Jump to content

A Test Rig for Assessing Tonal Properties of Solid Body Electric Guitar Construction


Recommended Posts

@mistermikev's thread a few weeks ago regarding an experiment he noted on a private Facebook group got me thinking about practical and meaningful methods of testing factors that may be important in a solid body guitar in terms of sonic impact. I'm sure everyone can point to numerous experiments and comparisons floating around the interwebs, but I'm also sure just as many people will highlight deficiencies in the testing methods that could void the conclusions presented.

As with most of the tonewood tests I've personally seen, one of the single biggest hurdles to overcome is the lack of consistency in playing the instrument to demonstrate the effect some component of a guitar has on the resultant sound. It would be ideal if this could be eliminated, or at least reduced to a level where some confidence that the method used to excite the strings to vibrate is as consistent as possible. In this way isolating just one element of a guitar's construction may be possible. Below is my attempt to create such a test rig:

image.png

 

Ignore the white platform-y thing on the other side of the pickup for the moment. I'll have more to say on that a bit later on.

The 'guitar' is just a single plank of Tas Oak I've cut and rough dressed to arbitrary dimensions of 840mm x 65mm x 25mm. A single headless tuner and string lock have been mounted at each end, and a plastic nut placed near the string lock giving an effective scale length of 25". The single string can be quickly swapped out and replaced as required with another of different tuning or gauge.

The decision to utilise a typical nut for the string to pass over, rather than simply allowing the string to vibrate all the way to the string lock, was taken as it was found during initial testing that any slight misalignment of the string leaving the string lock could potentially allow it to contact the sides of the hole on the face of the string lock. This caused issues such as curtailed sustain, sitar-like buzzing or unexpected changes in tonal response. By using a nut any interference caused by the threading hole is eliminated, and the strings vibration is allowed to occur relatively naturally and consistently. The downward pressure exerted on the nut as the string passes over it may also assist in any potential 'tonal transfer'  of the string's vibrations into the timber that might be expected in a real guitar:

image.png

The pickup is a single coil of an unknown vintage retrieved from an old Yamaha RGX electric guitar mounted to a wooden plate that in turn is screwed into a rebate cut into the oak plank. This allows the pickup to be quickly detached from the body, and to be re-mounted in the same position as accurately as possible. The end-most polepiece is aligned under the string. Other pickups could easily be substituted if required:

image.png

The pickup is wired directly to a 1/4" jack mounted on a short section of angle aluminium at the far end of the rig. A single 240Kohm 1% tolerance resistor is soldered across the hot and ground leads of the pickup to approximate a typical 250Kohm volume pot wound up to maximum. A ground wire is also run between the output jack and underneath the headless tuner base plate, which grounds the string to help with noise suppression, just as in an actual guitar. The output is then connected to a Presonus FP10 8-channel Firewire audio interface and recorded direct to disk at 44.1kHz 24bit using Reaper digital audio workstation software, which can then be used to quickly compare recordings or run analysis on the waveforms if required:

 image.png

To excite the string into motion it would make sense to try and replicate the action of plucking the string in a manner as realistically as an actual player would, but with as little variability as possible. The method utilised here is to employ a plectrum attached to an axle that rotates around a full 360 degrees of motion, plucking the string as it passes overhead. A small block of wood has a shallow notch cut into one end, into which a 0.7mm 'medium' Dunlop nylon pick has been press-fitted into. The block has in turn been press-fitted onto an 8mm axle supported by a pair of pillow bearings to allow it to rotate freely:

image.png

The other end of the shaft is coupled to a small stepper motor. The use of such a motor allows precise rotational speeds to be applied to the shaft, largely irrespective of the resistance that might occur as the pick begins to contact the string as it swings past. This allows for a method of controlling the relative strength and repeatability of each pick strike:

image.png

To control the stepper motor's motion, an Arduino programmable microcontroller development board has been employed to issue a series of control pulses to a stepper motor driver. This particular driver has been selected primarily because it features a smooth motion option for lowest possible noise and vibration, which could otherwise acoustically and electrically couple back into the test rig and influence the measurements unnecessarily:

image.png

Additional control of the motor is provided by use of a single push button to trigger the motor to begin rotating and execute a pluck of the string. A secondary limit switch attached to the plucking rig is actuated when the pick completes a full rotation and signals the Arduino to halt the motor's rotation. Pressing the start button again recommences a further  pick strike. The picking mechanism takes approximately 5 seconds to complete one full revolution before stopping:

image.png

The plucking rig may be freely positioned up and down along the length of the test rig and is attached using rubber pads to grip on to the sides of the rig. It may also be set at any height above the string, allowing as much or little of the pick to contact the string as desired. The use of rubber pads also helps to acoustically decouple the plucking rig from the test rig to minimise the possibility of mechanical noise from the motor or bearings influencing the produced signal.

The test rig itself is also resting on soft rubber bumpers to minimise any possible acoustic coupling from the desk underneath that may have an effect on the resultant output signal:

image.png

To ensure the pickup is set at a consistent distance from the string between tests, a 2mm thick wooden spatula is used as a makeshift feeler gauge. Wood is used as it will not be attracted to the pickup and interfering with the measurement, but other non-ferrous materials could also be employed:

image.png

Some of the possible characteristics that could be tested with such a rig could be whether the output of the pickup is influenced by whether it is direct-mounted or decoupled via a pickguard or mounting ring, what differences there may be between top mounting and through-body stringing, differences in tone due to scale length variation, or (dare I say it) what differences may be audible in the output signal between materials used in a solid body electric guitar.

More later...

  • Like 1
Link to comment
Share on other sites

Thanks! I fully expect this to potentially open a can of worms, but I am genuinely curious about how some of this stuff could be measured and assessed, and I think as builders without ties to a larger company interest we should all be curious about it as well. I've been guilty in the past of saying that something is so because I think it should be, but with this experimental base I'm more interested now in finding out if there could be any basis to such claims, whether they point one way or the other. I'm trying to approach this with an open mind.

I think it's also important that if anyone wants to add their own opinions and ideas to the matter they should feel free to do so, whether in this thread or elsewhere. As long things remain civil, of course :)

Link to comment
Share on other sites

3 hours ago, curtisa said:

Thanks! I fully expect this to potentially open a can of worms, but I am genuinely curious about how some of this stuff could be measured and assessed, and I think as builders without ties to a larger company interest we should all be curious about it as well. I've been guilty in the past of saying that something is so because I think it should be, but with this experimental base I'm more interested now in finding out if there could be any basis to such claims, whether they point one way or the other. I'm trying to approach this with an open mind.

I think it's also important that if anyone wants to add their own opinions and ideas to the matter they should feel free to do so, whether in this thread or elsewhere. As long things remain civil, of course :)

I nominate you my official r&d department... you don't have a budget... but you don't have a mandate either!! 

I am genuinely interested in whatever you find whether it supports my own bias or not... in fact I'll prob learn more if not - seldom pleasant but always healthy.

Link to comment
Share on other sites

You can pay me in Kempers :D

The writeup at the top of this thread is actually a few days out of date. I do have recordings from at least one experiment, but I haven't had a chance to compile the files in any great detail yet. Drawing my own conclusions based on what I can hear or see is one thing, but my intentions are to provide some write-ups on what I'm doing here and release a bunch of comparative recordings so everyone can judge for themselves what they can or cannot hear.

Based on the one (as yet unreleased) comparison I did with the above rig though, I'd say it's really hard to do this kind of work. A true scientific study, which this most certainly is not, is a huge undertaking to get right. The work involved to account for or eliminate as much of the unwanted data, and focus on the variables you want to look at must be extremely labour-intensive.

Link to comment
Share on other sites

Genius is one percent inspiration and 99 percent perspiration,” said Thomas Edison.
 
That's a nice approach! One could argue, though, that the resistor doesn't transfer sound equally to a <brand name> volume pot on 10. That one is not me!
I dont' know if it would matter or not, but the wooden plectrum holder doesn't mimic the flex of the flesh of the fingers holding it. A wider slot padded with semi-soft rubber, maybe? Then again, we're after consistency so don't bother!
Link to comment
Share on other sites

One thing that would be interesting to test with this is sustain with different setups eg. string trees behind the nut vs. no string trees, more/less string after the nut (inverted headstock vs. normal) and definitely string stuck to the bridge vs. strings going through body/wood.

Awesome setup, this is as scientific as it gets, looking forward to some results :) 

Link to comment
Share on other sites

1 hour ago, Bizman62 said:

I dont' know if it would matter or not, but the wooden plectrum holder doesn't mimic the flex of the flesh of the fingers holding it. A wider slot padded with semi-soft rubber, maybe? Then again, we're after consistency so don't bother!

I guess we'd have to draw the line somewhere in terms of practically mimicing a hand holding a pick vs picking the string. I suppose you could go to great lengths to design and build some kind of robotic hand with synthetic 'flesh' to hold the pick that accurately mimics someone's picking action. But you could also begin splitting hairs and say, 'well that's not the way *I* play guitar, your tests are all null and void!' :D This is more about getting the strings to move in a way that is repeatable and using a tool that a typical player would use to do so. Assessing tonal effects due to a player's picking hand technique is well beyond the scope of this test rig.

I was explicitly trying to avoid the variations that might be encountered in trying to excite the strings in some of the Youtube tonewood comparison videos that seem to have received lots of attention over the years:

  • This one by our old mate Ola Strandberg is interesting, but the method that was used to get the strings moving was to allow a ruler to swing down under gravity and 'hammer' the string, effectively turning the guitar into a piano or dulcimer, which is not how anyone plays an electric guitar. There's also inconsistency between some string strikes within the same wood samples that (to me at least) raises some doubt that it's possible to conclude each timber sample is actually sonically different.
  • This old favourite is commonly cited, and while the claim is that the two guitars differ only by body material, the biggest difference I can see is that there's two different people playing the instruments. You'd also need to consider any differences in the electrical characteristics of the pickups and electronics and the proximity of one player to the amp vs the other (acoustic feedback?), amongst other things.
  • Another one that seems to pop up occasionally also suffers from the human factor. It's clear that the same riffs are being played differently each time, which changes the validity of the comparison. Sometimes the samples sound the same between timbers, sometimes they sound noticably different.
  • Like 1
Link to comment
Share on other sites

1 hour ago, Gogzs said:

One thing that would be interesting to test with this is sustain with different setups eg. string trees behind the nut vs. no string trees, more/less string after the nut (inverted headstock vs. normal) and definitely string stuck to the bridge vs. strings going through body/wood.

There's lots of different things you could try with this kind of rig. I'll add some of your ideas to my 'to-do' list :thumb:

  • Like 1
Link to comment
Share on other sites

6 hours ago, curtisa said:

I guess we'd have to draw the line somewhere in terms of practically mimicing a hand holding a pick vs picking the string.

I fully agree. That's why I told you not bother.

Link to comment
Share on other sites

Attached is the first recording worth sharing. All this WAV file is intended to demonstrate is if the 'robo-picker' can operate in such a way that each pluck of a string sounds consistent. A set of 10-46 strings was used, starting with the low E 46 first. This was tuned to pitch using a Peterson StroboPlus tuner set to normal equal temperament mode, and then a recording of 10 robo-plucks made. The string was then detuned and removed, the next string (A/36) installed and tuned, and a further recording of 10 plucks performed. The process repeated until all 6 strings had been recorded.

Each recording of 10-plucks-per-string was then sliced up into 2 second soundbites and spaced equally apart. The resultant recording attached here only demonstrates 5 of the 10 samples of each string (to save yourself from going mad ;)), but there should sufficient examples to be able to make a sufficient judgement as to whether they sound uniform enough to continue using the robo-picker rig.

Don't judge the tone too harshly at this stage - there is absoutely zero processing on these recordings, nor is there any attempt to make the 'guitar' sound like it would normally through an amp, whether simulated or real. Consequently the recording sounds extremely lifeless and uninspiring from a tone perspective. The string-to-string comparison is also not what should be focussed on at this point; rather focus on how consistent (or not) each of the 5 plucks of the same string sounds.

01 Test Rig Consistency Check.wav

Link to comment
Share on other sites

I liked the sound but the tune wasn't too catchy. In fact it was so boring I stopped halfways in the A part...  Nice calm tempo, though, maybe I'll add this to my go-to-sleep playlist.

Just out of curiosity, I came to thinking that on the first picture the pick was deep in the holder but later it was just barely pinched by the edge. Was that to prevent grabbing, allowing the pick to flex enough to release the string?

Link to comment
Share on other sites

Something to compare for yourself - can you tell the difference in sound between a pickup that is direct-mounted on to the body, and one that is completely decoupled from the instrument?

The theory is supposedly that if you have the same pickup directly attached to the body it will sound different compared to a pickup not in direct contact with the body via a scratchplate or pickup mounting ring; that the pickup may be able to respond to the material vibrating. I needed some way of taking the pickup out of the test rig and re-positioning it such that it was no longer touching the body, but still in the same location relative to the string. The 'white platform-y thing' I mentioned at the top of this thread finally gets a chance in the spotlight here. The pickup remains attached to the timber mounting plate, but the mounting plate itself is unscrewed from the test rig and re-attached to this new jig on the other side of the guitar. This holds the pickup in the same relative position as it would be if attached directly to the guitar, only this time it is above the string pointing down, rather than underneath pointing up:

image.png

image.png

The height of the pickup when held in the inverted position is adjustable using M5 allen bolts. The cradle itself is also mounted on soft rubber strips to minimise any acoustic coupling from the test rig, through the desk and up into the pickup.

image.png

To ensure the pickup polepiece is at the same distance from the string the same 2mm wooden spatula is used to fine tune it:

image.png

So. Interactive time.

Attached are three WAV files for your listening pleasure. A low-E, a D and a high-E have been recorded 10 times each using the 'pickup direct-mounted' rig. The same three strings have then been recorded a further 10 times with the 'pickup free-floating' rig for a total of 60 individual samples across three pitches.

  1. File 02a - Contains three pairs of samples of each string chosen at random from the two banks of 30 samples, giving 3x2 strikes of the low-E, 3x2 strikes of the D and 3x2 strikes of the high-E. The first sample in each played pair is always the direct-pickup version (A), the second sample in each pair is always the floating pickup version (B). The playing order of each pair is always A-B. Can you hear anything could differentiate one sample from the other within each pair?
  2. File 02b - Similar to file 02a except that this time the order of each pair of strikes is now randomised. The first sample in each pair may be either the direct-mounted pickup (A) or the free-floating pickup (B). The second sample is always the opposite version compared to the first. Based on what you could hear in file 02a, can you identify the playing order of each pair of samples, A-B or B-A?
  3. File 02c - A sample of each of the three strings is played three times. The first (A) sample is always the direct-mounted version , the second (B) sample is always the free-floating version. The third (X) sample could be a recording from either the pickup direct-mounted or free-floating. Can you identify the third (X) sample in each group of three based on what the previous two samples (A-B) sounded like?

02a Pickup Mounting, Direct vs Floating.wav 02b Pickup Mounting, Direct vs Floating, Randomised Order.wav 02c Pickup Mounting, Direct vs Floating, ABX.wav

Link to comment
Share on other sites

My wooden ears could hear no difference! At times I thought there might be something but then that something disappeared. If there's any difference it might be in the attack. Even so it feels more about imagining a difference when the ears can't hear any.

I wonder if the difference could be heard better with an overdriven pickup. Would extra gain accent such very subtle differences? Or would it hide any nuances into buzz?

Link to comment
Share on other sites

A slightly simpler, and perhaps more obvious one this time around. What difference does scale length make?

Same test rig as before, except that the string lock and nut at the 'head' end of the oak plank are being shifted towards the bridge by an inch at a time, providing scale lengths of 26", 25" and 24":

20210502_135111.jpg

20210502_140513.jpg

20210502_141106.jpg

Just two strings played this time, a low-E (0.042 gauge) and a high-E (0.010). One pluck of each at 26", 25" and then 24" scale lengths.

Also attached is the same recording through the moderately gainy amp sim. Interestingly I think the differences between the three variants get a bit harder to hear when played through the amp. Try them for yourself!

03 Scale Length Variations, 26-25-24 + Amp.wav 03 Scale Length Variations, 26-25-24.wav

Link to comment
Share on other sites

After listening all the samples I dare say that neither the pickup mounting nor the scale length really matters, the sound is mostly in the fingers of the player!

Back when I was at school they had some recordings to support the music book. On one of the tapes they cut the attack off of the sound samples. You couldn't tell if the sound came from a guitar or a piano without hearing the very start. This is similar, the sounds are very similar and the only difference I imagine hearing is during the first nano-second.

Yet another thing to test is to move the picking location from neck pickup (1/4 of the scale length from the bridge), mid and bridge in the locations where one of the harmonies get eliminated due to counteracting waves - along the maths described here: https://www.physicsclassroom.com/Class/sound/u11l4d.cfm

I'm suggesting this as I've learned that an essential part of the sound of the bridge pickup is just because the second harmonic gets muted on that spot. How much might picking at the third harmonic at 1/6 scale length from bridge change the sound?

And then the pickup location at the harmonic muting spots...

I feel like I'm losing the focus!

Link to comment
Share on other sites

Interesting. I can hear a change in tone between the differing scale lengths. There's a high frequency peak that I can hear getting lower as the scale length reduces on the low-E string plucks. Kinda like those Tuvian throat singers who can excite harmonic overtones on top of their base drone vocal to sound like 'iiii', 'eeee' and 'oooo'. To my ears the high-E also gets less bright as the scale length gets shorter. That could be a contributing factor to the Strats-are-brighter, Les-Pauls-are-warmer philosophy, the two guitars having scale lengths that differ by around 3/4 of an inch. I really struggle to hear those differences as soon as I add the amp simulation though.

Link to comment
Share on other sites

Hmmm? You say something? ;)

No, the peak I'm hearing isn't anywhere near up that high. It's subtle, so it probably won't stand out if you're just listening on the laptop speakers or mobile phone. I suppose it may only come across as 'slightly more treble'/'slightly less treble'.

Link to comment
Share on other sites

Low e sounds pretty much the same for me, but the high e does get less "trebly" or however you might call that... I need to pull this trough some spectral analysis when I get home, this is rather interesting :D I thought it would get more "sharp" the less the scale length is... 

Link to comment
Share on other sites

The falling 'trebly-ness' in sympathy with the shortening scale length pretty much follows what I was expecting. The string tension is reducing while the pitch and string gauge remains constant. If you followed it to an extreme situation, say making the scale length as short as a ukulele but kept the same string on it tuned to the same pitch, the note generated would sound pretty dull and it would feel like a rubber band. Try down-tuning a 46 low-E on a Strat to the same pitch as a bass low-E and you'll get the same effect.

  • Like 1
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

 Share

×
×
  • Create New...