Tonewood Differences Measured: Interesting Study

[s_wood]

I can't believe I am posting about this. As a long-time member of Talkbass, I know --Lord, I know-- that these threads never end well.

But this study is interesting and worth sharing. It was recently published in the Archives of Acoustics by several researchers in Department of Mechanics and Vibroacoustics at the AGH University of Science and Technology in Cracow, Poland. It concludes that that different tonewoods in an electric guitar produces a measurable and audible difference in sound:

Abstract (excerpt):
"The presented study examined how the sound of a simplified electric guitar changes with the use of various wood species. Multiple sounds were recorded using a specially designed test setup and their analysis showed differences in both spectral envelope and the generated signal level."

Conclusion (excerpt):
"The tonewood used in the construction of an electric guitar can have an impact on the sound produced by the instrument. Changes are observed in both spectral envelope and the produced signal levels, and their magnitude exceeds just noticeable differences found in the literature. Most listeners, despite the lack of a professional listening environment, could distinguish between the recordings made with different woods regardless of the played pitch and the pickup used. The conducted test does not allow any conclusions regarding a more holistic outlook on a guitar’s timbre, as the observed relations are complex beyond the scope of the obtained results. Further tests regarding the guitar’s entire frequency response are required. It should also be noted that these tests were conducted based on a simplified guitar model, which might have overemphasized the researched phenomenon.”

The whole article is here: Jasinski et al. Archives of Acoustics 2021, vol. 46 no. 4, pp. 571-578. https://journals.pan.pl/Content/121810/PDF/aoa.2021.138150.pdf?handler=pdf

Let's discuss the study. There's another thread going on here The Tonewood Debate - Finally Solved? that's a forum for subjective opinions about whether tonewood matters. I get it: some think it does, others think it doesn't. Let's not clutter this thread with that noise, and instead let's critique and discuss this interesting study.

That last paragraph reflects optimism that is almost surely misplaced.

 

[jd56hawk]

Well, case closed, right?
Seriously though, somewhat interesting article, but...
(And that's a very big but.)

Regardless, I just posted a video I made about tone after someone said they definitely prefer passive basses.
Here's his response:
Joe Delaney sounds like both pickups on full or maybe slightly biased to the front with the tone rolled mostly off. Good for accompaniment

(I was playing this.)

 

[Ghastly]

I'd be curious know more about their "specially designed test setup". Further, I'd like to know how they determined that the listener could distinguish the differences. Did they use the same recording randomly or just play the different recordings back to back, with the listener expecting something different each time? I have no strong opinion on the topic, either way. I'm just curious.

 

[bucephylus]

Has anyone yet stopped to consider that both sides of this discussion can be correct? One possible reason the discussions rarely end well.

Here are two aspects that are both true.

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.

At the same time, nuances in the differences between instruments can be highly important to both players and band mates. I readily distinguish differences; and getting these aspects where I prefer strongly affects my ability to concentrate and play parts well.

Between these two aspects, we can argue back and forth all day and forever; though it does get somewhat tiring. To say the least.

 

[s_wood]

I'd be curious know more about their "specially designed test setup". Further, I'd like to know how they determined that the listener could distinguish the differences. Did they use the same recording randomly or just play the different recordings back to back, with the listener expecting something different each time? I have no strong opinion on the topic, either way. I'm just curious.

That's all explained in the methodology section of the study.

 

[s_wood]

Has anyone yet stopped to consider that both sides of this discussion can be correct? One possible reason the discussions rarely end well.

Here are two aspects that are both true.

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

At the same time, nuances in the differences between instruments can be highly important to both players and band mates. I readily distinguish differences; and getting these aspects where I prefer strongly affects my ability to concentrate and play parts well.

Between these two aspects, we can argue back and forth all day and forever; though it does get somewhat tiring. To say the least.

I agree with all of this, especially that last paragraph! My point wasn't to join the debate, which as you rightly point out is tedious. Rather, I thought the study was interesting and unprecedented (at least to my knowledge), and worth our discussion.

 

[ac100X]

The differences in wood bodies may be small but I used to own an Amped plexiglass bass, no mistaking the difference there, you knew it wasn't wood. It had a very solid defined punch in the low bass, it sounded like plexiglass to me.

 

[Zoffy]

OP, thank you for linking to the actual study and not just referencing from it. The authors concede that the effect of the wood is a smaller consideration on an electric guitar compared to an acoustic guitar and they limit their sample woods to sapele, rosewood, plywood, and pine, however do conclude that "[m]ost listeners, despite the lack of a professional listening environment, could distinguish between the recordings made with different woods regardless of the played pitch and the pickup used." My ear is not so attuned so I would bet that (to me) they would all pretty much sound the same, but it was interesting to read the results of an actual controlled study.

 

[MultiScaleMale]

Tone wood certainly can have a difference in the tone of an instrument, the question is "How much of a difference does it make?" In their selection of radically different woods, only a single note plucked, pickup embedded in a large block of wood, it emphasizes the contribution of the wood. A better experiment would be using bodies and necks made out of different woods and using them in an actual guitar. Then play chords.

 

[truand85]

Well, if signal level was affected that invalidates things from a "functional" level... You can always strum harder... Or turn the volume up. It'd be interesting to see the results at the same output levels.

Edit: Ive read a lot of "audiophile" studies on speakers, cables, amps, preamps etc...
, double blind tests, and when volumes are matched, differences go out the window.

 

[StayLow]

Tone wood certainly can have a difference in the tone of an instrument, the question is "How much of a difference does it make?" In their selection of radically different woods, only a single note plucked, pickup embedded in a large block of wood, it emphasizes the contribution of the wood. A better experiment would be using bodies and necks made out of different woods and using them in an actual guitar. Then play chords.

And multiple wood samples from different sources.

As always, terribly flawed methodology throughout AND not at all reflective of real-world scenarios.

Even if for some reason someone were to accept it though, this part was most amusing: “The test also showed that certain wood combinations were much harder to distinguish.” It makes the “I selected these 9 different tiny strips of wood to impart these specific tonal characteristics…” statements even more funny.

Is that a win/win, or a lose/lose?

 

[frankie5string]

Actually, what the test shows is that there can be a marked difference between different planks of wood when the strings are plucked mechanically. It is a massive and unscientific leap to take that study and say it applies to a guitar in the same way.

Aside from the "Poop stinks" study, that is one of the shortest write-ups I've ever seen

They need significantly more research and testing to do that - but I doubt any organization would fund such testing.

 

[TyBo]

The fact remains that listeners could discern a difference in the sound generated with different wood samples when amplified. That calls into question the "it's all in the pickup" view ... we know that pickups are vitally important, I don't think there's much dispute about that. But perhaps it's worth considering that other factors can enter in, as seen with these different wood samples. So in that sense, relevant on that point, despite all the other myriad construction factors that enter into sound.

On a similar theme, we all know that woods have an effect on acoustic tones, clearly. And I think (or hope) that most of us have experienced the difference between a magnetically amplified acoustic instrument and a solid body instrument. It's not ONLY the pickup that generates sound.

As one acecdote , I changed the bridge on one of my basses to a Babiscz "Full Contact" bridge ... I noticed a distinct difference both amplified and unamplified. The whole instrument just resonated more, and that came through the speakers. Not intended as an endorsement, simply my experience. Should I believe someone's theory, or my own ears and fingers? It is what it is ... though it's nice to have some study to point to, I don't need that to understand what I've played and heard.

 

[James Collins]

Disclaimer: I am biased to believe what wood the necks and bodies are made of impacts the tone in some way AND that modern construction practices do everything to eliminate those nuances.

Regarding the study and my criticism:

I think that they demonstrated their stringed instrument which is somewhat similar to a guitar was affected by the instrument wood.

Overall they did a good job of controlling variables they didn't want to test (mechanism of plucking, same electronics harness, same setup, etc).

I didn't like the simulated guitar body. I would think it was possible to CNC route 4 identical neck through guitars, but i can understand the cost would be higher. Also, the added variables of frets and fretboard, truss rod, and others might create some problems that would make hearing the difference more difficult.

I was surprised how much that people were able to tell a difference. However, I cannot tell if there are 3 samples or just 2. Meaning are the playing Sapele as the known followed by a different Sapele recording and another tonewood, or are bother Sapele samples the same recording. My reading of the study sounds like it is 3 different recordings which is strong evidence that it can create a difference. If it is the same recording, that would be less impressive.

Also they make a wild conclusion that plywood sounds the same as rosewood when they said the correct responses were 76% of the time and 75% was considered a true difference. I hate when papers state that something approaches statistical significance. If it passes the thr

I have to take them at their word on the spectral analysis. That is the author's specialty.

I am left thinking that tonewood can play an important part in the tone generation of an instrument. However, like their final sentence of the conclusion, an actual guitar is a much more complex system across many more pitches than the 3 tested pitches.

Someone mentioned the difference in volumes on the recordings. Am I wrong, but they were talking about the different volumes of selected harmonic pitches, not the overall volume. Basically, I thought they were saying the overtones we're 5dB different between samples, not that the recording was 5dB louder. The fact that the overtones are different is what you would expect if tonewood made a difference.

If they do test this with an actual guitar, the biggest challenge would be controlling the construction tightly enough. Basically everything you build into the guitar is going to create a site of damping. Also, you have a separate issue of neck and body. Do you make an all rosewood neck and body, even though that is not often used, or do you use maple neck and different bodies? Do you do neck through with the one wood which would likely emphasize the effect of tonewood, but is not the most common construction method? How do you glue the fretboard on-do you use the most common or the best theoretical glue? Is there a truss rod? Etc.

I would be interested to see the results of that. My hypothesis is the additional construction methods will lessen the effect of tonewood. And the most practical part is: can you make different tonewood sound like other tonewood by construction methods, pickup location, pickup construction, electronic chain, plucking technique, etc.

I have some other soap box items, but I will not get into them because they veer too far off of the topic of discussing this paper and introduce my personal bias on the subject.

 

[lomo]

I didn't pour over this for long , but..........although it looks all "scienced up", I don't see a control group, or methodology of how the listening data were obtained. 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. The spectral analysis is interesting, but....what actually matters is whether humans can reliably hear differences when blinded, not whether a waveform looks different to the eye. How many times did they listen to each sample? There are some gross %ages listed, but no actual statistical analysis (which would express the likelihood that the results were due to real differences as opposed to random chance.......you know, like in an actual scientific paper). How about showing differences between multiple pieces of the same wood, and determining how many times a listener thought the same piece was different upon listening to it twice, blindly.....= real controls. No respectable scientific journal would publish stuff like this, regardless of how much window dressing it's packaged in. Lastly, in the tone wood debates on TB, I have never seen people argue that wood "makes NO difference". What I and others have argued is that differences between multiple pieces of the same wood vary greatly, so the idea that certain wood species affect audible tone in a predictable way (for eg "ash bodies produce strong upper mids, or whatever) is vapourware. Choosing body wood with the idea that one can control/ predict how the final product will sound is the Emperor's new clothes.

 

[James Collins]

Actually, what the test shows is that there can be a marked difference between different planks of wood when the strings are plucked mechanically. It is a massive and unscientific leap to take that study and say it applies to a guitar in the same way.

Aside from the "Poop stinks" study, that is one of the shortest write-ups I've ever seen

They need significantly more research and testing to do that - but I doubt any organization would fund such testing.

If you want to test one variable (effect of wood on tone), you don't design a study with multiple variables (different plucks, uncontrollable differences in construction). I agree this is a far keep from saying: tonewood matters in the final sound of an electric guitar. Fundamentally, what they are testing is: does the wooden frame a string is anchored to affect the overtones of a vibrating string. The answer to this test was yes.

 

[TyBo]

The results showed a consistent difference in what listeners heard using different pieces of amplified wood. Whether that's predictive of those types of wood producing similar (different from one another) tones across a broad sampling may be an open question. But clearly, different pieces of wood can influence the amplified sound in an audible way.

 

[lomo]

A more reliable, scientific way to evaluate whether the contribution of wood to perceived tone is predictable would be better served by the following method:

Obtain 10 basses, 1/2 with bodies of 1 species (say alder) and 1/2 with another (say ash), all with the same neck/ fingerboard woods. Set 'em all up the same, with the same new strings. All eq flat as possible. Have the same player play the same riff ,while BLINDFOLDED, 3X on each bass while handing the player the basses in random order, which would yield 30 random samples. Then play the samples in random order to 3 blinded listeners and ask them to guess wood type for each sample-total 90 data points. When the dust settles, see how many times each wood was correctly identified and compare it statistically to 50%, the figure expected based purely on random chance. You could also increase the numbers such that different listeners could be evaluated individually, and/or more wood types. Absolute numbers matter, Small samples don't mean much.

 

[HaphAsSard]

Thanks to the OP for the article. I regard it as a step in the right direction, but, as said above, a more comprehensive study would be very expensive to conduct. One only hopes that one day a gazillionaire with an interest in music and the science thereof will forgo or postpone an orbital trip and instead finance such a study.

What the article demonstrated is those specific wood samples made an audible difference. A couple of questions come to mind:
- would different pieces of the same species have yielded the same responses? Is "species" a substantial category, or is it that no two pieces or wood sound alike? Or can two different pieces, from different species and with different looks, happen to have the same tone?
- how does wood density factor in? did the listeners pick up density rather than species, given how densities were reported to be markedly different across the samples?
EDIT: in fairness, densities of the sapele and the plywood slabs are not so far apart but, as the study shows (if I'm brows-/ahem, reading it correctly), both spectral analysis and percentage of correct answers telling them apart seem to point to density, at least in this case, not being the determining factor.

 

[TyBo]

A more reliable, scientific way to evaluate whether the contribution of wood to perceived tone is predictable would be better served by the following method:

Obtain 10 basses, 1/2 with bodies of 1 species (say alder) and 1/2 with another (say ash), all with the same neck/ fingerboard woods. Set 'em all up the same, with the same new strings. All eq flat as possible. Have the same player play the same riff ,while BLINDFOLDED, 3X on each bass while handing the player the basses in random order, which would yield 30 random samples. Then play the samples in random order to 3 blinded listeners and ask them to guess wood type for each sample-total 90 data points. When the dust settles, see how many times each wood was correctly identified and compare it statistically to 50%, the figure expected based purely on random chance. You could also increase the numbers such that different listeners could be evaluated individually, and/or more wood types. Absolute numbers matter, Small samples don't mean much.

Problem is, each bass there would have (at least subtle) differences in their construction (as can be seen trying out different examples of the same bass model). For their purposes, better to remove all those confounding factors and narrow it down as much as possible.