"Wood and tone" experiment

[Josh Emmons]

So I post this out of sheer curiosity, and the desire to know just how much woods actually have to do with the tone of a solid body bass. let me start off by saying, I am not asking for opinion, I asking for help in how to test it. But I digress, every forum post on this topic I have ever read has ended in some variant of a "flame war", with both sides interjecting some opinionated factually inadequate statement. I also read a lot of posts saying how there has been no scientific test to back the tone wood hypothesis up. This brings me to the point of this post- I feel it would be interesting to approach the debate using a less bias viewpoint. It would be amazing to create an experiment to test how much woods have to do with tone. So, lets begin...

Using a logical approach, it seems it would be possible to construct an experiment to test the variables of the problem. As per the rather generic scientific method, I've just asked the question.

Hypothesis:

Based on many years of playing bass, I suspect that woods would have an affect on the tone of the instrument. I suspect that the differing densities of woods would be the primary culprit. Having stated this, I understand that there are a lot of factors that make a bass, a bass- I.E. strings, scale, pickup placement, electronics, ecc. I find that even playing identical models of instruments produce obvious tonal variants. Granted that the aspect of strings, and pickup hight can also play a role in the tonal differences. But I feel it isn't quite all left to that. It seems that the nuances in the densities of the wood could be part of the cause for these differences in tone. Also, don't take my word thus far as anything completely factual. So far I have stated my opinion. But this is my hypothesis after all, and that is what a hypothesis is: an educated guess.

Experimentation:

Now this is to be one of the hardest stages: testing my hypothesis. Im not entirely sure how I may go about testing it... But this is why I've posted this thread, after all. There must be several factors in this experiment. The first being consistency. This includes the ability for this experiment to be retested, and the ability to attain a controllable testing method that minimizes error. I will test how sound, (the vibrations) of the strings affects the wood, and how it changes how the string is vibrating. Secondly, the testing of different materials with the same specifications. I will have several different woods, metals, and even some others as test fodder. There will be one bass pickup with identical bridge, string, tuner, and electronic connections. The pickup will be identical heights, and the placement will be identical on the whole of the set up. (see included diagram below). But this is the tedious part: I do not want to just use my ear, as this is completely subjective and hardly accurate. There must be a way to accurately rate what is happening. IE an electronic ear that can measure the differences in the materials.

Yes I understand that this may be a bit ambitious, however I feel it could be feasible to accomplish. Secondly, this I hope will not turn into a flame war, that is not the goal of this post. Also, To who ever might say why knowing this matters, I say: The goal is to learn, and when seeking light on a subject that hasn't reached any type of conclusion, I see it as "worth it".

Lastly, Im not sure how to proceed experimenting, help is something I need...

How would you guys go about testing this idea?

 

[cstar]

I'm a tonewood advocate and I've thought about this before. I thought about doing an engineering project on it when I was studying that.

I think you need to learn something about sound; how it's produced, how it is conducted, how it is affected. How do the different wood densities affect some specific, measurable pitch output?

That's the thing though... usually with engineering hypotheses you're looking to prove one very specific, hard-fact kind of thing. "Does instrument wood affect tone?" is a very open ended question. In what way are you suggesting it affects tone? How can that be measured? These are all very difficult issues in this experiment, and probably the reason why no one has attempted this.

I think what you need to do is this: have two basses of the same design (perhaps j or p, whatever as long as they match) which different body or neck woods, but all other elements the same (as much as possible obviously). Any differences between the basses beyond the one variable (say, body wood) could throw your results off (theoretically). It's a little hard to control all those variables, particularly because to sets of Fender custom shop jazz bass pickups could be slightly different in a number of ways, and two maple necks are made from two different pieces of wood which may be slightly different in the way they resonate or transfer vibration, etc...

But, with two basses as identical as possible, with one defined variable (say body wood), if you put some sort of vibrational input into them (say you strike the low e string on both basses), and you have some way of measuring some kind of vibrational output from the body (like if you had some kind of device that could measure or map the frequencies the bodies were playing back after you struck the e string on both), then you'd have some hard factual parameter to compare. On this parameter, you could probably deduct some conclusions about differences in "tone" as caused by body wood. But, you'd have to find a hard, definitive way of relating your outputs to "tone".

Edit: I think the clearest relationship between vibrational output (which we are assuming depends on body wood) and tone would be to prove that different body woods are different in bass, mids, treble, etc. I feel like that would be RELATIVELY easy to measure. It should be pretty easy to show that say, perhaps, ash resonates better/ more at higher frequencies, whereas a mahogany resonates better/ more at lower frequencies, and that because of this ash sounds "bright" (trebly) and mahogany sounds "dark" (bassy).

I hopes that makes some sense, I'm sure I explained it terribly.

 

[Lonnybass]

Why don't you just go on YouTube and listen to a bunch of MTD 535 demos? All have the same custom preamp and pickups, and differ in the woods. Lots of examples already out there to compare.

Lonnybass

 

[wneff]

Actually, it is not only the sound, but also the "feel" of an instrument that counts. What we want to know: Does the sound and feel of an instrumnent depends on:
- Wood type
- Wood density
- Simply the weight

How about this DOE:
- Use two different woods that are presumably much different:
Maybe Ash and Basswood.
- Select these woods such that they have different densities - one high, one low
- Build the basses with different body thicknesses so that the the weight of a "thick" bass with a less dense wood is the same as that of the "thin" body with a dense wood.

Total layout like this:
Bass # Wood Density Weight* (adjust body thickness)
--------------------------------------------------------------------
1 Ash low low
2 Ash low high
3 Ash high low
4 Ash high high
5 Basswood low low
6 Basswood low high
7 Basswood high low
8 Basswood high high

Output:
- Sustain
- "feel"
- .......what can we get a number off....attack time? Ratio first to higher harmnics?

Use Anova to analyze this and see what happens

 

[Hopkins]

This is a very good video as far as body wood is concerned.

 

[FBB Custom]

Make your hypotheses more specific and you'll find them easier to answer. Why not start with an easy one: "can people reliably tell the difference between two of your test instruments" made from a different wood?" Make a bunch of recordings, get a bunch of subjects, and make them say whether randomized pairs of recordings come from the same instrument or different.

You can move on to more sophisticated experiments later, but let me add my advice on your test instrument. I would not make it a 2" thick beam, I would try to get closer to the geometry of an actual bass. Why? Because some people whose opinions I respect believe that some of the character of an instrument comes from the way in which the neck flexes under tension and during use. So, you will want your instrument to resemble a real bass in that way. Also, I'd put 4 strings on it for the same reason and also so you can get simple recordings (where the strings are plucked mechanically) consisting of more than one note.

 

[bucephylus]

For the lay bass player, the evidence is as you have stated: an experienced player can tell the difference in sound between two instruments having the same components (pups, pots, circuitry etc) but different woods.

If you wanted to see data describing this, you would need to invest 7k$ to 10k$ in an instrument called a spectrum analyzer. There may be less expensive alternatives now for the PC using various A/D converters; but, I am not current with those. I have not seen any such spectrum analysis for stringed instruments; but, it could be out there.

 

[Beej]

You're genuinely asking so I'm willing to genuinely help. The scientific method generally tests the null rather than for the presence of a difference. Reading up on null hypothesis should help you to formulate your hypothesis. I agree it needs to be more specific than what you have up there. Generally as well, I really recommend using the simplest, cleanest language you can accurately describe your problem in. The flowerier it gets, the less clarity you will have in determining your subject, method and results.

Are you trying to measure whether there is any difference between woods, or measure the magnitude of the supposed difference? Or want to determine whether there is a predictable difference?

Once you have your question sorted out, I'm more than happy to help with experimental design. There are many variables to control for, but many ways to do it...

 

[bucephylus]

The scientific method generally tests the null rather than for the presence of a difference. Reading up on null hypothesis should help you to formulate your hypothesis.

What this means is that if you want to study the effect, you would hypothesize that there is no effect, then go make measurements, look at the results, and determine whether or not the results disprove the null hypothesis of there being no effect.

Which sounds like a very complicated way of saying that you have to measure something that shows a significant difference. The reason for the Scientific Method is that if you propose that there is a difference, observing any difference does not prove that there will always be a difference. It's a logic thing. I'm not sure how applicable it is to this question.

 

[pilotjones]

Bouncing back what's been stated as criticism/help, and adding a bit:

Problem definition

- Like Beej said. Must be clear, fully defined
- If your hypothesis is that the result (whatever that is) is a function of the density, then you need to calculate the density of the actual boards, not just published numbers for the species. For a geometrically simple board, this is easily done.
- Also, to prove that a correlation is due only to density and not to any other of the many differing properties of the different species, you must test samples that are different species but are the same density, or samples that are the same species but differing densities, or both.
- Analyzing content ratios of high/mid/low is not useful in the absence of

- frequency content vs. time analysis
- justification for the arbitrary divisions, stemming from some kind of study that legitimately links the divisions to a psychologically useful premise - otherwise it's just math without meaning, statistics without significance​

Experimental setup

- That 2 x 6 x 24 board may be difficult to obtain in every wood you want to test
- it is not representative of the instrument (bass guitar) about you which to gain insight
- Plucking machine is a must. This gives you a consistent mechanical input pulse, and allows you to use raw sound file results, eliminating the normalization that others have done in a recent test, which invalidated any results
- if your problem definition is anything in the direction of "quality of effect on musicality" then you can't just test a ~20" scale .045, it would have to be a range of bass-like strings and pitches. If your problem definition is simply "is there any discernible difference" then maybe the one string setup is OK
- you need to put a nut in there. A string vibrating directly off a tuner post is both unrealistic, and a needless vibrational/mechanical complication​

 

[Deep Cat]

I've put more thought into this conundrum than is is reasonable. The trick is to make your test measureable and repeatable, and to take as broad a sampling as humanly possible.

The variables of this kind of experiment boggle my mind. First you have to find a way to eliminate the player from the the test. Then rather than say "Wood Matters" Or "Wood Doesn't" you kind of have to figure our what actually does contribute to the tone, and in what degrees.

Imagine if you will that ash vs alder matters, but only if you are using nickel rounds and a bent tin bridge with medium high frets. Does it still matter? Maybe.

And what if the resonant properties of the woods only work if you are using a J with the bass rolled off and the volume at 9/10. Does it still matter? Maybe.

I really think that wood matters, but even the best luthiers will tell you with all of their years experience and tap wood tests that they can't predict what will make one guitar magical and another midling at best when it gets assembled and plugged in.

There is so much variation in wood, even in the same species and the same board that it becomes impossible to predict, based upon presumptions of a wood's resonant characteristics, whether or not the board will actually manifest those in an instrument.

So... If I were to undertake this herculean feat, I'd rig a "plucking" device that simulated a variety of finger and plectum strikes by angle and striking pressure. Then I'd build a "bassline" bass with a passive single coil, and an active soapbar pickup rigged to separate outputs. To round out the "baseline" I'd include a piezo pickup with separate crystals for each string on a third output.

I'd want to go direct with all three channels into a PC recording program that'll let me go 3d with EQ. Basically the point of this is to determine the differences between players in as broad a sampling as I could get and the plucking machine. Then I have a variety of options, such as fine tuning to make it more like the aggregate players, or not.

Once the variables of the plucking machine are known I could begin to test identical basses with the same setups and the same woods and the same electronics to get as broad a sampling as possible. 50 bodies, same neck. Same body, fifty necks. That sort of thing. And on and on.

 

[Duplo42]

First, an opinion.

I am advocate of bit paradox thinking when it comes to wood and tone on electric instrument: Bodywood does affect tone of electric instrument importantly, but its not that important at all.

EDIT: deleted huge chunk cause others said very similar thing

You can do scientific test on how much tone wood affects, but in a long run, it will prove not much. So many factors determine crucially on what bass guitar is, so many.

Thus said: DO IT!! : D I'm always for experimenting and doing stuff that enrich our lives!

I propose same thing that was already said: bass with top notch stuff and just change the bodies. No finish on bodies. Record straight into recorder. Use robotic hand to play notes to be consistent in volume etc.
Once you have recording, spectrum analyze it.

And as many would agree - I have couple of instruments made of plywood body and thick finish and they resonate as best tonewood!

Cheers

EDIT: Btw - does any luthier makes basses outta PLYWOOD?? I want one!

 

[Duplo42]

Which sounds like a very complicated way of saying that you have to measure something that shows a significant difference. The reason for the Scientific Method is that if you propose that there is a difference, observing any difference does not prove that there will always be a difference. It's a logic thing. I'm not sure how applicable it is to this question.

Hmm how about this test, I believe this would show much.

No body or neck - just strings, nut, bridge, tuners and pickup - positioned EXACTLY the same as on bass. This would determine what sound you get with no wood (doh...).

Then use same hardware on bass with no neck. This will determine the tone of body.

Then add neck.

Then play around with various types of wood.

No sure how to mount hardware onto nothing thou, lol - maybe thinnest metal/whatever rods/spikes as possible?

And play only open notes since theoretically any given number of tunes can require just open notes. This will eliminate task of putting frets on no-wood bass.

 

[Hopkins]

This thread is making my head hurt.

 

[Beej]

Full disclosure:
I'm a tonewood agnostic. I'm also a research scientist.

So I post this out of sheer curiosity, and the desire to know just how much woods actually have to do with the tone of a solid body bass....... It would be amazing to create an experiment to test how much woods have to do with tone.... I suspect that woods would have an affect on the tone of the instrument....

From these few comments, it looks like we should be trying to create an experiment that will have a hypothesis sort of like: "solid body instruments utilizing different wood species will not impact the sound produced such that a listener will be able to determine a difference between different species". I know I said to keep it simple, but this is a complicated topic. Correct me if I'm way off on your intentions here.

If you can show that people can tell a difference beyond the probability of chance, you can argue that the wood species used creates a difference. You can't argue it's predictable, or other inferences from the experiment, but this would at least be good enough to demonstrate what we intuitively already know - that is, that there is "some" effect of the wood on tone. For the record, I think much of the tonewood argument is around the "magnitude" of the difference, and the "predictability" of the difference imparted by different species.

Talking about the method itself used to test the hypothesis will give us information on what kinds of variables need to be controlled in order to single out the wood species, and how to exercise those controls.

I understand that there are a lot of factors that make a bass, a bass- I.E.

• strings,
• scale,
• pickup placement,
• electronics,
• I find that even playing identical models of instruments produce obvious tonal variants.
• pickup hight
• nuances in the densities of the wood could be part of the cause for these differences in tone.

These are all variables that will need to be controlled - controlling all of them should be possible, although some may be complicated.

Experimentation:
Now this is to be one of the hardest stages: testing my hypothesis. Im not entirely sure how I may go about testing it... But this is why I've posted this thread, after all. There must be several factors in this experiment. The first being consistency. This includes the ability for this experiment to be retested, and the ability to attain a controllable testing method that minimizes error.

Replicability is good.

I will test how sound, (the vibrations) of the strings affects the wood, and how it changes how the string is vibrating.

I think this would be quite difficult to do, and I don't know if there is any face validity to this measurement, if its operationally definable. It could probably be eliminated from the experimental design as its impact on the hypothesis is not direct (can people tell the difference).

Secondly, the testing of different materials with the same specifications. I will have several different woods, metals, and even some others as test fodder.

These are your experimental variables. They are the ones you manipulate to see if there is an effect.

There will be one bass pickup with identical bridge, string, tuner, and electronic connections.

It must be the same one, you cannot use a different pickup, bridge, etc, even if they are from the same manufacturer with the same tolerances - it introduces too much error. These are part of your variables that require close control.

The pickup will be identical heights, and the placement will be identical on the whole of the set up. (see included diagram below).

They must be set identically, as closely measured as possible. If screws or bolts are used, they must be applied with a torque wrench.

But this is the tedious part: I do not want to just use my ear, as this is completely subjective and hardly accurate. There must be a way to accurately rate what is happening. IE an electronic ear that can measure the differences in the materials.

This too could be eliminated from the experimental design as it would not impact the hypothesis directly (although it would probably be interesting collateral data).

If you think some of what is above sounds reasonable, we can look at methods for gathering and analysing the data...

 

[Luthier Atlanta]

Subscribing...............

 

[Big B.]

This thread is making my head hurt.

Just wait till the discussions about the "plucking machine" gets into full swing.

 

[NAVET]

This kind of project is great to talk about but would require several disciplines to achieve. I would guess you would need:

1) Mechanical and/or electrical engineers to design and build the plucking machine. Maybe even a computer software programmer type to ensure the computer code for "Paul the plucker" was the same each and every time. If you went with a computer controlled machine that is.

2) Noise or sound analysts of some type to ensure the sound measurement and equipment needed is correct.

3) Bass or stringed instrument musicians to help the others in the wood selection, pup setup, string selection, etc.

This is no small task in any way. Doable for sure but not small. It would be a great project for a college engineering department to undertake if they could partner up with a few bass, string and pup manufacturers.

 

[SirRolfe]

I just wanted to say that to conduct an experiment like this with as few confounding variables as possible will definitely require a good amount of money and resources. It's absolutely possible, just keep in mind the expenses involved. Skimping on any part of the experimental setup would weaken any conclusions that can be made. Definitely something more suited for a college engineering/music department or corporation, but hey, anything's possible.

 

[Bassamatic]

For such testing to be of value, all variables EXCEPT the wood must be completely eliminated.
Same hardware
Same Strings
Same tension
Same plucking (machine is a must)

A major concern is that there are so many variables in the wood itself.
Density
Grain tightness
moisture content
the EXACT way it was sawed

These are all going to vary a lot from one piece of the same wood to another. You can't use just one sample of each. For such testing to be valid, you would have to use MANY pieces of each wood - maybe even a 100 if there is enough variation between the wood samples.

Your first test would be between samples of the SAME wood to see how much tonal variation there is, and then use the proper statistical formulas to determine how many samples are required.

Finally - what is your evaluation method? Your ear? That is the biggest variable of all, unless you do multiple blind tests where you don't know what wood is being tested and the tests are mixed several times randomly. The judging should probably also be spread out among many people.

You might also want to use a proper acoustic analyzer that can produce something like a 3-D "waterfall" plot of multiple frequencies over time to have something to quantify.

Of course there is also the issue of finish. Guitars are finished. The finish MIGHT have some different effect on different woods. Your samples would need to be tested with a proper finish, and probably without a finish just to determine if it has any effect. And, of course, there are variations in the finish material and thickness that might affect the sound as well.

Having performed and planned many such tests in the past, these are the few things that come to mind to make such a test even remotely valid.

It's a nice idea, but it will take a lot of work to come to a valid conclusion.