What does your guitar REALLY sound like? An experiment.

[lpdeluxe]

Quite apart from all the rational, scientific explanations, I have learned over the years that an electric guitar is only a viable musical instrument when it's plugged into an amp, or equivalent.

Until recently, I owned what I considered one of the very best sounding electrics around, a '70 Les Paul Deluxe. Alas, the tone didn't translate when I tried to recording it DI. I learned to use a distortion box or POD or whatever to emulate the coloration that amp lends to the sound.

While I could NEVER get a good DI sound, I could easily get something musically useful with almost any stomp box or whatever.

This led me to believe that there's absolutely no point in debating the "pure" sound of an electric guitar: it's the combination of guitar and non-linear amplifier that makes the sound we want to hear.

Consequently, the sound of the guitar by itself is, while not unimportant, but of no use.

 

[c7sus]

Really, I just wanna bend the string like reen toon teen toon teen too tee too teeeeeeeeeeennnnnnnnnnnn!

 

[thebigcheese]

Quite apart from all the rational, scientific explanations, I have learned over the years that an electric guitar is only a viable musical instrument when it's plugged into an amp, or equivalent.

Until recently, I owned what I considered one of the very best sounding electrics around, a '70 Les Paul Deluxe. Alas, the tone didn't translate when I tried to recording it DI. I learned to use a distortion box or POD or whatever to emulate the coloration that amp lends to the sound.

While I could NEVER get a good DI sound, I could easily get something musically useful with almost any stomp box or whatever.

This led me to believe that there's absolutely no point in debating the "pure" sound of an electric guitar: it's the combination of guitar and non-linear amplifier that makes the sound we want to hear.

Consequently, the sound of the guitar by itself is, while not unimportant, but of no use.

Yeah, but lets say you know what the "pure" sound of your guitar is. If you could compare that to the "pure" sound of other guitars that you were considering purchasing, don't you think that would be useful? Obviously it's still best to go out and try out guitars, but you might be able to narrow down what you are looking for. I guess, as others have mentioned, in order for a true database like that to exist, we would have to have some standards in place. But even so far, we can see that my guitar, for instance, is different from Cristoph's guitar, and the differences are at least somewhat quantifiable.

So here's a question for Muttley/Mshilarious: Let's say the same person was to play the same tune on several guitars, all strung with the same strings (well, not literally the same set, but same brand/gauge) and run that all into the same interface with all the same settings for each recording. Would there be any other things to think about in terms of standards for comparison? Maybe it's not that useful to have a whole bunch of people contributing, but if, say, I was to put together a database where I keep everything except the guitar the same, that could create a useful comparison database, right?

 

[kcearl]

Quite apart from all the rational, scientific explanations, I have learned over the years that an electric guitar is only a viable musical instrument when it's plugged into an amp, or equivalent.

Until recently, I owned what I considered one of the very best sounding electrics around, a '70 Les Paul Deluxe. Alas, the tone didn't translate when I tried to recording it DI. I learned to use a distortion box or POD or whatever to emulate the coloration that amp lends to the sound.

While I could NEVER get a good DI sound, I could easily get something musically useful with almost any stomp box or whatever.

This led me to believe that there's absolutely no point in debating the "pure" sound of an electric guitar: it's the combination of guitar and non-linear amplifier that makes the sound we want to hear.

Consequently, the sound of the guitar by itself is, while not unimportant, but of no use.

sorry, with respect, this is absolute nonsense...there are a wealth of tunes out there that you would not have a clue whether they were real amps or software..and every time Ive seen what you have said some half accomplished player has stuck samples up and persons like yourself have failed in identifying the sources of them..

just because you cant get a decent tone out of a POD (neither can I) and some nameless software doesn't mean to say others cant..

 

[mshilarious]

So here's a question for Muttley/Mshilarious: Let's say the same person was to play the same tune on several guitars, all strung with the same strings (well, not literally the same set, but same brand/gauge) and run that all into the same interface with all the same settings for each recording. Would there be any other things to think about in terms of standards for comparison? Maybe it's not that useful to have a whole bunch of people contributing, but if, say, I was to put together a database where I keep everything except the guitar the same, that could create a useful comparison database, right?

It depends on how much data you accumulate. There are basically two ways to do scientific measurements: control all of the variables and keep the measurement within the tolerances of your measurement gear. If you can manage that, a single set of data should be quite powerful, provided it is repeatable by others using your methodology. That doesn't work if the performance is not repeatable, and no human is ever capable of repeatable performances.

The other way is you collect a large amount of data and do statistical analysis of the resulting data sets. As the data gets larger, you may have greater assurance that your conclusion is more accurate--although it's entirely possible and not at all rare that your accurate conclusion is that you can say with statistical certainty that there is no correlation between the dependent and independent variables you were testing. Or you conclusion might be that you can't conclude anything from your data set.

Which reminds me of my undergrad thesis, actually My profs were all excited that I'd found statistical significance . . . of my null hypothesis

So the question becomes how large does the data set need to get? It depends on the number of variables, again. So you have to control some of them. If you played all of the guitars yourself in your own setup, each for an extended period of time, that would be a start, but you'd really need several other players to attempt the same exercise. And then you've only described whatever variables you left free.

For example, you could take your The Paul up against a Les Paul Standard, make sure the pickups, electronics, strings, and amplifier are the same, and then have say 30 players play them both under the same conditions. As mutts pointed out, you are looking for differences in fundamental and overtone series (from the strings/wood/whatever) and differences in frequency response (from the electronics). So you'd want to design the playing to capture those differences without generating a whole lot of extra data that is hard to analyze. So no jamming is allowed! You could assign the players a score (repeated a few times for each player), with a scale covering the length of the instrument, and several well-chosen chords. Then using FFT, you can measure the overtone series and frequency response, and that is one data point.

Once you have a large number of players with similar data, you might achieve statistically significant results.

That's all a lot of work and again in the end you'd probably end up with the well-known anecdotal descriptions of most instruments.

 

[mshilarious]

I still say that you guys are looking at this entirely too scientifically, but I suppose the differences among interfaces could negate some of the slight differences between guitars. I realize that you guys probably know more than I do, but please don't take the high road and treat me like I don't know anything. My degree includes a minor in audio technology, so I should hope I understand at least the basic differences. Maybe it wasn't the best thought-out plan, but I had figured that if everyone had a reasonably transparent preamp, we would be able to get some somewhat useful results.

That's actually the trouble, a transparent DI will reduce differences between guitars because it isn't loading them. When you plug the same guitar into an amp that may or may not have a very high, purely resistance input impedance, the tone can change in a complex manner.

So you need to start off deciding if you want to test electronics or acoustics, and try to eliminate one or the other from your test.

 

[Lt. Bob]

the thing is .... an electric guitar is not a stand alone thing. It's actually only part of the instrument. The amp is every bit as much of the instrument because it's a system.
A guitar and an amp are a total system. Since different guitars work differently with different amps it just doesn't matter how they sound thru any interface. It STILL won't tell you which is better because depending on which amp you run it thru, the one that sounded the best thru the interface might be the worst thru the amp and even more important, another guitar that sounded identical thru the interface might sound totally different thru the amp.
Amp/guitar are a single system so you can't leave either one out.
The only thing you'd really find out with such a test is how the guitars compare thru that interface.
And before kcearl jumps on this ..... the same would be true of a POD. Regardless of whether you use an amp or a sim ..... it's still a total system ..... git/ git amplifying whatchamacallit.

 

[c7sus]

You may not be able to tell the difference listening to an MP3 but I think WAVs would be much more revealing.

I sure as hell can tell a digital grand piano from a real grand piano.

 

[muttley600]

.......................
So here's a question for Muttley/Mshilarious: Let's say the same person was to play the same tune on several guitars, all strung with the same strings (well, not literally the same set, but same brand/gauge) and run that all into the same interface with all the same settings for each recording. Would there be any other things to think about in terms of standards for comparison? Maybe it's not that useful to have a whole bunch of people contributing, but if, say, I was to put together a database where I keep everything except the guitar the same, that could create a useful comparison database, right?

That would certainly lend itself to a more considered comparison. As you see there is little value in simply doing it independently of any control. The main problem you have would be to maintain the attack and strike position on the string as well as the age of the string. You would have to consider whether you would be looking to achieve the same amplitude of response or the same energy of attack when picking. Not the same thing! The more you examine this the ore you will find you need to qualify.

When I was involved in musical acoustic research I both published abstracts and looked to get support for them and commented on those submitted by students. Two things must always be present in any abstract without which the paper would be rejected out of hand. First, what the desired outcome is and the objective measures taken to prove it and second any observations on what may have been considered to achieve it or that may interfere with the results. All we have done is point out some of those issues to you or where these considerations are lacking.

Again I salute your intentions but would encourage you to examine the why and how of this very carefully. Two different guitars of the same type and vintage can and will give very different results. The very nature of the instruments makes this unavoidable. As I mentioned that is what makes our instrument such a wonderful thing and also why attempting to quantify any useful data in this way such an impossible task. The only real data you can collect would be to confirm the broad distinctions we are already aware of.

If you do collect a significant number of samples I would be interested to see how you intend to use the data and how you would be able to validate any conclusions you may make from those results.

 

[Christoph]

Is it just me or are people overthinking/taking this a little TOO seriously?

 

[muttley600]

Is it just me or are people overthinking/taking this a little TOO seriously?

How seriously would you like us to take it?

Would you like us to ignore obvious flaws in the concept and blindly forge ahead wasting time and effort doing something that is ultimately worthless. Or have you spotted some useful purpose that has eluded those with experience in both working with and exploring these things?

I'd be interested to hear what you have to say both on why we should ignore the obvious and also what we can gain from doing so.

 

[mshilarious]

Is it just me or are people overthinking/taking this a little TOO seriously?

Well here's the question: are we trying to tell apart a Tele from a LP (not very hard), or a LP from an SG from an Explorer? The latter is going to be much more difficult.

 

[thebigcheese]

At first, I thought they were taking it too seriously, but now I am more seriously considering how I would go about putting together such a database.

Ok, so let me clarify what exactly the goal of this project would be. The project would not be for me to have graphs of the frequency response of various guitars. I don't think the average guitar player would find that very useful. What I want is a collection of sound files from a variety of guitars (as many as possible), each with their stock equipment, so that someone trying to achieve a particular tone would have a decent idea what kind of sound they'd end up with if they picked this guitar or that guitar. The end product would be a web site where you could load up two sound clips and listen to them back to back for differences.

Using stock equipment, the pickups will obviously be different, but the goal is just to have a sound file that someone could listen to and expect that if they went to the store and bought that guitar, it would sound fairly similar to what they heard. The web site is more of a shopping tool than scientific data.

Mshilarious: I imagine different amps put different loads on the guitar, right? I'm trying to see if there's a way that I could have the guitars sound more like what they "should" sound like without introducing all the variables that go with using an amp--tone settings, volume settings, tube choice, age of tubes, age of speaker, choice of speaker, etc.

I think some of the variables become sort of irrelevant given the purpose of the project, but here are the ones I see:
-Year, brand, and model of guitar could all be put in a menu selection so that a sound clip from my 1979 The Paul would be a different clip from a 1981 The Paul.
-Only using stock equipment, so that set of variables (pickup choice, vibrato vs non-vibrato bridge, etc) can be ruled out. Hopefully you can see now why I am ruling them out.
-Position of the pickup selector could also be put in a drop-down selection.
-Performance is probably the biggest variable. I suppose that, ideally, I could build a machine that would always play everything the same, but otherwise, I guess I'd just have to get "close enough" that the differences would be practically negligible.
-Strings. Easy enough, I could just pick one brand/gauge and put a new set on right before I record. 10s seem to be the most common, so I'd probably go with those.
-Pickup height. I could just measure and adjust that, though I would need to figure out what a normal height is first.
-Leave all tone and volume knobs at max so they aren't a factor.
-Probably the most troublesome variable will be that guitars from the same year sometimes have different wood options, so I might have to create an additional menu for that.

Am I missing anything?

 

[lpdeluxe]

Post deleted.

 

[thebigcheese]

Post deleted.

You are completely misunderstanding. I'm not trying to define tone. I don't WANT to define tone. I am trying to give people a tool to compare tones so they can make an informed decision about the tone they want. I'm trying to give people the online equivalent of going to the store and plugging in a bunch of different guitars to see what they sound like, which sounds like exactly what you are so adamant about continuing to do.

 

[Lt. Bob]

You are completely misunderstanding. I'm not trying to define tone. I don't WANT to define tone. I am trying to give people a tool to compare tones so they can make an informed decision about the tone they want. I'm trying to give people the online equivalent of going to the store and plugging in a bunch of different guitars to see what they sound like, which sounds like exactly what you are so adamant about continuing to do.

It's a cool idea but it won't work, unfortunately.

The problem always comes back to the amp that ends up getting used.

No matter how a git sounds thru your interface, or even an amp if you chose to use one, it still will not tell you how the very same guitars would sound thru a different amp.
You get into things like the input impedance of the amp which will react differently with different guitars. Loading ..... gain stages .... all those design parameters affect how the guitar behaves thru it. Two guitars might behave the exact opposite of how your test showed when plugged into something different. I'm not talking about just an uniform difference for all guitars. I'm talking about even complete reversals of behavior compared to the test results.

It would be like testing a bunch of tires on a single car and then making a blanket description of how the tires work. Obviously they might behave very differently on a Ferrari than on a Neon. Using one car as a test bed for tires would be totally useless for any car other than the one you tried the tires on. You COULD NOT test some tires on a Buggatti and then make any meaningful decision on how they would perform on a F-150. ....... same thing here.

Let's say that you used your interface and found that guitar A was very 'thin' sounding and guitar B had a fat muddy sorta sound.
If you go to an amp you might find it to be the exact opposite. The 'thinner' sounding of the two guitars could be the 'thicker' sounding of the same two guitars when plugged into a different amp. ..... and then on yet another amp you might find that it does sound sorta like your test and then on a 4th amp BOTH guitars might be thin or thick or whatever.

The guitar and whatever it's plugged into, whether an amp or an interface, make up a complete electronic circuit and changing anything in that circuit can change the characteristics of the sound.
So no matter WHAT you use, amp or interface, the very next amp/interface it get's plugged into can easily behave completely differently than the tests would make you expect.

 

[loveofjazz]

i never said they sounded like wallowing in misery bacon dodgers dancing 5 to 5 with one foot on the barstool now did i?

Funniest thing I've read all day!

 

[muttley600]

You are completely misunderstanding. I'm not trying to define tone. I don't WANT to define tone. I am trying to give people a tool to compare tones so they can make an informed decision about the tone they want. I'm trying to give people the online equivalent of going to the store and plugging in a bunch of different guitars to see what they sound like, which sounds like exactly what you are so adamant about continuing to do.

I have tried not to discourage you from trying to get involved in the fascinating field of musical acoustics but I'm afraid you really aren't getting this. Your latest premise is that by collecting raw samples of the output of a guitar we can use this as a base on which to formulate a start point from which to base an informed choice when selecting a guitar. This simply won't work.

Look at it this way. You want to select from a restaurant menu. You can approach it one of two ways. You can base your choice of menu on the cooked and prepared dishes as you know them to be from your experience and comparison to other known finished dishes and the base ingredients or you can arrive armed with knowledge of how all the raw ingredients taste and base your selection on that. I know which way I would be inclined to go. Same with guitar tone.

As a guitar builder I constantly make judgments on what will make the finished guitar sound the way I and the customer want. They tell me what they are after and I draw on years of experience of the materials to try and get to that result. The result is ALWAYS greater than the sum of its parts.

Good luck with the project but I fear it will not get the support it needs to get it off the ground.

 

[Zaphod B]

the thing is .... an electric guitar is not a stand alone thing. It's actually only part of the instrument. The amp is every bit as much of the instrument because it's a system.

Let's not forget an equally important part of the system: the player. Two different guitarists playing the same rig will often obtain recognizably different sounds.

 

[muttley600]

Let's not forget an equally important part of the system: the player. Two different guitarists playing the same rig will often obtain recognizably different sounds.

Thats exactly what I was getting at when saying that you need to consider attack and where on the string, and the angle and how hard, and with what...