Can You Tell The Difference??? ASH vs. ALDER

[dvh]

i don't think so at all. for one, never use heavy ash, and for two, it's never important to me to know the botanical names of every species of wood. i build guitars, i don't grow trees.

" The word "swamp" does not refer to a species of wood but actually to a "zone" where ash grows. There are different ways to describe what is actually the same tree, which by growing in different areas of the USA (north or center or south) develops almost totally different characteristics. So we have "red ash" also known as "pumpkin ash" (fraxinux prufunda) which is what we guitar makers call "swamp ash", a
tree growing in the swamp areas of Mississippi and Alabama, and other heavier species such as green ash (Fraxinus Pennsylvanica, "white ash", Black Ash (Fraxinus Nigra), which are more common in the centre and in the north of the country.

Ash tree.

All the above are actually names which are given to the same tree to distinguish the very different characteristics (weight, colour, density) the tree develops by growing in different areas and environmental conditions.

So is it the place where it grows that makes "swamp" ash so light? Possibly, but not only. Also, how and where the woodworker cutsthe blank from the tree has an importance.

CUTS: Everything else being the same, a cut in the upper part of the tree will result in a heavier blank of wood. Conversely, a piece of wood obtained from the lower part of the tree will be lighter. The rule here is that: the further north the ash tree grows, and the higher and the more internal the cut into the tree, the heavier the wood will be, and vice versa.?

I was simply pointing out that swamp ash and "hard" ash are different species. I wouldn't expect a woodworker to know the actual botanical names but to at least appreciate they are different species, not the same species growing in two different places.

That last part baffles me. Never heard that one before. There's no reason the upper wood should be heavier. More northerly, maybe. Shorter growing season, more compressed growth rings.

The point beside all this: even within similar species there will be differences. So, no broad generalizations can be made about wood and tone except that you can't make any hard and fast rules.

 

[mongo2]

Ok, so then doesn't this shatter the entire premise that Rosewood fretboard necks are more "mellow" than Maple fretboard necks?

It means the results of a test that doesn't isolate the variable being tested are ambiguous at best and a causal relationship cannot be be determined.

In other words, it's not really a test of what you seem to be attempting to prove.

 

[-=DanAtkinson=-]

Don't care much about the 'wood' discussion, but care very much about experimental design, which is basically my formal training and is a large part of my consulting business.

Just FYI, the OP did a great job isolating as many confounding variables as possible. One of the keys of 'correct' experimental design is to isolate the variable of interest by holding as much of everything else constant. So, an 'A' there... same strings, same player, same everything other than the body.

That is the good news. The bad news is, we have a sample size of 1 (actually could be considered 1 or 2, depending on how you think of the design... within or between instrument(s). As we all know, just like human beings, wood is 'organic' and there is a lot of variation among different pieces of ash and different pieces of alder (this is a HUGE point, by the way).

The only way to really determine the 'answer' is to do the same experiment the OP did, with a larger sample (lets say 50 pieces of randomly selected ash and 50 pieces of randomly selected alder) cut into the identical body shape. Then, you would statistically control (covary) for weight differences, and have another reasonably large panel of 'blind listeners' do what is called a 'triangle test'... which basically is a task for choosing which 'sample' is different among three executions (e.g., ash, ash, alder), rotated randomly. The outcome of interest is if there is a non chance identification difference between the 'different' versus 'same' correct and incorrect rates.

Most probably don't care about this, but unfortunately, a poorly designed or executed experiment often leads to worse conclusions than just 'qualitative judgement' (i.e., the opinion of luthiers who have informally compared hundreds if not thousands of similar executions of, in this case, ash versus maple bodies).

My guess is, over many many samples, there would be a small but statistically significant difference in the 'perceived tone' of ash versus alder bodies, controlled for weight, everything else being constant. However, the problem is, there will likely be individual alder bodies that sound more like the ash average than the alder average, and vice versa.

Complicated stuff

Yes, you're exactly right. Only this type of complicated, extensive testing could possibly determine (with certainty) tone variations between these wood types.

Which simply proves the point that wood tone variations are so subtle, and so nuanced, that 99% of this "tonewood" stuff is gibberish.

 

[mongo2]

That last part baffles me. Never heard that one before. There's no reason the upper wood should be heavier.

I've heard that before, though I have no way of knowing where from the tree the pieces I use are cut.

 

[-=DanAtkinson=-]

It means the results of a test that doesn't isolate the variable being tested are ambiguous at best and a causal relationship cannot be be determined.

The conventional wisdom is that a Maple fretboard neck is "snappy" and a Rosewood fretboard neck is more "mellow".

By your logic, this conventional wisdom is complete nonsense, because the rest of the neck wood has as much (if not more) impact on the tone.

 

[mongo2]

Which simply proves the point that wood tone variations are so subtle, and so nuanced, that 99% of this "tonewood" stuff is gibberish.

Sorry, that's an incorrect conclusion.

 

[-=DanAtkinson=-]

As my last couple of posts have shown, there seems to be a strange pattern from the "wood matters" naysayers. Every argument they present against the validity of these tests actualy backfires and undermines their claim.

 

[mongo2]

As my last couple of posts have shown, there seems to be a strange pattern from the "wood matters" naysayers. Every argument they present against the validity of these tests actualy backfires and undermines their claim.

LOL, nice fantasy.

 

[aborgman]

If tricking someone into answering a question the way you want them to in order to discredit them is your idea of a fair test, then you are a sociopath

So the doctors that use placebos in double blind drug testing are sociopaths?

That is an interesting viewpoint.

 

[unclejane]

Ok, so then doesn't this shatter the entire premise that Rosewood fretboard necks are more "mellow" than Maple fretboard necks?

In all seriousness, mongo's and the others' criticisms in terms of the experiment are actually well placed so I think we're kind of forced to acknowledge them.

In fact, because of the lack of controls on the other variables (which by themselves havn't been well defined yet) I'd say that, no, we actually can't yet conclude that the premise has been shown to be false.

I.e. to argue their position for a sec: it's conceivable that the wood type interacted with one of the other variables to actually _negate_ the effect it might have otherwise exhibited had those variables actually been held constant. perhaps unlikely, but definitely within the realm of possibility.

So I agree that the test, while I think it does stand as a data point in favor of the non-contribution of the wood species to tone, doesn't yet actually _prove_ anything. It's a point in evidence, but I'm forced to agree with the others that it's still a long way from a proof.

Truth is, the experiment would have to be significantly more rigorous before it could be said to be conclusive. You'd have to control for all the other variables besides wood type in a well-defined way, you'd have to define exactly what constitutes a "different tone" (i.e. well-defined object of study) and, most importantly, the result would have to be robust. I.e. just one instance of a particular result doesn't do by itself - the result has to be repeatable and easily (or even blatantly) detectable on a repeated basis.

So I thought the experiment was a good start, but at the same time we do have to honestly acknowledge the criticisms, because I think they're generally valid.

LS

 

[KJung]

Yes, you're exactly right. Only this type of complicated, extensive testing could possibly determine (with certainty) tone variations between these wood types.

Which simply proves the point that wood tone variations are so subtle, and so nuanced, that 99% of this "tonewood" stuff is gibberish.

I would agree with you here that, when choosing an individual instrument, the 'plank to plank' variability in wood seems to overwhelm the ability to 'predict' the impact of one piece of wood on one instrument. +1.

Again, didn't mean to dog your test... pretty cool actually.

The variability across individual pieces of wood both somewhat invalidates your little experiment while at the same time possibly supporting your larger point

 

[Russell Bergum]

So the doctors that use placebos in double blind drug testing are sociopaths?

That is an interesting viewpoint.

That's not a valid comparison.

In a placebo controlled double-blinded study in medicine, the researchers are required to inform the participants that they may get the real drug, OR they may get the placebo. Deception to try to prove a point is not a part of a valid placebo controlled experiment.

 

[aborgman]

Realize that people given multiple choice tend to choose only from answers that fit the preconditions stated in the test. So it shouldn't come as a shock that many people would rule-out options 3, 4 & 5. Despite the fact that people should have only used their ears, you cannot suddenly discount your preconditioning deception for analytical convenience.

Sure you can in analying results.

In a drug trial, they tell everyone they're getting the same thing... but it's a lie. Yet, if the drug is actually effective, it still works just fine on those given the real drug and not on those given the placebo. If it doesn't show greater efficacy than the placebo - it doesn't work.

Just like the color example someone brought up. If we were asking everyone to identify colored balls - no one would argue that the slightly deceptive suggestion about the data would cause a large change in results.

The fact that psychological pre-conditioning has so much effect shows that the sound effects of the two body woods are so miniscule as to be masked by plain old mental suggestion... which means the difference must be tiny relative to our perceptual threshold.

 

[aborgman]

But there is no doubt about the difference of sound between the two tone woods of my Sadowskys.

That is all.

Unless you've done double blind testing with a third party playing the basses, yes there is a doubt.

We're ALL subject to psychoacoustics.

 

[unclejane]

The fact that psychological pre-conditioning has so much effect shows that the sound effects of the two body woods are so miniscule as to be masked by plain old mental suggestion... which means the difference must be tiny relative to our perceptual threshold.

Agreed and this brings up another point, namely the definition of what the difference is. I.e. when we say tonewoods "make a difference" what do we actually mean? What kind/magnitude of difference are we actually looking for? Does a difference minor enough that it can be masked by our perceptions pass muster? Or does it have to be big enough to overwhelm our biases?

This is also something that hasn't been fully defined by the tonewood advocates and I think it's necessary. They should also be able to consistently describe the effects of the certain types of woods. I.e. maple always has X effect on the tone, ash always Y, etc..... Again, this is all moving towards a well-defined object of study....

LS

 

[aborgman]

It means the results of a test that doesn't isolate the variable being tested are ambiguous at best and a causal relationship cannot be be determined.

In other words, it's not really a test of what you seem to be attempting to prove.

It does say something about the "degree" of the variable being measured though.

If you look at black and a red ball - you'll identify them just fine, regardless of pre-conditioning or test design.

If you take a large dose of methamphetamines - it doesn't matter whether you've been told it's placebo or quaaludes - you'll still be bouncing off the walls.

If body wood had a large, unmasked effect on tone - it would be EASILY identifiable by a large majority. The fact that it isn't, bounds the degree of its effect.

 

[Russell Bergum]

Sure you can in analying results.

In a drug trial, they tell everyone they're getting the same thing... but it's a lie. Yet, if the drug is actually effective, it still works just fine on those given the real drug and not on those given the placebo. If it doesn't show greater efficacy than the placebo - it doesn't work.

Just like the color example someone brought up. If we were asking everyone to identify colored balls - no one would argue that the slightly deceptive suggestion about the data would cause a large change in results.

The fact that psychological pre-conditioning has so much effect shows that the sound effects of the two body woods are so miniscule as to be masked by plain old mental suggestion... which means the difference must be tiny relative to our perceptual threshold.

Wrong.

In a drug trial, you are told that you may get the drug, or you may get the placebo. Psychological pre-conditioning is not a part of a valid placebo controlled trial.
If the OP had said, "the clips may compare ash vs alder, or they may be 2 clips of the same wood" do you think the results would be the same? I doubt it. I bet well over half would get it right, probably more like 3/4.

 

[KJung]

+1 to the previous couple of posts. Many 'laymen' confuse 'statistical significance' (i.e., the results being 'non chance) with 'effect size' (the magnitude of the effect).

If there were little or no variance in wood, and a relatively large, consistent impact of wood species on tone, the 'effect size' part of the power equation (i.e., the part of experimental design that determines how large of a sample you need to determine a given effect size 'non chance') would make the answer to the question pretty easy to determine with a relatively simple study.

Think of it this way... if being a Republican versus a Democrat (in the US) meant exactly the same thing to each and every member of those groups (i.e., each member of each group felt exactly the same about every issue), then we would only need to talk to one Democrat and one Republican to represent each group (versus the 1000 or so sample size it takes to determine the feelings of each group versus the other on various issues because of all the variation in beliefs and opinions).

 

[BassyBill]

I'm just listening to some lovely Astrud Gilberto tunes. Her lad was a member on here a while back. Anyway, I don't think I can really tell what species of tree the instrument that the bassist was using came from.

 

[aborgman]

Agreed and this brings up another point, namely the definition of what the difference is. I.e. when we say tonewoods "make a difference" what do we actually mean? What kind/magnitude of difference are we actually looking for? Does a difference minor enough that it can be masked by our perceptions pass muster? Or does it have to be big enough to overwhelm our biases?

IMO -it has to be big enough to overwhelm our biases and be so to a majority of people LISTENING (not playing) the bass.

They should also be able to consistently describe the effects of the certain types of woods. I.e. maple always has X effect on the tone, ash always Y, etc..... Again, this is all moving towards a well-defined object of study....

Bingo. As soon as someone shows me they can identify wood types, consistently, from audio samples - I'll buy it.