A device for sound analysis?

[lar]

To avoid changing strings often you could string the guitar with 3 pairs of the same string (2 E's, 2 D's, 2 A's for example). Then cut each saddle in half and tune up 1 of each of pair on each of the two saddles. I'm not sure how well a split saddle will work though. You won't be playing music like this obviously, but if your measuring only the quantitative effects this may work.

In order not to over-stress the guitar maybe use 2 pairs of strings instead of 3?

With the frequency analyzer you should be able to see if certain frequencies are accentuated more or less. But this won't tell you which saddle you like more, only that differences exist between them, and that these differences are measurable.

Sounds fun - go for it. Your results may not be what you expect, but you'll learn something (if not about saddle material then about frequency analysis).

[Bax Burgess]

I've noticed that cutting a tusq saddle in half has what may be a diminishing effect. Dropping it on a hard surface, a whole saddle has the tinkling ring - with half a saddle it's absent. Since the material is the same, I wonder if there's a minimum mass required for the 'tusq character' to show, but I haven't done a comparison to know much of anything at all.

[Bax Burgess]

Originally Posted by Alan Carruth

Welcome to the world of audio measurement.

If it's 'just' a matter of bass balance, check out the mass of the saddles. Generally speaking adding mass at the bridge reduces the response at all frequencies, but more so the higher you go. A heavier saddle can sound 'bassier', not because there's more bass, but rather there's less treble. It can be surprising sometimes how much of an effect even a couple of grams can have. What happens and what it sounds like can be very different things.

If you're more interested in the perceptual differences, then a careful comparative test can be useful. I did that several years ago when checking out the effects of string height off the top and break angle on tone. I used wire break plucks to synthesize open string strums for three different setup cases. These were played back in random pairs through headphones, and listeners were asked to rate them as 'same' or 'different'. I got about 100 readings on each possible combination. Basically, it turned out that when the only change was the break angle people were guessing: they said 'same' almost exactly half the time. When the height of the string off the top was changed they heard it as 'different' virtually every time. Note that everybody has a different idea of what constitutes 'full' or 'treble' or any other tone descriptor: the safest question is 'same' or 'different'. With that data in hand you can go looking for what it was that people heard in the spectrum, or rise and fall time, or whatever.

I see the limitation of any result I may come to identify, using my action, string type, guitar, etc.

[yellowesty]

Originally Posted by Bax Burgess

I've noticed that cutting a tusq saddle in half has what may be a diminishing effect. Dropping it on a hard surface, a whole saddle has the tinkling ring - with half a saddle it's absent. Since the material is the same, I wonder if there's a minimum mass required for the 'tusq character' to show, but I haven't done a comparison to know much of anything at all.

Bax: When you drop a Tusq saddle on a hard surface, you excite many of its natural modes of vibration. Considering the stiffness and mass distribution of a saddle, it's lowest frequency mode likely has a wavelength equal to the length of the saddle. If the saddle is 3" long, that frequency will be around 6khz (assuming that the speed of sound in Tusq is around 1.5 times the speed of sound in air). If you cut the saddle in half, the lowest resonance would become 12khz. Can you hear a short ping at 12khz?

[jwayne]

Wouldn't you also need a robotic string plucker to ensure reproducibility?

[ManyMartinMan]

Evaluating the effects of different saddles, as you've read, will be nearly impossible to do empirically due to many factors already mentioned. Here's what you need to evaluate the effects of different changes to your guitar..... What matters is what you like best.

[Bax Burgess]

I've been going over the spectograms (three plucks for each string, three seconds duration for each pluck: light, firm, strong). Easy to see is a string's intensity of attack and drop off rate, also if the hz lines are defined or blurred, or showing a staccato effect.

I'm guessing that tusq hz lines blurring together, whereas micarta's lines are defined, correlates with my hearing tusq as producing a comparative airiness...

...and that corian's quick drop off rate is in line with its clarity and seeming volume boost.

Bone, compared to micarta, overall displays a greater density of sound (especially on the 5 and 6 strings) and a stronger initial attack on strings 2 thru 6, but on 1 it shows a lot of staccato, whereas micarta has a more uniform readout. Doubtful that the bone is imperfect - it's a beautiful piece of unbleached cow bone, not a single deviation when held up to a light.

[Alan Carruth]

jwayne asked:
"Wouldn't you also need a robotic string plucker to ensure reproducibility?"

Yup. I described the 'wire break' pluck in my first post. So far as I know it's the gold standard, even though there's some variation. When I measured the break strength of the wire I've been using it was 119 grams, plus or minus 2; pretty darn close.

Joe Curtin pointed out years ago that one of the qualifications of a good player was that they can get the sound they want from any reasonable excuse for an instrument. They do this automatically, and you can't get them to not do it. That's one reason it's hard to hear the differences between instruments when you hand them to players. It's also why you need to use things like the wire break pluck.

Told'ja it was fussy....