Official Luthiers Forum!

The kind of A-B comparison many of us have often thought about, but here's a well-documented example. Michel Pellerin is a talented young builder in Thetford Mines, south of Québec City.


Wish I could figure out how to post a YouTube video with a screen shot! (Thanks John for the tutorial!)

Very interesting!

The video has to be in this format

Note http instead of https. The BB will not recognize anything else.

_________________
http://www.Harvestmoonguitars.com

Many thanks for that, John. I've edited my original post.

Great sounding guitar!

Reinforces my strong preference for TUSQ...

I thought the bone and Ivory sounded best but the bone was good enough to not promote the sale of ivory of any kind

They all sounded okay though.

Great sounding guitar indeed and well done video. With all the usual caveats about listening to a youtube recording on computer speakers, I also think all of them sounded good. On the first listen, the only one that stood out clearly to me was the graphite which sounded darker than the others either because of less high end or more low end and I didn't like it as much as the others. The mammoth ivory seemed to have a tiny bit more high end, but otherwise the non-graphite ones were pretty much indistinguishable to me. That was listening through my computers speakers. Listening through pretty good ear buds confirmed that the graphite sounded distinct from the others but I actually liked it because the low end sound a bit better to me, especially noticeable in the single chord section of the video. For the mammoth ivory in that same section, the middle strings sounded boxy, but I couldn't hear that in the "theme" section. If I was going to select among them for tonal difference, it would be between the graphite and the rest.

_________________
Once in a while you get shown the light in the strangest of places if you look at it right - Robert Hunter

Bone would be my first liking but the granite defiantly stood out in a good way. That was a very nice guitar and they did all sound good. That was very well done!

I am with Jay, the graphite is the only one that stood out, I thought the highs had less sparkle or sheen or some other term. I can not say for better or worse. The guitar really sound good as recorded as the player plays really well.

_________________
http://www.Harvestmoonguitars.com

It's an interesting video, and I definitely commend him for the great efforts he put in to this. If I may suggest however, it should also serve to demonstrate the challenges in doing controlled comparisons.

In the first portion, the graphite saddle definitely stood out as the most distinctly different - in fact, a bit too distinct to be comfortable in the notion that the differences heard were directly and solely due to the change in material.

Of course if you look closely, it's quite clear that the right hand is shifted a bit further to the soundhole with the graphite sample than with others, and this is a major, major, major, major influence.

He did an excellent job mind you - I am in no way at all criticizing this work as poorly done. Things like this however, are just the tip of the iceberg in challenges to maintaining controls in comparisons like this. He obviously went to great efforts to maintain the same posture, position to the microphone, aggressiveness and angle of attack, hand positions. Even for the best players with the best of intentions though, this is pretty close to an impossible challenge.

It's not just drive position, but the tiniest difference in angle of attack, how your pick or nail slips off the string, tiny difference in drive force, and so on. These confounds can be so influential as to easily and completely outweigh any changes that may be affected by the material, especially with the all of the operations involved in changing components between recording samples - even taking your hands off and putting them on again holds potential to affect consistency.

The goal of accomplishing this sort of comparison with perfect consistency in playing, is truly akin to shooting 8 arrows at 50 meters, and driving each one through the center of the last. It is simply an unrealistic demand.

There could be ways to account for this, but the scale of complication in testing increases very rapidly. One method could of course be to use a mechanical drive system (which can be surprisingly difficult to engineer as perfectly consistent, accurate to a natural finger stroke, as well as acoustically silent).

Another method is to increase the sample size, and introduce practices of blind controls. This means a lot of sampling, a lot of listening, and introduces concerns of both player and listener fatigue. One example approach may be to limit the comparison to two options at a time, let's say bone vs Tusq. You put 6 black marbles and six white marbles in a hat. The tester draws a marble, and if it's white they put a bone saddle in, if it's black then Tusq.

With the saddle installed they hand the guitar to the player, who is blindfolded. They record a sample to a backing metronome track, then repeat this process recorded on parallel tracks until the hat is empty. Now we have 6 samples of each material with natural playing, with no way for the player to have subconsciously been affected by knowledge of the variable.

Now we get a listening group to determine the most average representation of each material and eliminate notable outliers. Each listener will listen to a looped sample, and be able to click 1-6 to flip between tracks in real time. They choose the one that sounds most different, then that track is eliminated. Shuffle the tracks to a different order, then repeat the process with 5, shuffle again, and they can narrow from 4 down to the three most average. Now repeat this with 6 tracks of the other material, and do this with several listeners independently, and evaluate the results to pick out the three samples of each type that were most consistent with each other.

Now on the second listening test you could do the same real-time track selection, and do an odd-one-out survey where you play 2 or three of one type against one of another, and see how reliably listeners could identify the odd one out. Or you could use the modified Levitt Up-Down method, which is probably more precise. Here you use only two tracks at a time, randomly selected to either be of the same material or different, and the listeners are asked to determine if they are the same or different in several sequential tests.

Then you would have some data to confirm if the materials affect any significant tonal difference audible within the scope of ideally controlled variances in real playing. Then you could repeat this with other materials, and through further testing select a single example of the most representative example of each, and use those for a display of variation in tone that could be credited solely to the material change with a reasonable degree of confidence.

Sounds ridiculous, doesn't it? That's how it is though, and really just the tip of the iceberg when you want to achieve objective and reliable data. Except of course at this point you will only have tested this on one guitar, with one playing style, one string type, leaving open the distinct possibility that there may be other differences that could be revealed in different environments.

Of course I don't see anyone embarking on this kind of test anytime soon, unless you find someone willing to give a grant to fund it.

So I would give this man an A for effort and good intention, and admire the work he put in to it. In the end though, I have to walk away with near zero confidence that the results consistently and reliably reflect any qualities which could be solely and directly credited to the materials tested.

_________________
Eschew obfuscation, espouse elucidation.

I think the bottom line for this type of comparison is whether whatever tonal differences there are between the materials are even big enough to be resolved definitively. That guitar sounds great with any of those materials. The graphite was only slightly, but noticeably, different from the others which were all essentially the same to me by that test. I can't imagine that if I had that guitar with, say, the bone nut and saddle, that I would be thinking dang, I just wish I had the xxxxxx nut and saddle and I would be so much happier. It would be interesting to have included nuts and saddles made from distinctly different materials like brass, horn, ebony, or stone (onyx for example) to see if a blatantly different tone could have been produced. If yes, then you would conclude that all the materials that were tested in the video above are basically the same tonally. If not, then you could conclude that the material used doesn't really matter too much.

_________________
Once in a while you get shown the light in the strangest of places if you look at it right - Robert Hunter

There is also the issue of even if one can determine or thinks that they can determine which material sounds..... better than the others does what we hear or think that we hear on THIS guitar always translate to be true on any other guitar.....

Years ago back in around 2006 IIRC there was a thread on here about different bridge pin materials "sounding better" than others. I'm remembering this conversation because it changed my thinking AND I actually invoked the explanation provided back then the other day with a client who wanted to know what bridge pin material would improve the tone.... of his Martin.

As mentioned we shy away from any and all tonal discussions with our customers because tone is so very subjective. But I did relate what I learned back in 2006 to this client and likely left them more confused.

When I was thinking that materials such as BRW would make great "sounding" bridge pins Al C. came along as he does and set all of us straight. Al's argument was that it was not necessarily the bridge pin material that might be responsible for any perceived "difference" in tone from bridge pins but moreover the mass differences of the "change" of bridge pin material.

More specifically the bridge is located on some of the most important real estate on the entire guitar. When we change mass in that area of the pins from say pl*stic to bone it may be that this specific guitar experiences a tonal change that is audible. It also may be that the change in mass detracts from something desirable and notable for that specific player. And it also may be that no change is perceivable either as well.

Anyway this thinking applies to at least the saddle change here I would think too. Some materials are going to be more or less massive and again in an important location on the beast. With that said is it the material change that is perceived or the change in mass of the material change?

It's something to chew on for another decade I suppose.

PS: Great playing, great guitar

Good effort --- unfortunately a sample pool of "one" is rather meaningless unless the goal was to hear how the "parts on hand" sounded on the guitar being used for the test. Its possible/likely especially with the natural materials, to have different audible results when a different pair of like material components are installed.

_________________
Ken Cierp

http://www.kennethmichaelguitars.com/

This understandably casual statement of observation is mentioned several times above, but I think wording matters here. To say "the graphite was", implies confidence that the graphite caused. Rather I would say -

The sample which included graphite was the most notably and clearly different tone. There is nothing in this particular demonstration however, to show or even suggest whether the material contributed toward this 'warmer' tone, or had no significant influence at all, or perhaps even affected a change away from 'warm' but was overridden by notable variations in the player's attack.

So in spite of good effort and intention, there is nothing in this demonstration to show if, how much, or in what way any of these materials directly affect a change.

_________________
Eschew obfuscation, espouse elucidation.

As David says, this is a huge can of worms. Just a few points.

One big issue is that our senses are set up to detect small changes at the threshold of perception. We can miss fairly large changes in things like power/volume when the spectrum of the sound is unchanged, and hear very tiny changes when it comes to adding in some energy at a high frequency. It's sort of axiomatic that there are things that are easy to measure, and hard to hear, and things that are hard to measure and easy to hear.

You HAVE to control pretty closely for mass. As Hesh says, the bridge is where the rubber hits the road, so to speak, and small changes can have outsize effects.

The best way I know of to 'pluck' strings controllably is the 'wire break'. You pass a length of fine magnet wire under the string and pull it upward until the wire breaks. The actual force is the same within a percent or so, and the location and direction of the pluck are easy to control.

Violin maker Joe Curtin pointed out once that handing an instrument to a good player usually won't tell you anything about the instrument. Good players compensate automatically for the characteristics of the instrument to produce the tone they want. On the other hand, poor players can't make anything sound good, so it's no use having them do it. Recourse to mechanical means is tedious, but reliable.

So,
We make a case for how unreliable the human hand is for reproduction of movement and accept a mechanical means of testing as controllable; and I get it. However, unless a robot is going to be playing said guitar, I can't see mechanical means being valid either.
Each individual has a different touch and this thought of compensating to produce a sound that they would like to hear leaves me wondering. I guess that's why each guitar has its own voice.
..... This stuff is frustrating to read as a beginner! : (

_________________
Wake up and smell the rosewood!

It's easy to get discouraged when you began to develop a clear picture of the confounds and complexity of testing. I would much rather recognize the challenges and uncertainty however, than naively believe something as established fact, when there is in fact no reliable evidence to suggest such.

The mechanical drive methods such as the wire pluck method Alan described, are not only valid, but can be incredibly valuable. No, they do not accurately reflect the tone that any particular player will get in application, but this method is not meant to be used as a product demonstration.

The value instead, lies in the ability to control any influences outside the test variable, to offer a clear and objective appraisal of a variable's direction and degree of influence. This is more practically done by measurements and FFT analysis, rather than by ear. You of course will rarely leave with a clear tonal painting of a final embodiment. Instead you can gain valuable insight in to what does or does not affect significant changes, where the changes are focused, and incorporate that knowledge to make more intentional educated decisions with better odds of affecting more predictable results.

The blind listening tests I tend to be fond of are useful as well, but toward slightly different goals. In spite of generally requiring larger pools of samples, they still will deliver less precise quantified results than the mechanical drive approach. They are meant more as qualifying rather than quantifying tests though, which can demonstrate the existence and perceived qualities of variables in application. They are very challenging to do well though, yet incredibly easy for audiences to mistake entirely unreliable tests as valid.

The two methods - quantified measurement tests, and qualified listening tests - are meant to be complimentary to each other, and neither one fully substitutes for the other. It's the engineer's wind tunnel tests, vs driving tests on the track. Each method has its place, and both can be much more complicated to design and execute for reliable results than many may assume.

I think the best value in the type of comparison attempted above, is the opportunity to critically pick it apart, recognize the many ways in which results could be compromised, and keep these lessons in mind when viewing claims of evidence or attempting to form your own.

_________________
Eschew obfuscation, espouse elucidation.

Nice discussion. I've changed out the tusq Saddle and nut on my 1997 Martin OOO-28, to bone for me a noticeable improvement.

However I've been taking guitar lessons for couple of years from an absolute monster musician. After my first lesson my 16 year old daughter who thinks "daddy" is the best guitar player in the world, asked "What did you learn" I told her "Sweetie I'm not that good".......anyway a two years later I've learned a thing or two especially with picks, different material's, thicknesses, etc and I'm beginning to be able to get "My" sound out of any of them. I believe when your approaching a certain level of competence you can get your sound out of just about any decent instrument.

As David says, things like the wire pluck are there to reduce the number of variables, so that you can isolate the thing you're interested in. It is possible to combine mechanical plucking and listening tests, though. I did that a few years ago when I wanted to investigate the relative contribution of string height off the top and break angle to tone. I used the mechanical pluck to get consistent input into a Classical guitar with three different setups:
A)strings 11 mm off the top and a 25 degree break angle,
B) strings 11 mm off the top with a 6 degree break angle, and
C) strings 18 mm off the top with a 25 degree break angle (don't try this at home!).

Using the wire pluck I got six samples of the sound the guitar produced for each open string in each setup (so 108 samples) recorded on my computer. I made up a 'synthetic strum' for each setup condition using (iirc) about .7 seconds of sound for each string. These were played back in random pairs through headphones for a large number of listeners (mostly my long-suffering students) who were asked to say for each pair whether the sounds were 'the same' or 'different'. Asking for timbre descriptions would have been too confusing, as everybody has a different way of describing those. Basically, when the string height off the top was the same, people were guessing whether they were the same or different, but nearly everybody heard a difference when they compared the 'high' setup (case C) with either of the other two. This tallied nicely with the differences, or lack thereof, in the sound produced by the different setups. I've been trying to work up a paper on the whole thing on and off, but it's pretty complicated in the details, and really needs to be three papers. At some point, when I have the time...

Thanks David and Al for the responses.

David, I dont disagree with you and understand the signifigance of establishing baseline.
I personally find the video persuading and consistent.
Given the fact that Bakas is a professional musician, has a degree in sound engineering (honours), and a masters degree in music compostion, lends a lot of credibility to the QA of his experiment.
It would have been nice if he detailed his thoughts or findings at the end of the video.

Without even seeing or hearing a cheaper guitar we say "upgrade to a bone nut and saddle" and the implication is that it will sound better.
I could go on with other examples where acoustic faith is assumed rather than proof.

And no Im not arguing, just letting the thoughts outta my head.
gregor

_________________
Wake up and smell the rosewood!

I understand the desire to weight as many factors as possible in one's favor in the effort get a great sounding guitar, but jeeza meeza, many people in the guitar world fret so much over factors that are only a very minor influence at best on the overall sound. IMO, the video does a good job within its limitations. The guitar in the video sounds great with any of the saddle/nut combos that were compared on that guitar. I wouldn't lose any sleep at all over which one it was equipped with. It may not be a rigorously scientific study, but I have no problem concluding for me at least that worrying over bone vs. TUSQ vs. mammoth ivory vs. graphite is well beyond the point of diminishing returns and I don't believe that, in this instance, a more elaborate study would change that conclusion. For me, that's useful information so I'm glad to have seen the video.

There definitely are aspects of guitar design and construction that are worthy of in depth studies especially when testing long held lore that has never actually been supported by evidence, but doing tests with elaborate setups to attempt to isolate and demonstrate minuscule differences that can be completely swamped out by other aspects of a guitar or by player technique seems futile to me. In the world of high energy particle physics research, the greater the lengths you have to go to measure tiny tiny differences, the more likely it is that the conclusion will be profound. The harder you have to work to demonstrate a difference in a guitar, the more likely it is to be insignificant as a practical matter. (IMO)

_________________
Once in a while you get shown the light in the strangest of places if you look at it right - Robert Hunter

It seems to me there are so many aspects of building that outweigh the ever so slight differences in nut materials. The limitations for players are their abilities. A good player will not be daunted by the nut material. Perhaps a stupid example... Many years ago I brought an old Silvertone saxophone into a shop for inspection. I thought it was a crappy instrument, but the guy at the shop picked it up an started wailing on it. I couldn't believe how good it sounded. I realized that it was not the instrument but my limitations keeping me from sounding good on that instrument. Maybe not a good analogy with a guitar... but, my point is that I don't think most players are going to go "you need to change to _____ nut material..."

_________________
Steve Sollod (pronounced sorta like "Solid")
www.swiftcreekguitars.com

Yup - a never ending conversation about 6 open notes. Weird.

_________________
"Act your age, not your shoe size" - Prince

Points made on the relatively minimal influence factors like these may play are all good points. From the perspective of a repair person however, even if some component makes no difference, this information can be just as valuable as knowing what does have some effect.

The role of a repair person is often one of a trusted advisor, and a problem solver. When a client shows up with a particular problem or a change they wish to affect in an articulable way, they are relying on (and paying for) our trustworthy knowledge and skills to accomplish their goals. When they come in with a particular goal, and bring up one of the countless solutions they've seen in advertisements or discussion boards, I am supposed to know how things really work, and be capable to advise them as to if or in what way the proposed change may help.

Not all shops follow this philosophy, and often some may just go with what they've heard through marketing or anecdotal accounts without question. My problem is that if the intended goals are not achieved, the customer is often going to be less than thrilled. We want all of our customers to be thrilled.

So, I find it very important to understand these little details, so that we can confidently make realistic promises and meet or exceed customer expectations. If they come in with a problem or desire to change things in a certain way, I want to be able to reach that goal in the most direct way possible rather than embark on a journey of unknown length and destination as we resort to speculative trial and error. This means understanding as many of the little details as possible, including awareness of what you don't know, and not taking on projects unless you're sure you can guarantee the results.

So if someone comes in and wants me to replace their bone saddle with a graphite one because they want to warm up their tone and that's the impression they got from this video, I'm going to make clear that the demonstration referenced was irreparably flawed, and any conclusions they may have arrived at from it I feel are completely unreliable.

Knowing what doesn't work is just as important as knowing what does. Believing something to work based on bad, unreliable evidence can be even worse in terms of guiding you to set false expectations that may not be met in the results.

_________________
Eschew obfuscation, espouse elucidation.

Hello guys!

Somebody let me know there was this discussion for my video here and I thought to stop and say thank you for sharing it Mr Tim Mullin and thank you everybody who watched and wrote your thoughts!It seems like a creative forum here, full of ideas!
Of course as some people said above these changes applying exactly like that only to my guitar but how I am thinking is that similar differences between the materials will be to another instrument as well with different tone woods etc or using a pick or nylon strings, you ll get the same kind of darker or more transparent sound, more/less sustain etc but in a different frequency spectrum, so you can imagine, but nothing beats the experience of trying for yourself.
Just my opinion is that small piece of material on the guitar matters because it is not only about sound but feel as well and that was something I couldn't show in a video.Having more pronounced bass/mid/ or high freqs or more/less sustain really affects your playing and maybe even your compositions.
A lot of people having bone as the best default choice but for me it is hard to believe especially after this test that bone or any single material is the best choice for every guitar and every player.I think especially the players should experiment more with that and find their choice because depending the saddle/nut material you getting differences from bass/mid-high freqs to sustain and overtones.Some really nice choices for the player right there to search I think.That was the goal of the video and also to show that clearly there are differences for anyone that isn't really sure and maybe thinks to search more about it.
Also I didn't mention my opinion on the video at all because I didn't want to influence anyone.
Everybody has another taste, a sound that is looking for and another instrument, strings etc.
I see many different choices from mine above here and that is great!
Thank you so much, appreciate your time to check it!

Enjoy your day!

Thanks for your thought Sotos.
As others have mentioned, great playing and nice guitar. : )
gregor

_________________
Wake up and smell the rosewood!