Official Luthiers Forum!

How would you rate various top woods in terms of overtone production?
Please take my list (my guess) and add to it, move woods up and down etc.
Starting with the most and ending with the fewest - Thanks!

Cedar
Redwood straight
Redwood sinker
Redwood Figured
Lutz
German spruce
Engelmann
Sitka
Walnut
Koa
Carpathian
Adirondack

_________________
"An adventure is only an inconvenience rightly considered. An inconvenience is an adventure wrongly considered." G. K. Chesterton.

Good thread Steve my friend.

OK here's my thoughts:

Cedar 1
Redwood straight 2
Redwood sinker ?
Redwood Figured 3
Lutz 4
German spruce ?
Engelmann ?
Sitka 7
Walnut ?
Koa ?
Carpathian 6
Adirondack 5

Please note that I am commenting on the question as asked - overtones. Many of the top woods that I rated lower for overtones have other attributes that are desirable and if rated for these attributes would exceed the front runners for overtones.

What evidence suggests that topwood dictates overtones?

_________________
JJ
Napa, CA
http://www.DonohueGuitars.com

Are you saying it doesn't? What evidence suggest that it doesn't?

Just kidding!

I don't see that anyone suggested that it did.

Obviously maple back and sides dictate overtones also.

OK - For the purpose of this discussion, let's assume that the top is on EI Rosewood.
Assume your standard bracing scheme and someone is asking you to build a guitar with lots of overtones, and he wants you to rank all his possible tops in terms of overtones on EI rosewood.
How would you rank them?

_________________
"An adventure is only an inconvenience rightly considered. An inconvenience is an adventure wrongly considered." G. K. Chesterton.

Steve, if you haven't read it yet, Dana Bourgeois wrote what has become a "classic" on the subject:
http://web.archive.org/web/20050308152013/www.bourgeoisguitars.com/tapping_tonewoods.htm
He doesn't discuss Carpathian (perhaps not yet en vogue at the time or the lecture). The guitars I've built with it have been pretty powerful with a lot of projection and close enough to German (or European) spruce in tone.
Of course the usual caveat, as often discussed here, is that there is sometimes considerable variation within the same species. Also IMHO if you seek a tone laden with overtones a lightly braced thin top will help toward that goal, regardless of species.

_________________
Laurent Brondel
West Paris, Maine - USA
http://www.laurentbrondel.com/

Steve, you obviously have an ear for this kind of thing.

Can you describe to me what you mean though by "more overtones"? I am curious because scientifically it is within the realm of a single builder to measure this kind of thing pretty easily. I see in your profile that you are a scientist, have you done any testing on finished guitars?

Sorry I can't rate them myself, I have only used engelmann.

_________________
Ken

Here is a comparison of an OM style guitar that I made and a 1930 Gibson A style mandolin. I happened to have these two in my computer room. But you can see the difference it the overtones.

Graphs show a mandolin on top and a guitar on bottom. The X axis shows the same scale and range for comparison. the Y is relative amplitude. The way they were generated was by playing each note on each string of the instrument and then performing the fft analysis on the whole recording.

Some interesting things to note here are the groups of peaks determine the timbre of the instrument and there is a BIG obvious difference. Notice the group of peaks in the 1.0 to 1.5 hz region on the mando and the big wide peak at .25 Hz on the guitar and then another formant (group of peaks from about .3 to .6 Hz on my guitar. This is a little unusual because of the construction of my guitar is not the usual X brace. Do this on your guitars, preferably at the same time with the same mic set up and you will notice some interesting things.

It is very easy to set up for this testing using a computer and a free download of sound analysis software.

_________________
Ken

Ken,

Interesting stuff. Did you mute all of the other strings as you played the notes? I sometimes wonder what people mean by "overtones" in a guitar's sound as opposed to other strings coming into play via sympathetics.

_________________
Dave White
De Faoite Stringed Instruments
". . . the one thing a machine just can't do is give you character and personalities and sometimes that comes with flaws, but it always comes with humanity" Monty Don talking about hand weaving, "Mastercrafts", Weaving, BBC March 2010

Oh and by the way, look how good these two instruments would sound together. There would be very little overlap in resonances! And the sound together would be full.

_________________
Ken

IMHO what most people understand as overtones would be above 4K, as the last notes in the series. Wounded strings generate more overtones, whereas plain ones are more "fundamental". Every acoustic guitar has peaks in the 100/250 range, and depending on how the instrument is miked, it creates problems even in the simplest mixes. This is what most people refer as "boomy" I suppose, some dreadnaughts being particularly hard to record for this reason, unless the mikes are placed over 2ft from the guitar in a fairly neutral room.
With a spectrum analyser I think one would need to magnify what's happening between 4K and 12K to perhaps have a visualization. Most good quality analyzers can boost the peaks by up to 20db, and separate the spectrum in more or fewer bands. The more bands, the more faithful and detailed the spectrum. The more peaks, and the sharper they are, the less fundamental the tone is (a pure tone is a sinewave).

_________________
Laurent Brondel
West Paris, Maine - USA
http://www.laurentbrondel.com/

Ken's test is similar to what I do with an 'impulse spectrum'. The impulse takes away the need to play every note with the same amount of force, which could mess up the relative heights of the peaks, but the played spectrum might give a truer picture of what the instrunent sounds like. There is no perfect test.

The FFT will show the total power in a series of 'bins' with uniform frequency spacing. For example, if your sound sample is 2 seconds long, the bins will be 1/2 Hz wide. Most of the spectrum analysers I've seen in studio type setups are for 1/3 octave bands, so a lot of the high frequencies get dumped into the same bin. As usual, there's something to be said for each. The FFT is more likely to show you all of the resonant peaks, but the studio setup is closer to what you hear in terms of critical bands. I think the former is more useful for a maker.

In general the sorting of the wood types seems to follow the average damping factor of them, as one would expect. Damping refers to the amount of energy dissipated by the material per cycle as it vibrates, and is often expressed as a Q value. If a material has a Q of 50, then 1/50th of the total energy in the system is dissipated per cycle. High frequencies have more cycles per second, so they tend to lose energy faster. Thus low Q/high damping materials tend to lose high frequencies faster. Redwood and cedar tend to have much lower damping than the spruces, with some redwood samples being as low as Brazilian rosewood in that respect.

As always, there's lots more we want to learn about this stuff, and the work progresses slowly.

Can't say as I have enough experience yet to rank those, but I did notice that mahogany is not listed on there.
dave

Alan, when you do your impulse test do you produce an impluse with an audio speaker, a transducer, or just tap? And do you have strings on or off?

Interesting info Laurent about the "overtones" being perceived as those above 4K. Of course in the true definition, an overtone is any resonance above the fundamental and is usually a whole number multiple of it's frequency (though not always). I recorded a few chords and filtered out everything below 8k and above 12 K and listened to the chords. It sounded like playing behind the bridge on an archtop.

_________________
Ken

ok, here's a dumb question. How do know what you are looking at in the graphs is overtones and not fundamentals?

_________________
Jeremy Douglas

At least all that you see over 1 KHz is a bunch of overtones, the 19th fret high E string is 950 Hz or so.

_________________
Build log

I wonder also what is meant by the statement that a guitar has "overtones". I think Laurent has some inside info on that.

I don't mute the strings because I am measuring ALL of the resonances of the guitar. It is a "long time signal". One way to do it is to have a "peak hold" function and the peaks just keep adding to the spectrum as one plays the notes.

I would like to be able to do impulse testing as Alan Curruth does but I seem to get enough energy to build the equipment.

Interesting thing is that iphone has a plug in to do sound analysis using the accelorometer that is built in or the little mic which I am told is pretty flat. If anyone is interested I could look it up.

An overtone is a resonance above the fundamental that sounds when the string is plucked. So Alexandro is right.

_________________
Ken

That was simple enough

_________________
Jeremy Douglas

Ken,

Thanks. I only asked this as I think that sometimes what players call "overtones" in a guitar's sound has more to do with the "complexity" of sound that comes from sympathetic string resonance as opposed to Laurent's scientific definition of overtones. If this is the case with Steve's customer then the hunt for true overtones may not deliver the expectation. It would be interesting to repeat the analysis you did but this time muting all but the plucked string and compare the graphs - although not true "overtones" the sympathetics will be in the chart.

_________________
Dave White
De Faoite Stringed Instruments
". . . the one thing a machine just can't do is give you character and personalities and sometimes that comes with flaws, but it always comes with humanity" Monty Don talking about hand weaving, "Mastercrafts", Weaving, BBC March 2010

For a definition of overtones (or harmonics):
http://en.wikipedia.org/wiki/Harmonic_series_(music)
http://smu.edu/totw/overtone.htm
I think that what defines the tone of an instrument (beyond the envelope -attack, sustain, decay-) is the overtones content (or what "sticks out" beyond the fundamental), because there is nothing else. A "pure" tone is not very pleasant, it is a sinewave (like the telephone tone, without the distortion…).
A good exercise is to hit one note on the guitar, say an open A, eyes closed, and try to hear the harmonic series.

_________________
Laurent Brondel
West Paris, Maine - USA
http://www.laurentbrondel.com/

Just to get this straight for those who may not give a hoot, but want to be able to talk acoustics at a cocktail party, The definition of a guitar overtone is not those resonances from 8=12 K, but rather any resonance above the fundamental. And the beautiful thing about this whole wide world of acoustics is that the fundamental can be pretty weak and person can "perceive" it because the first harmonic overtone is higher. The cello and double bass take advantage of this phenomenon. So regarding what Laurent said about the resonances in the 8-12 K range is important to understand but must be transferred into a truer technical definition in the luthier's mind. I wouldn't correct the one wanting you to build a "guitar with lots of overtones, man, so I can sound like Steve Howe" but know then that the Steve Howe wannabee wants a "brilliant" guitar.

As Alan will tell us, I hope, is that brilliance is hard to achieve. It probably has a lot to do with the damping of the back and side woods. But to perceive more brilliance, if the mid and low range are not as loud relatively, then the high region becomes more pronounced. So tone down the lows to get more highs.

It would be nice to get a clear definition of guitar sound types. Similar to what Martin Schleske has put up on his excellent website on violin acoustics. http://www.schleske.de/en/our-research/handbook-violinacoustics/klangfarbe-und-resonanzprofil.html

Maybe one exists that I am not aware of.

So, enjoy that cocktail...it is better than talking about the economy anyway.

_________________
Ken

Ken, I wrote that the overtone content is most obvious over 4K, 8K is a little high and misses a lot of frequencies that we perceive as "rich". You can experiment with a high pass filter to hear what's happening up there.
Regarding bowed instruments, theirs is a sustained sound, so the harmonic content is much more "constant" so to speak.
Plucked strings tend to produce a flurry of overtones at the short moment of the attack, especially steel strings, and those decay faster than the fundamental, or this is what I hear anyway.
While it is possible to voice a guitar to enhance the middle to upper range, especially a small body, voicing to enhance a "rich" harmonic content is different.
As you point out I am sure the back and sides choice plays a big role in this as some harder woods are more reflective and have less damping, but IMHO the top choice, and the way it is voiced, is the determinant factor.

_________________
Laurent Brondel
West Paris, Maine - USA
http://www.laurentbrondel.com/

The impulse test is the easiest one to do, and requires the least 'technology'. All you need to do is hit some part of the guitar with something and record the sound on your computer for later analysis.

The frequency content of the signal will be limited by how long the hammer stays in contact. As long as the hammer is touching it damps out the vibrations of the anvil/guitar. If the hammer stays in contact for 1/1000 of a second there will be essentially no energy above 1000 Hz, and the energy will be uniform at every frequency up to half of that, 500 Hz. Thus, to see what's going on at high frequencies you need a hard hammer. The safest place to hit a guitar with a hard hammer is on the saddle, but that's OK, since that's where the strings are driving the top from anyway.

The hammer will set in motion every mode of vibration that is moving at the place it hits. No matter what you do, you can't get a see-saw going by pushing on the fulcrum. Each of the these modes will 'ring' on for a while, and put out some sound. Some of the lower modes will radiate sound all around, but most of them will be at least somewhat directional, so even if you get everything else the same you'll get a different power spectrum if you move the mic even a little.

With moderate care you can get quite reprodicible results with a simple setup. I use a hard plastic or wood ball, 5/6" or 3/4" diameter, swinging on a string or a light shaft. I like to hit the saddle between the 4th and 5th strings, to activate the 'cross dipole' mode as well as the others. I pick up the sound from 3' out in front, on the center line, with the mic positioned between the bridge and soundhole. I use a normal soundcard and 'Cool Edit' to do the recording, and an old DOS freeware program called 'FFT4WAV3' to break the signal down into it's frequency components. This program allows me to export the 'real' and 'imaginary' parts of the transform as comma delimited text files, so that I can massage them in a spreadsheet.

The hard part in all of this is figuring out what is 'room' and what is 'guitar'. It's less important if all you want to do is compare the guitars that come through your shop: just use the same setup every time and look for differences. I've recently spent some time putting together an 'anechoic closet', and am testing that out, and getting stuff going with that. Hopefully I'll be able to get some 'cleaner' data. So far it looks promising.