An attempt at an unbiased string comparison test

Choice of classical guitar strings and technical issues connected with their use.
astro64
Posts: 934
Joined: Sat Aug 06, 2016 7:43 pm
Location: American Southwest

An attempt at an unbiased string comparison test

Post by astro64 » Sat Feb 10, 2018 10:22 pm

Instead of trying to rely on memory as to how strings sound, when making comparisons between different strings, I am trying a more objective test. You have to have patience for this one, but I figure if I can observe the differences clearly, I will save lots of time in the future...

One of my guitars is currently strung with 6 different second string brands. Why test the second string? On the one hand it is the least picky string to sound good, it is always warmer than the first and easier to get good vibrato, etc, and it is never the highest tension string in the set. But I am always interested in an extra bit of clarity in the lower positions, where it seems the string sometimes struggles a bit to keep up with a good first string and a carbon third. So I put 6 second strings on the guitar. I measured all the string gauges with a micrometer, on the string part not under tension, just for information. The strings are installed in the following order:

1 = La Bella PEPESR, supposedly high tension, but feels medium. 0.79-0.80 mm thick.
2 = D'Addario J45, medium tension, 0.80 mm thick (not 0.816mm as the package info would give)
3 = Aranjuez Suave 800 (LT), 0.79 mm thick, this one is possibly closer to a "titanium" type of string, although Aranjuez calls it differently
4 = Savarez Alliance MT, 0.66-0.67 mm thick ("carbon treble")
5 = Savarez Crystal MT (but definitely stiffer feel than J45s), 0.79 mm thick
6 = Augustine Regal (HT), 0.81 mm thick

The strings will be kept on for a few days to make sure they are fully settled. Initial impressions are that differences are more subtle than one might assume, and that the PEPESR is least bright in the lower positions, not the J45. We'll see in a few days.

Laudiesdad69
Posts: 2442
Joined: Tue Nov 29, 2016 1:16 pm

Re: An attempt at an unbiased string comparison test

Post by Laudiesdad69 » Sat Feb 10, 2018 10:46 pm

This sounds like an interesting experiment.

User avatar
rojarosguitar
Posts: 4661
Joined: Sat Sep 19, 2009 12:24 pm
Location: near Freiburg, Germany

Re: An attempt at an unbiased string comparison test

Post by rojarosguitar » Sat Feb 10, 2018 11:14 pm

Interesting experiment, though I guess there is a lot of asymmetry in the response in different places these strings occupy. Guitars are by no means symmetrical.

Maybe you would need to permutate them in all combinations in order to get a feel for that...
Music is a big continent with different landscapes and corners. Some of them I do visit frequently, some from time to time and some I know from hearsay only ...
My Youtube Channel is: TheMusicalEvents

User avatar
petermc61
Posts: 6430
Joined: Sun Apr 08, 2012 12:11 am
Location: Sydney, Australia

Re: An attempt at an unbiased string comparison test

Post by petermc61 » Sun Feb 11, 2018 2:55 am

Very interesting experiment. How would the total tension compare to a normal set of strings? I presume it might be lower?

astro64
Posts: 934
Joined: Sat Aug 06, 2016 7:43 pm
Location: American Southwest

Re: An attempt at an unbiased string comparison test

Post by astro64 » Sun Feb 11, 2018 3:03 am

I would indeed think the tension is lower than a normal set. Usually the 4th and 1st string have the higher tension, as I recall and the 2nd string the lowest. As to the difference due to where the string is placed, I think that is secondary effect. The entire top is engaged in producing sound generally, and it would be wrong to assume there is a treble and a bass side to the top. But there could be subtle effects, which I am not going to explore[*]....This will be confusing enough. One thing that will not play a role is the string distance above the top; for this particular guitar the saddle has equal height over the top from bass to treble, so the torque on the top will be the same for all strings. Action is of course different. I was surprised to find out that although saddle and nut are compensated, intonation for the strings is good, the only ones going a little sharp by the 12th fret are 5 and 6. The alliance, which needs less saddle compensation than the nylon trebles sits on the 4th string slot so it too plays in tune because the saddle edge is more towards the bridge.

[*]PS. What I will do in a couple of days is interchange 1 and 6, since they already seem to show among the largest difference, just to see if string placement affects those two.

User avatar
rojarosguitar
Posts: 4661
Joined: Sat Sep 19, 2009 12:24 pm
Location: near Freiburg, Germany

Re: An attempt at an unbiased string comparison test

Post by rojarosguitar » Sun Feb 11, 2018 9:36 am

astro64 wrote:
Sun Feb 11, 2018 3:03 am
As to the difference due to where the string is placed, I think that is secondary effect. The entire top is engaged in producing sound generally, and it would be wrong to assume there is a treble and a bass side to the top.
I didn't mean to imply there is a bass and a treble side. There is an asymmetry, though, and so every one of the strings will meet with different boundary conditions and stiffness parameters during the transient phase, which may have quite an effect. I mean, we are talking about quite subtle differences anyway...
Music is a big continent with different landscapes and corners. Some of them I do visit frequently, some from time to time and some I know from hearsay only ...
My Youtube Channel is: TheMusicalEvents

es335
Posts: 1470
Joined: Thu Aug 30, 2007 2:12 pm
Location: Germany

Re: An attempt at an unbiased string comparison test

Post by es335 » Sun Feb 11, 2018 11:04 am

Hi astro64, interesting experiment but the measured diameters seem all quite low even though they are just for information!

User avatar
souldier
Posts: 1087
Joined: Tue Dec 08, 2015 4:45 pm

Re: An attempt at an unbiased string comparison test

Post by souldier » Sun Feb 11, 2018 1:31 pm

Will be interesting to see what you come up with. Some potential pitfalls is that you may prefer the second string of one brand but the first string of another, etc. The other issue is that you won't be able to play any music, which in the end is what we play the guitar for. I would find it quite difficult judging strings purely from their sound when it is detached from playing actual repertoire. The third issue is that in the end, string sets need to be judged as a whole and how they collectively impact the sound of the song you're playing. I could imagine this might be helpful in figuring out what second string could potentially be your favorite in terms of sustain, tension, etc. but there is a certain collective X factor that can only be experienced when you have a complete set on.
"Success grants its rewards to a few, but is the dream of the multitudes.
Excellence is available to all, but is accepted only by a few." - Christopher Parkening

astro64
Posts: 934
Joined: Sat Aug 06, 2016 7:43 pm
Location: American Southwest

Re: An attempt at an unbiased string comparison test

Post by astro64 » Sun Feb 11, 2018 4:32 pm

es335 wrote:
Sun Feb 11, 2018 11:04 am
Hi astro64, interesting experiment but the measured diameters seem all quite low even though they are just for information!
All I can say is that I measured them carefully with a Stewmac digital micrometer, taking an average of about 3-5 measurements at the loose end of the string at the head of the guitar after the strings were installed.

astro64
Posts: 934
Joined: Sat Aug 06, 2016 7:43 pm
Location: American Southwest

Re: An attempt at an unbiased string comparison test

Post by astro64 » Sun Feb 11, 2018 4:37 pm

souldier wrote:
Sun Feb 11, 2018 1:31 pm
Will be interesting to see what you come up with. Some potential pitfalls is that you may prefer the second string of one brand but the first string of another, etc. The other issue is that you won't be able to play any music, which in the end is what we play the guitar for. I would find it quite difficult judging strings purely from their sound when it is detached from playing actual repertoire. The third issue is that in the end, string sets need to be judged as a whole and how they collectively impact the sound of the song you're playing. I could imagine this might be helpful in figuring out what second string could potentially be your favorite in terms of sustain, tension, etc. but there is a certain collective X factor that can only be experienced when you have a complete set on.
Indeed, the goal is not to find the one best string but a couple that then can be combined with a matching 1st, ideally. But e.g. I will probably not select a HT first string even if the 2nd string is a HT version of some brand. The difference in tension between 1st and 2nd string is always very large and I see no issue in combining HT 2nd with MT first. So I would rather select a string of the same brand but MT for the 1st string. And basses can easily be decoupled from trebles without any problems. Lastly, the 3rd carbon treble is very neutral and works well. So, from my perspective, it is mostly the 1st string match that matters. Third string nylons work fine on this guitar but I hate the extreme temperature sensitivity given the thickness of the string. The sound is also more "inharmonic" on a nylon third due to that thickness, but it can add some character so that can work out fine.

es335
Posts: 1470
Joined: Thu Aug 30, 2007 2:12 pm
Location: Germany

Re: An attempt at an unbiased string comparison test

Post by es335 » Mon Feb 12, 2018 1:21 pm

astro64 wrote:
Sun Feb 11, 2018 4:32 pm
es335 wrote:
Sun Feb 11, 2018 11:04 am
Hi astro64, interesting experiment but the measured diameters seem all quite low even though they are just for information!
All I can say is that I measured them carefully with a Stewmac digital micrometer, taking an average of about 3-5 measurements at the loose end of the string at the head of the guitar after the strings were installed.
A digital micrometer or calliper? In case of a micrometer make sure the torque of the ratchet is as low as possible. In case of a calliper don‘t use the front blades for measurement as recommended. They may cause indentations in the soft Nylon strings, which leads to smaller values. Use the flat section right behind the blades and apply as few pressure as possible. I use a high definition dial gauge calliper that way and the measurements are usually in good conformity with the suppliers data, provided there is no remarkable ovality. There are deviation occasionally, even once a complete (non-fake!) EJ43 package, but it's not the rule particularly in case of d‘Addario. They are usually spot on, one of their unquestionable strengths! :!:

BTW measuring the ends before installation is a bit more comfortable! :wink:

astro64
Posts: 934
Joined: Sat Aug 06, 2016 7:43 pm
Location: American Southwest

Re: An attempt at an unbiased string comparison test

Post by astro64 » Mon Feb 12, 2018 3:58 pm

es335 wrote:
Mon Feb 12, 2018 1:21 pm
astro64 wrote:
Sun Feb 11, 2018 4:32 pm
es335 wrote:
Sun Feb 11, 2018 11:04 am
Hi astro64, interesting experiment but the measured diameters seem all quite low even though they are just for information!
All I can say is that I measured them carefully with a Stewmac digital micrometer, taking an average of about 3-5 measurements at the loose end of the string at the head of the guitar after the strings were installed.
A digital micrometer or calliper? In case of a micrometer make sure the torque of the ratchet is as low as possible. In case of a calliper don‘t use the front blades for measurement as recommended. They may cause indentations in the soft Nylon strings, which leads to smaller values. Use the flat section right behind the blades and apply as few pressure as possible. I use a high definition dial gauge calliper that way and the measurements are usually in good conformity with the suppliers data, provided there is no remarkable ovality. There are deviation occasionally, even once a complete (non-fake!) EJ43 package, but it's not the rule particularly in case of d‘Addario. They are usually spot on, one of their unquestionable strengths! :!:

BTW measuring the ends before installation is a bit more comfortable! :wink:
Stewmac digital caliper, sorry for the mixup in terms. I did use the thicker part of the caliper blades. I think the measurement differences are consistent among the different strings. Whether the caliper squeezes the strings a tiny bit, I can't rule or our verify easily. The ends of the strings after putting them on are still quite long since I didn't cut them for this experiment.

es335
Posts: 1470
Joined: Thu Aug 30, 2007 2:12 pm
Location: Germany

Re: An attempt at an unbiased string comparison test

Post by es335 » Mon Feb 12, 2018 6:54 pm

I suppose you always perform a zero adjustment before the measurement!?
Anyway the results are consistent and represent the relative diameter ranking which is sufficient for information purposes!

I’m excited to hear the outcome of you experiment! :wink: :D

astro64
Posts: 934
Joined: Sat Aug 06, 2016 7:43 pm
Location: American Southwest

Re: An attempt at an unbiased string comparison test

Post by astro64 » Fri Feb 16, 2018 1:02 am

Time to summarize, I can't take this any longer ;).

To recap I tried 6 different B-strings all installed at the same time. Guitar: Greg Byers, cedar/EIR. In order of string position on the guitar:

1 = La Bella PEPESR, later switched to Augustine Regal, then to Savarez Alliance
2 = D'Addario, J45
3 = Aranjuez Suave 800
4 = Savarez Alliance MT, later switched to Augustine Regal
5 = Savarez Crystal MT
6 = Augustine Regal HT, later switched to La Bella PEPESR

Procedure:

Robert was correct that it would make a difference where the string would be mounted. I clearly heard more volume from the strings mounted on the positions of strings 4, 5, 6. This seems a promity effect, the ear is closer to the initial attack on the top for the lower strings. To minimize this effect I played the guitar holding it low on my leg facing a corner. This pretty much eliminated the most glaring differences in volume. I also exchanged a few strings to confirm that impressions were not due to string location. Still, this test is by no means "exact science". I played both single note to note comparisons, while damping the other strings, and single string melodies repeated on different strings. I did this once or twice a day for the past 5 days and took some notes of what I thought I heard.

Comments/results:

- Do this experiment at your own risk... It is a good way to test your ability to maintain your sanity...

- The guitar (+player) dominate 95+% of the sound, not the brand of string. The differences in volume, tone, sustain were subtle.

Identical notes played on successive strings were VERY much the same in character. Variations in note timbre from one fret to another on a particular string were much larger than variations in timbre for the same note from one string to another string. So, e.g. the character of a C on all strings is pretty much the same, whereas the character from C to C# can change much more on any one string due to resonances in the guitar top and body and listener position. I would not want to identify any of these strings in a blind test.

- Variations in sustain among the strings were minimal. Not a deciding factor. The Alliance might beat the others by a bit.

- Some trends emerged and largely confirmed some impressions from before the test, based on many years of playing various types of string: harder tension strings can sometimes produce a "pinging sound" (hard to describe, a matter of nail length and technique so it can be avoided), they can be a bit louder, they do not have longer sustain. Timbre is rounder, not necessarily less bright but can lean to a more artificial sound (nasal?). This tends can be worse playing closer to the bridge. This is the main reason why in the past I have tended to prefer medium tension strings, other than ease of playing.

Generally, the J45 and Aranjuez were quite close though not identical in character and overall produced a nice sound, round tone, no lack of brightness, few rough edges to the timbre. The Alliance was the most different in timbre. It has somewhat more volume, a leaner yet at the same time still quite solid sound. It can have more of a bright edge to the sound, that needs care in right hand. Playing with right hand close to the soundhole it was close in character to the nylon trebles, the difference is more obvious when playing towards the bridge. I think the reputation of brightness that carbon trebles have is mostly true for the first string, which is thin and generally is a big jump up in tension, less for the 3rd or 2nd. (I find carbon third stringsless bright in general and more neutral than nylon thirds).

My conclusions of this test:

- Guitar dominates the sound character, not the strings, by a large factor (who would have thought ;)?). This seems totally obvious but it is good to have this reinforced; Whenever we might feel something lacking in the sound, odds are it has more to do with the time of the day and how our attention and hearing are doing, shape of the nails, etc, than what string is on the guitar.

- All of these strings ultimately worked fine, and did not give a dramatic difference to the sound. The tension difference may be more a factor to decide what to use than the sound will be, although for the B string, even the tension differences are nothing to write home about. The Regal is probably stiffest but not by that much.

- I had tried the Crystal MT B before and it worked fine according to my earlier notes. The J45 is fine too and I have used it many times; the Aranjuez 800 will undoubtedly be too and it works very well on my Kono. I am going to give the Alliance B string a try on the Byers to test its character over a longer period.

No, I am not going to repeat this experiment for the other 5 strings!

User avatar
souldier
Posts: 1087
Joined: Tue Dec 08, 2015 4:45 pm

Re: An attempt at an unbiased string comparison test

Post by souldier » Mon Feb 19, 2018 11:47 pm

Thanks for putting your mental health on the line for our learning! I enjoyed reading through your findings. I am not surprised by the major conclusion that the differences in string brand are not that dramatic. I've often said on this forum before that if you don't like the sound of your guitar, chances are you probably need to change the guitar rather than the strings.

I've also found that the difference in string brands is quite subtle, and the inherent sound of the guitar will always come through. This is why I ended the endless string search.
Last edited by souldier on Mon Feb 19, 2018 11:54 pm, edited 1 time in total.
"Success grants its rewards to a few, but is the dream of the multitudes.
Excellence is available to all, but is accepted only by a few." - Christopher Parkening

Return to “Classical Guitar Strings”