel, we think about this wood vs. that wood, which makes the differences more subtle and introduces the possibility that other factors might override the density difference. But I don't care about that argument. All other things being equal, the wood should contribute in some way. Whether what it contributes is worth the other tradeoffs (weight, primarily) is another question.

 

[bucephylus]

If you mean can someone tell that it's an ash body vs. alder, maple, mahogany, etc. probably not. Can they tell the difference between two otherwise identical instruments differing only in the woods? That's what this study was investigating and their conclusion implies people can indeed hear the differences. It's already been well established in many studies that it can be heard in non-electric string instruments. Whether this study was rigorous enough is of course open for discussion.

Understood. I’m just saying that I have a hard time hearing those differences; especially if we are talking about recorded tunes. I certainly can’t tell you whether a particular tune has a rosewood board or a maple board on the bass going by the recorded sound. Yet Roger Sadowsky will tell you that is important; and I have a lot of faith in his perspective.

So, I’m just relating what I can and can not hear. I do feel / “hear” the differences while playing. I don’t hear the differences in general on recordings. Perhaps, just my reality. Perhaps, there are those out there with better hearing than mine.

 

[mambo4]

Unless I am misreading it, only 1 sample of each wood type was used (that critical detail is glaringly absent, as are soooo many others) . If this is the case, the conclusions are laughable.

That was my thought too.
I want the Venn diagram of the variance within a species compared to the variance between species. Plus Plywood? why test that at all? Are builders seriously using plywood for guitars?

 

[Christophe C]

And then you add a compressor, an overdrive, a preamp with some eq,…
In the end : Select the best looking and most comfortable bass

 

[Christophe C]

Are builders seriously using plywood for guitars?

yes in the seventies and eighties for fender copies made in Japan. It sounded good … really … but heavy.

 

[-Asdfgh-]

That was my thought too.
I want the Venn diagram of the variance within a species compared to the variance between species. Plus Plywood? why test that at all? Are builders seriously using plywood for guitars?

Were there multiple audio samples of each construction, and was the FP rate for change measured. That is, could the subject be played A1, A2 in order where each of A1..N is a particular construction and frequency of string and be asked to select if there is a difference? That gives you a false positive rate if people say A1 and A2 are different, as they aren't in terms of construction, although there might be differences with each pluck from the picking machine.

 

[James Collins]

Were there multiple audio samples of each construction, and was the FP rate for change measured. That is, could the subject be played A1, A2 in order where each of A1..N is a particular construction and frequency of string and be asked to select if there is a difference? That gives you a false positive rate if people say A1 and A2 are different, as they aren't in terms of construction, although there might be differences with each pluck from the picking machine.

This is the gist of the audio sample listening statistics:
https://en.wikipedia.org/wiki/Two-a...ed choice (2AFC,versions of the sensory input.
They would always be present with A0 and then be presented with An and Bn and forced to make a decision of which matched A0. I don't understand how identifying the instrument correctly is a false positive.