Comparing one model guitar to another

[Bluedog52]

I see a lot of opinions and videos of comparisons of one model guitar to another, and also from one maker to another maker. The biggest thing that throws up red flags in most of the comparisons I read, are the age and type of stings begin used in the comparisons. For example, if you are looking at opinions, or even video comparisons of a Martin XYZ to a Martin XYZV, rarely do you know for sure if the stings used in the comparison were put on and played the same amount of time. I can think of exactly one video where they did it right. This can make a huge difference. So if you're in a guitar store and pick up one guitar and compare it to another, and you're a serious buyer, you might ask the manager to whip on some new strings on the two you're serious about. If you're looking at 4-5 models, I don't think I'd bother the manager about that. Narrow it down and have the cash in hand.

[charles Tauber]

Sure, new strings are nice. But a great guitar still sounds great with not-new strings and a lousy guitar still sounds lousy even with new strings. Chances are, that most of the time you'll play any given guitar with not-new strings anyway, so you're just hearing what it will sound like should you take it home and live with it: you are assessing what it sound like "on average" rather than only when it has new strings.

There are lots of attributes for which to listen that one can still assess even if the strings aren't new.

If the subtleties for which you are listening can't be assessed without brand new strings, then you shouldn't be making assessments from on-line recordings anyway: you'll want to play the instrument yourself and listen to it live the way you play it.

[geordie]

As Charles says, with the restrictions of the internet audio AND what you, the end listener is monitoring on accepted there are other aspects to consider viz, type of microphones and placement and room characteristics to be considered.
Add to that the maturity (or lack of it) of the instrument - is it played in / opened up, has the sound board material been lying seasoning for ten or twenty years or is it 'fresh' timber that is on the guitar ?????
and so on.
Ad up all these and the sum shouts - ya need ta play it.

[Bluedog52]

Honestly, I was not really talking about comparing lousy guitars. I was talking about comparisons from say, something like a Martin HD-28V to a Martin D-28 Marquis. They are both great guitars. But if you just put new stings on one of them, and the stings on the other are a month old, you can't really acurately compare them, even if the guitar with the month old strings still sounds good.

Here's an example of a pretty good comparison:

At about 2:40 is where the comparison starts:

http://www.youtube.com/watch?v=BQZ3npQ5wTo

[Bluedog52]

I absolutely agree that you need to play the guitars. BUT, lets say you're looking for info on just which guitars you even WANT to play. Would be helpful if the information out there were accurate to beging with, so that you can make an informed decision on just which guitars you might want to try. Does that make sense?

[Bluedog52]

Here's a useful comparison if you just happen to be looking at these two guitars:

http://www.youtube.com/watch?v=RaRwIs3DbY4

[charles Tauber]

Originally Posted by Bluedog52

Honestly, I was not really talking about comparing lousy guitars. I was talking about comparisons from say, something like a Martin HD-28V to a Martin D-28 Marquis. They are both great guitars.

And, therein lies the problem. Some Martin HD-28V's will be great instruments, some will be lousy. Some Martin D-28 Marquis will be great instruments, some will be lousy. That's true of many/most makes and models. There are many "average" instruments within any one make and model. There are some that are duds and some that are great. It's a Bell curve.

Compare a "great" HD-28V to a "dud" Martin D-28 Marquis and what have you learned? That the HD-28V is great and the D-28 Marquis isn't? Based on that sample space of 1, should one rule-out a D-28 Marquis? Whether the strings are new or old won't radically alter that.

If you want a really good one, rather than an average one or a "dud", statistically speaking for many makers, you'll either need to be lucky or need to try a bunch of any one make or model. And, we can't all have the above-average ones of any one make and model - its statistically impossible.

Originally Posted by Bluedog52

Would be helpful if the information out there were accurate to beging with, so that you can make an informed decision on just which guitars you might want to try. Does that make sense?

Not to me. It isn't "an informed decision". It's a preconceived bias that "this guitar sounds like this and that one sounds like that" based on what is often too small a sampling from which to draw much sensible conclusion. Really good instruments defy that preconception. All an A/B comparison will tell you is, "I like A better than B". It doesn't tell you that you like all A's better than all B's: it just tells you about the two individual, single instruments that were compared. Based upon that one comparison it doesn't seem "sensible" to me to extrapolate that to "I like all A's better than all B's".

<RANT ON>
By and large, guitarists get way, WAY, WAY too hung up on the effect of this or that on sound. They get buried in the details, most of which aren't really relevant. What it comes down to is quality of sound and quality of playability (including intonation). The rest is noise, unless one is buying a work of art that is playable, which is a somewhat different animal. It doesn't matter how the maker arrived at that sound - what bracing, what top material, what back material, what radius this or that... - the sound - and playability - is there for you or it isn't. In my opinion, how the maker arrived at that probably ought to be an idle curiosity for most players, not something to spend hours and hours agonizing over, as all too many do.

It seems that many guitarists buy based upon cosmetics - it looks like this or that - or based on specification - "Oh, it's made of Brazilian, it must sound great". And, of course, many buy based on the brand name.


Instead, go play lots of guitars. You'll know within a minute or two of playing which individual instruments you do or don't like. If you can't distinguish between them, then it won't matter all that much what you buy, unless you want to own a specific brand, model or specification for the purpose of bragging-rights.

Personally, unless it is a very rare, unique instrument, I'd never even consider buying a guitar on-line or based on anything on-line. I need to hear it and feel it first-hand. But, that's just me.
<RANT OFF>

[Bluedog52]

Originally Posted by charles Tauber

...Some Martin HD-28V's will be great instruments, some will be lousy. Some Martin D-28 Marquis will be great instruments, some will be lousy. That's true of many/most makes and models. There are many "average" instruments within any one make and model.

When you say "some" and "many", do you have some verifiable percentages, like 1% or 30%, or something like that, that we can look up?

[charles Tauber]

Originally Posted by Bluedog52

When you say "some" and "many", do you have some verifiable percentages, like 1% or 30%, or something like that, that we can look up?

Just experience over the past 40 years of playing and 30+ years of making and repairing guitars.

Go on a "road trip". Make it your goal to play 10 of the same make and model. Then you tell me how many of those fell into "great", "average" and "not so much". If one is not very discerning, "they're all great". Statistics suggest that about 70% will be "average" and maybe 15% great and 15% "not so much". That's probably not far off my experience. (Better makers sift the mean upwards but still, generally, have some that are better than others and some that are worse than others.)

[Bluedog52]

Originally Posted by charles Tauber

J Statistics suggest that about 70% will be "average" and maybe 15% great and 15% "not so much". That's probably not far off my experience.

Where can I find these statistics?

[charles Tauber]

Originally Posted by Bluedog52

Where can I find these statistics?

Pick up any introductory book on statistics. The most basic - and one of the most common distributions of data - is the "Bell Curve" ("normal" distribution).

A quick on-line search finds this illustrated introduction: http://www.mathsisfun.com/data/stand...tribution.html



Assuming guitar "quality" (sound, playability, whatever you chose to measure) is a symetric normal distribution, about 70% of the guitars will be within one standard deviation from the mean, leaving about 15% each beyond that and below that. It has been my experience that guitar "quality" more or less falls into that.

This is standard stuff in industrial manufacturing facilities where it is part of statistical process control. http://en.wikipedia.org/wiki/Statist...rocess_control

[Glennwillow]

I do think having relatively new strings on guitars for comparison is helpful. But this also assumes that the new strings on the guitar are the ones the player would choose for that particular guitar. If you are comparing two new Martin dreadnoughts, each with Martin's recommended Martin SP Lifespan Phosphor Bronze Medium Gauge (MSP7200), having the same strings does provide you a reference. But what if you prefer 80/20 strings on a rosewood dreadnought?

No matter what you do, the player is going to be making some reasonable assumptions and guesses. Having a lot of experience and playing a lot of guitars does help remove more of the risk from the evaluation process. I think this is what Charles Tauber is getting at.

- Glenn

[Bluedog52]

Originally Posted by charles Tauber

Assuming guitar "quality" (sound, playability, whatever you chose to measure) is a symetric normal distribution, about 70% of the guitars will be within one standard deviation from the mean, leaving about 15% each beyond that and below that. It has been my experience that guitar "quality" more or less falls into that.

This is standard stuff in industrial manufacturing facilities where it is part of statistical process control. http://en.wikipedia.org/wiki/Statist...rocess_control

I'd really be more interested in seeing actual statistical data, done by industry professionals, rather than your personal statistical analysis. Nothing personal. Also, I don't have a degree in stats, but I do have a masters in finance and business, and have about 9 very painful hours of stats under my belt. :-) So if you really want to conduct some statistical analysis, then you're are going to need a lot of prefessionally gathered data to enter into your models. Which brings me back around, if you know where to get that data, let me know. Sweeping generalizations/opinion, are fun, but in the long run, not very useful.

[Bluedog52]

Originally Posted by Glennwillow

I do think having relatively new strings on guitars for comparison is helpful. But this also assumes that the new strings on the guitar are the ones the player would choose for that particular guitar. If you are comparing two new Martin dreadnoughts, each with Martin's recommended Martin SP Lifespan Phosphor Bronze Medium Gauge (MSP7200), having the same strings does provide you a reference. But what if you prefer 80/20 strings on a rosewood dreadnought?

No matter what you do, the player is going to be making some reasonable assumptions and guesses. Having a lot of experience and playing a lot of guitars does help remove more of the risk from the evaluation process. I think this is what Charles Tauber is getting at.

- Glenn

I think the point I was trying to make has been lost, somehow. You actually touched on it exactly when you mentioned a point of reference. That's really all I'm talking about. An equal point of reference. In most cases, strings are the engine that drives the guitar. So if you have the same strings, that were put on new, on two martins you're comparing, even if they are not what you would normally use, then you have something reasonable to compare. On the flip side, if you have a Taylor that you're comparing to a Martin, then it is reasonable to compare them twice. Once with identical strings, and once again with what the manufacture suggests. But there are many many variable that can give bad results in a comparison. Just sayin!

[charles Tauber]

Originally Posted by Bluedog52

Sweeping generalizations/opinion, are fun, but in the long run, not very useful.

You're right.

I have no hard statistical data on the distribution of sound quality of guitars.

Just as you say, it all boils down to ensuring one has new strings on the guitars one compares.