Acoustic Addicts -- Spectral analysis of woods and style

[Bernie Daniel]

No mandolin content!

BUT if you have not seen this SERIES on tonal analysis of guitars-- different woods and different construction formats -- you might find it a fascinating as I did.

There are several videos in this series. This is only part one.

Try to stick with this video until at least the 20 minute mark where they get in to spectral comparisons of the different guitars (and woods).

One limitation I see is they do not isolate the effects of different woods from the co-variation caused by the different construction methods.


http://www.youtube.com/watch?v=ej-2_p83mWg

[OldSausage]

"Hear that bloom"

[Bernie Daniel]

"Hear that bloom"

...and SEE that boom! :)

[michaelcj]

So a salesman, who doesn't play guitar, and a first position, meeeehhh, "guitar player" get together to make a vid to explain all there is to know about woods, style and their "sonic effect". HE Shoots..... HE,... oh well............

[Bernie Daniel]

So a salesman, who doesn't play guitar, and a first position, meeeehhh, "guitar player" get together to make a vid to explain all there is to know about woods, style and their "sonic effect". HE Shoots..... HE,... oh well............

I guess you did not watch the video. I missed the part where did they claimed to explain ALL THERE IS TO KNOW about anything?

Trash talk them based only on their professions? Guess they forgot to get your permission before posting their video and presenting their data? Oh well...

[shortymack]

Nothing wrong with the first position, it is after all, first.


Where this test seems a little week in the legs IMO is the use of different makers and body styles when comparing. They should of just stuck with one brand/style for a more even playing field to base hypothesis on IMO. I could of done without the wine thing too, beer wouldve been much much better. :) Still, it was none the less interesting.

[Bernie Daniel]

Nothing wrong with the first position, it is after all, first.


Where this test seems a little week in the legs IMO is the use of different makers and body styles when comparing. They should of just stuck with one brand/style for a more even playing field to base hypothesis on IMO. I could of done without the wine thing too, beer wouldve been much much better. :) Still, it was none the less interesting.

Thanks and yes! You are correct in my view.

That was my main critique too. That is what I meant by "not controlling co-variation". That is there are two major sources of variation in the sound waves -- like you say - both the wood AND the construction.

If you vary both at once you do not know the source of the variation.

But in the second video by the Acoustic Addicts they address this by testing only Martin guitars in an effort to control the construction component of the variance. Even then it is not a perfect solution because all Martin flat top models are not exactly the same construction but they are certainly very similar. Here it is.


http://www.youtube.com/watch?v=UHbXJ69K8b0

[padawan]

Nothing wrong with the first position, it is after all, first.


Where this test seems a little week in the legs IMO is the use of different makers and body styles when comparing. They should of just stuck with one brand/style for a more even playing field to base hypothesis on IMO. I could of done without the wine thing too, beer wouldve been much much better. :) Still, it was none the less interesting.

My thoughts exactly. I gave up on the video as soon as they brought out the second guitar and I saw the drastic difference in size/shape.

I applaud their effort but not their methods. Just like Mythbusters. ;)

[Bernie Daniel]

My thoughts exactly. I gave up on the video as soon as they brought out the second guitar and I saw the drastic difference in size/shape.

I applaud their effort but not their methods. Just like Mythbusters. ;)

Again as I noted previously (and in the OP also) there are limitations to the experiment -- that it is not a "perfect experiment".

The Acoustic Addicts attempt to address issues of co-variation that in their following videos by limiting their analysis only at Martin guitars thus attempting to hold the construction variable more constant. For example they compare the D-18 and D-28 -- these guitars differ only in back and side woods -- the top woods and the bracing are the same. All guitars with new identical string sets. So they answer your critique directly.

They also compare a 1968 D-35 (Brazilian) with a 2011 D-35 (Indian) -- so there is another solid assessment of back wood only.

Short of having specially made guitars that are "identical" except for the top wood or the back wood (not very likely) one will have be be content to do the best you can.

Sorry if you gave up too soon and did not continue on to see the possibilities of methods of instrumental analysis they were attempting to apply.

I suggest that this is potentially a step beyond just sampling someone's opinion about how a guitar sounds. They make physical measurements of the tonal spectrum.

It seems to me that having a policy of accepting no assessment of tone unless it is the perfect experiment means the experiment will never happen? You have to work with what you have available -- because an experiment is not perfect does not mean that some useful information will be obtained. :)

I suggest that you take a look at the second video maybe some of your concerns are addressed.

[foldedpath]

The Acoustic Addicts attempt to address issues of co-variation that in their following videos by limiting their analysis only at Martin guitars thus attempting to hold the construction variable more constant.

No, that's not enough. It might tell you something about Martin guitars, but not about how different tonewoods are used by other builders, and what they sound like in those other guitars.


I suggest that this is potentially a step beyond just sampling someone's opinion about how a guitar sounds. They make physical measurements of the tonal spectrum.

Nope, still not enough. Leaving aside the very large issue of individual guitars varying off the assembly line, the differences in build technique are still more important.

Notice how in that first video, the guitar player comments on how he can feel the Santa Cruz guitar vibrating against his body? That's a characteristic of those guitars, and of many others in the "small shop" and individual luthier category that are built in a more lightweight style. They have different liability exposure to warranty issues, compared to the output of Martin and Taylor, so they can afford to push the construction a bit more towards the lightweight side. Martin and Taylor can still make great-sounding guitars, at least at the higher end of their output, but you will never see them risk a lightweight, "on the edge" building style. They can't afford it. Too much volume out there.

So, cross-comparisons like this are meaningless. The build styles are too different, even ignoring body shapes. And comparisons within one builder's output like Martin are also meaningless, because you can't tell how any of these tonewoods would sound under a different building technique. Or for that matter, under a different regime of tonewood selection, which will be different for a small shop that is hand-selecting less than 100 sets a year, vs. the kind of quantity buying that Martin and Taylor do every year. What does "mahogany" or "rosewood" even mean, when the wood supplies and selection criteria are so different?

[Bernie Daniel]

No, that's not enough. It might tell you something about Martin guitars, but not about how different tonewoods are used by other builders, and what they sound like in those other guitars.



Nope, still not enough. Leaving aside the very large issue of individual guitars varying off the assembly line, the differences in build technique are still more important.

Notice how in that first video, the guitar player comments on how he can feel the Santa Cruz guitar vibrating against his body? That's a characteristic of those guitars, and of many others in the "small shop" and individual luthier category that are built in a more lightweight style. They have different liability exposure to warranty issues, compared to the output of Martin and Taylor, so they can afford to push the construction a bit more towards the lightweight side. Martin and Taylor can still make great-sounding guitars, at least at the higher end of their output, but you will never see them risk a lightweight, "on the edge" building style. They can't afford it. Too much volume out there.

So, cross-comparisons like this are meaningless. The build styles are too different, even ignoring body shapes. And comparisons within one builder's output like Martin are also meaningless, because you can't tell how any of these tonewoods would sound under a different building technique. Or for that matter, under a different regime of tonewood selection, which will be different for a small shop that is hand-selecting less than 100 sets a year, vs. the kind of quantity buying that Martin and Taylor do every year. What does "mahogany" or "rosewood" even mean, when the wood supplies and selection criteria are so different?

I think you maybe somewhat missing the point or at least part of the point. A long journey starts with a single step -- this is one of those steps.

It would foolish to try to completely solve the question of tonal differences in wood with a single experiment and that is not what they are trying to do in these videos. They are not saying that they did solve it. They are merely shedding light on the topic. They are making a start to the solution using their analytical approach.

Did you watch the video on the Martin comparisons?

I agree with you that cross guitar comparisons like video #1 (e.g., Taylor vs Santa Cruz) are kind of "meaningless" as to the question of wood effects.

In contrast, I think you are wrong about the Martin comparisons. These tests clearly are meaningful in that regard.

Now to get into the question of how the woods sound in a different construction (i.e., different makers) is no longer just addressing the woods -- but rather it addresses a more complexed question of the effects of construction on wood tonal differences .

So in that regard I think you are asking a much different question -- one that is not addressed here.

[Steve Sorensen]

Whether or not the video is good "science", the presentation of the visual displays was very interesting.

I think that the idea of starting with guitars which are overtly similar and sound moderately similar when played separately pays off nicely as the four samples are brought together. With the samples being played more and more closely together, the differences rather than the similarities come into highlight.

The use of wine varieties and goofy terms to describe the guitars is the sort of marketing make-believe that it seems instrument seller feel compelled to do. Perhaps for a viewer who doesn't play, build, or work with instruments this sort of fluff is what helps the guitars to feel like special works of art where the mystique is what makes them interesting.

Perhaps the best quote, as the differences came the focus, was the delightful epiphany, "So, Carl, you mean the craftsmanship is as important as the wood?"

I'll admit, the first thing that came into my mind was, "I sure would like to see this sort of controlled comparison for a bunch of Loar-period Gibson mandolins." Wouldn't it be interesting to have a more controlled version of this test with samples from that set of mandolins where many chords, and notes across the neck were sampled?"

Wouldn't it also be interesting to see one top-quality mandolin sampled with several different picks?

How about - Same mandolin, same pick and Grisman, Reishman, Bush and Marshall right hands?

Or - Montleone, Dudenbostle and Gilchrist Loar copies?

Or - current $1000 "The Loar", Eastman and Kentucky offerings . . . perhaps with a top quality mandolin thrown in for perspective?

While these sorts of presentations do nothing to illustrate the true playability of instruments, they do highlight, in a very simple way, that the range of choices for string instruments is stunningly broad these days and that the tools to objectively assess basic instrument traits are easily accessible.

And that's a good thing.

Steve

[jackmalonis]

This could be really cool (not that it isn't already) with some more hardcore statistical analysis. Analyses of Variance and whatnot.

[Bernie Daniel]

Whether or not the video is good "science", the presentation of the visual displays was very interesting.

I think that the idea of starting with guitars which are overtly similar and sound moderately similar when played separately pays off nicely as the four samples are brought together. With the samples being played more and more closely together, the differences rather than the similarities come into highlight.

The use of wine varieties and goofy terms to describe the guitars is the sort of marketing make-believe that it seems instrument seller feel compelled to do. Perhaps for a viewer who doesn't play, build, or work with instruments this sort of fluff is what helps the guitars to feel like special works of art where the mystique is what makes them interesting.

Perhaps the best quote, as the differences came the focus, was the delightful epiphany, "So, Carl, you mean the craftsmanship is as important as the wood?"

I'll admit, the first thing that came into my mind was, "I sure would like to see this sort of controlled comparison for a bunch of Loar-period Gibson mandolins." Wouldn't it be interesting to have a more controlled version of this test with samples from that set of mandolins where many chords, and notes across the neck were sampled?"

Wouldn't it also be interesting to see one top-quality mandolin sampled with several different picks?

How about - Same mandolin, same pick and Grisman, Reishman, Bush and Marshall right hands?

Or - Montleone, Dudenbostle and Gilchrist Loar copies?

Or - current $1000 "The Loar", Eastman and Kentucky offerings . . . perhaps with a top quality mandolin thrown in for perspective?

While these sorts of presentations do nothing to illustrate the true playability of instruments, they do highlight, in a very simple way, that the range of choices for string instruments is stunningly broad these days and that the tools to objectively assess basic instrument traits are easily accessible.

And that's a good thing.

Steve


Yes lots of interesting things could be tried. Too bad their first video was so "unfocused" in that it mingled variables (the woods both top and back with the guitar maker and construction) -- thus, impossible to tell much except that they have an interesting system of tonal analysis.

The second video where they just look at Martin guitars and do a D-18 vs. D-28 vs D-35 is more satisfying intellectually.

Then they do a 2011 D-35 (Indian rosewood) versus a 1968 D-35 (Brazilian) versus a 1998 HD-35 (Indian with scalloped braces). On these comparisons the factors (other than back woods) are more controlled. More tractable story?

[Bernie Daniel]

This could be really cool (not that it isn't already) with some more hardcore statistical analysis. Analyses of Variance and whatnot.

Yes that is always the problem -- it is fun to do a single test on a two guitars -- but very hard work to do the test on 20 D-18's versus 20 D-28's so you could run an ANOVA on the frequencies and their magnitudes for example.

Of course when you do that bigger experiment you might end up finding that the intra-variance is as large as the inter-variance and then you are back to nowhere! LOL.