Official Luthiers Forum!
http://w-ww.luthiersforum.com/forum/

Acceptable intonation accuracy
http://w-ww.luthiersforum.com/forum/viewtopic.php?f=10101&t=18682
Page 1 of 2

Author:  ncovey [ Sat Sep 06, 2008 10:51 am ]
Post subject:  Acceptable intonation accuracy

Hi everyone;

How many here use a Peterson 490ST strobe tuner or similar strobe tuning device to tune their instruments?
More importantly though, How accurate are your ears, in comparison
to what a strobe tuner suggests?
If you take any off the shelf guitar, and most hand builts (mine too) tune it with a strobe, you'll find that in conventional set-up (90 degree nut face to compensated saddle, with fret spacing set according to any given scale) [The Peterson 490 declares an accuracy of .01 cent/1-1000 of a semi tone.] the progressive notes of each string vary from flat to sharp by as much as 7 cents.
Buzz Feiten has a compensated nut system, a custom made cantilvered nut that has staggered break points for each string, which is supposed to remedy this, however, if I'm not mistaken, his system still allows for a 3 cent discrepancy after so many frets (10th or so) . I have had a recent request for a commission that is entirely focused around improving this accuracy to +/- .05 cent. So many factors are present in the nature of a guitar, action height, string gauge, neck relief, humidity, temperature, individual playing style, bad habits in fretting, etc.. and lastly, try turning your guitar on it's side, back and top while hooked up to the strobe, what was 2 cents flat on it's side is now reading 1 cent sharp. Not sure what causes this fluctuation, gravity?? .Who knows, but in any case, my question is,
just how accurate is accurate enough in terms of acoustic guitar intonation? I suppose it would be nice to have a guitar as finely tuned as a grand concert piano, but really, being a hand held, hand manipulated stringed instrument with many factors for error. ?? I wonder about the validity of such a goal. I never want to turn away a commission, and I am willing to work with this request as it seems very challenging, I like those kind.
Actually what I have come up with is intonating from the mid scale point (12th fret) each way, rather than the (earvana method of) 1st fret down, so this trail off in accuracy in terms of +/- 1 cent is reduced by half by spreading each percentage of error above and below 12th, instead of all discrepancies leading up the fret board. Does this make sense? I might need another cup of coffee.

Thoughts on this would be appreciated , as I start this build this week, I am confident that I can intonate each fret and string length pretty much dead on, but the process is very complex and time consuming. Also I have to rule out human error from building to playing for this to result in perfect intonation in every note.

Author:  Jim Watts [ Sat Sep 06, 2008 12:50 pm ]
Post subject:  Re: Acceptable intonation accuracy

I don't think you'll get there unless you want to build a guitar to just play in one key. The whole tuning thing (fret placement) is a series of compromises that produces a satisfactory (or unsatisfactory) result in all the keys. In the west these days we use an equal tempered scale that accounts for these compromises.
Also (I might be wrong here) but I think the fretting differences between players would throw you off that much.
You might try contacting Mike Doolin as he's done a lot of work in this area. He's also published some of his work in the GAL journals within the last couple of years I believe.
Good luck!

Author:  ncovey [ Sat Sep 06, 2008 1:28 pm ]
Post subject:  Re: Acceptable intonation accuracy

Thanks for the response Jim,
You see, that's what my point is too, given the differences in playing/fretting style alone, this pin-point accuracy attempt is nearly an impossible quest.
All things being what they are and the accepted averages of cent gain or loss in any key with the formulas we all are familiar with in properly, or at least commonly intonating an instrument, is in my opinion, simply good enough.
Spot on open "E" tuning (A-440Hz) with a fretted "B" note showing a loss of 1.25 cents is not even detectable by ear. At least not by mine.

I have to wonder if some people are really that keen of hearing?
I would much rather make a fuss over the complete sound and projection of a guitar than to worry about this detail. If you can call it a detail.
But, I know lots of players are better than me, and I wanted to see what the popular sentiment was regarding this level of accuracy.
Is it even understood or recognized as a discrepancy to the average guitarist (me) as compared to a grand concert pianist? I think it's apples/oranges. The instrument is only as good as the player.

Author:  Terence Kennedy [ Sat Sep 06, 2008 1:32 pm ]
Post subject:  Re: Acceptable intonation accuracy

I've always been a little confused about using perfect intonation with a strobe used as a selling point or demand by a customer. As mentioned, a few degrees change in temp or humidity, the position of the instrument, brand of strings, playing technique, etc. all change things. So far I've found that intonating each string separately at the saddle with a good quality tuner to have been adequate along with cutting the nut end of the fretboard as a fret which shortens it by around 0.012. No complaints so far.
Do fanned frets, Buzz Feiten, Strobe tuners really make a difference in the real world or are they more in the line of a sales technique?
Terry

Author:  Kent Chasson [ Sat Sep 06, 2008 1:49 pm ]
Post subject:  Re: Acceptable intonation accuracy

Did you really mean .05 cents or .5 cents.

Even .5 would be quite a feat and if you got there, you likely wouldn't stay there very long. Changing humidity (action) or changing strings alone would likely throw it off more than that. Slight deviations in stiffness from one set of strings to the next (even within the same brand) can effect intonation more than that.

And yes, the reason a guitar changes pitch when turned 90 degees is gravity. As you turn a guitar from a vertical playing position to horizontal, the weight of the neck pulls it back which stertches the strings slightly and sharpens them. So this goes to show yet another way in which playing technique causes intonation problems. Pressure on the neck as well as fretting technique can easily cause deviations of a few cents.

Also consider, as Jim said, that equal temperment alone leaves notes out of tune in some keys by way more than that (can't remember off the top of my head but it's at least 5 cents).

Author:  ncovey [ Sat Sep 06, 2008 2:31 pm ]
Post subject:  Re: Acceptable intonation accuracy

...I actually did quote that right. .05 cent.
This man is willing to pay for my full attention to this request. I made no promises to him except a total purchase "base" price and my effort to achieve +/- 1 cent being my unsubstantiated goal. (per note) based on A-440Hz. And I only made this offer after some simple experiments.
I really don't think there is anything positive to be gained in doing this, but I am actually very curious after making a few minor nut and saddle adjustments to a Goodall, then playing it on a strobe and comparing to a factory set up yamaha and one of mine. It was still off (according to the tuner) by about .50 cent on average with the worst case being 3 cent sharp on "B" string in a place or two. So like I said, I think I can hit +/- 1 full cent but the process is slow and easily the worth the additional charge. I do have to add that before the changes, the overall tone especially up the neck, which did sound off, improved afterward.
Keep the comments coming , I'll go get a beer.

Author:  ncovey [ Sat Sep 06, 2008 2:41 pm ]
Post subject:  Re: Acceptable intonation accuracy

...I actually did quote that right. .05 cent.
This man is willing to pay for my full attention to this request. I made no promises to him except a total purchase "base" price and my effort to achieve +/- 1 cent being my unsubstantiated goal. (per note) based on A-440Hz. And I only made this offer after some simple experiments.
I really don't think there is anything positive to be gained in doing this, but I am actually very curious after making a few minor nut and saddle adjustments to a Goodall, then playing it on a strobe and comparing to a factory set up yamaha and one of mine. It was still off (according to the tuner) by about .50 cent on average with the worst case being 3 cent sharp on "B" string in a place or two. So like I said, I think I can hit +/- 1 full cent but the process is slow and easily the worth the additional charge. I do have to add that before the changes, the overall tone especially up the neck, which did sound off, improved afterward.
Keep the comments coming , I'll go get a beer.

Author:  John Mayes [ Sat Sep 06, 2008 6:10 pm ]
Post subject:  Re: Acceptable intonation accuracy

accuracy within a few cents is deemed acceptable.

Author:  Rod True [ Sat Sep 06, 2008 10:48 pm ]
Post subject:  Re: Acceptable intonation accuracy

Everything I would have said, has been said already except for one thing,

I hate to say it, but this customer will never be happy! You may be able to get the guitar intonated in your shop to +/-1 cent (although I doubt ANYONE could get it to within 0.05 cent without a laser guided system of some sort) but as soon as you pick it up it will be lost.

The player will not be able to touch the instrument if he wants to keep it in that fine of tuning. As soon as his hands get warm and they warm up the neck it will change, as soon as he takes it from a colder room to a warmer room it will change. He will never be able to hold that type of accuracy, not even close.

Why does he want one of your guitars? Because they sound and play great? Do you know if or how many other builders he's gone and asked this same question of and been denied because it's such an impossible task?

If that were someone coming to me to ask me to build a guitar in order to intonate it to that type of accuracy, I would politly tell them that the level of stiffness in the neck and body required to hold that sort of accuracy would render a guitar which was too heavy and over braced and would not sound very good at all. I would also say that building to such a ridgid structure would compromise my style of building.

Now if he's willing to live with a heavy guitar that sounds bad (compared to lightly braced, great sounding guitars that move when the strings are plucked) than by all means build him the guitar. BUT, I can see something like this coming back to bite you in the rearend if you take it on.

I've heard it said that the average human ear can not detect a note out of tune by 5-8 cents. There may be some with better hearing than that but I highly doubt that anyone could detect a note being out of tune by any less than 2-3 cents. 0.05 cent is quite over the top if you ask me.

Author:  ncovey [ Sun Sep 07, 2008 12:00 am ]
Post subject:  Re: Acceptable intonation accuracy

Well, thanks for the replies everyone. I see I'm surely not alone in my opinion of "over the top" requests from some people.

I don't mind the average picky client, as long as everything is documented and agreed upon in the normal legal fashion, I state everything in writing that I can and will afford in terms of guarranteed delivery and end results, so I'm not worried about anything "biting" me later. But I was curious if any one else had such particular requests.
I think about Stevie Ray on stage playing Little Wing , and can't help but wonder. Man, is his guitar properly intonated... :roll:

Really though, I'm not trying to invalidate this persons or anyone elses notion of what is acceptable accuracy, but I have to view this from a builders standpoint. I just wanted to see if any of you here had similar
criteria specified in a build.
I'll still proceed if he holds up his end in what my "contract" stipulates.
I figure the willingness to oblige a customer as best I can, will be a positive.
Whether or not it nets +/- 1 cent, it will still play and sound as good as I can build.

Author:  Colin S [ Sun Sep 07, 2008 6:21 am ]
Post subject:  Re: Acceptable intonation accuracy

You should point out to the prospective purchaser that final intonation is the responsibility of the player, some players that you listen to are just astonishing in the accuracy of their playing, purely mechanical intonation will never be accurate accross the board and in all keys. Listen sometime to a player like Martin Simpson, he seems to be constantly aware of the pitch of every note he plays and adjusts it accordingly as he goes. My wife was a long term orchestral violinist and like most experienced unfretted string players developed perfect pitch, I use her ears rather than a strobe to adjust my intonation, but then at best it can only be an approximation.

When I make and play my lutes I'm even working with tied on frets and no saddle to set intonation on. I just have the method of pulling the string loop forward or back from the bridge to make slight adjustments to string length. It's down to the player (me!)to make the final adjustments to fret placing and fingering. Fretted players have become lazy and seem to expect everything to be done for them.

Also how does he intend to tune it, which method? Open string, fifth fret, harmonic, 'failsafe'? They will all make a difference.

I think he is chasing an impossible dream, and sounds like a nightmare customer!

Colin

Author:  Hesh [ Sun Sep 07, 2008 7:21 am ]
Post subject:  Re: Acceptable intonation accuracy

I agree with the others that this client may, I say may be problematic. But before I decided that they were I would discuss this with them and bring up some of the very fine points made here.

Let them know that the request is unrealistic and unreasonable and that you are not willing to commit in writing although you are willing to do the very best that you can toward excellent intonation. I would also only commit in writing to any spec if the methodology that will be used to determine "success" is clearly and technically defined in advance. When meeting an expectation specification the definition of success is very important.

Some folks do not know that they are making unreasonable requests and when it is pointed out to them and explained to them in a manner that they can understand they will change position. This exchange will also provide you with a better feel for this prospect who is a bit "suspect" at the moment now...... If the client is unwilling to change their expectation I would respectfully suggest that they seek another builder.

That's my 2 cents :D worth.......

Author:  Howard Klepper [ Sun Sep 07, 2008 8:44 am ]
Post subject:  Re: Acceptable intonation accuracy

Decline this commission. It's no brainer. I'd advise you to decline it at half a cent; your client is asking for the impossible.

Author:  TonyKarol [ Sun Sep 07, 2008 8:59 am ]
Post subject:  Re: Acceptable intonation accuracy

I would decline as well .. thats just insane. As has been said, finger pressure while playing and string differences from one set to another will kill it.

I havent taken the time to figure out why, its just an abservation, but I find that now matter whose gutiar, or how things are done, any chord or cluster with 3rds and 6ths are always off to my ear (and most others who I show this to).

On the Goodall, or Yamy or your own Nehemiah, if you get a G (with a 2nd string D note, not the open B, thats how I play em) and D chord sounding as good as you can (even there, the I find the G will be perfect, and the D will have a 6th in the f# and sounds sharp to me) then play a C, isnt the G flat sounding ??? (its the 3rd of the E in that chord). Fix the G in the C chord, and now the D chord is off because the A is sharp.. cant win. Now use a tuner to get the open strings best you can, does this get any better ??? I find not. 3rds and 6ths.. hate em.

I put an Earvana on a factory Strat a couple years back (strat was 70s I think) ... the first position tuning/chords were awful before we started. Once done, it was much better, but what blew me away was the fact that the G string was shortened by a whopping 4mm (nut to 1st fret). I would never have believed it if I hadnt seen and done the work myself.

Author:  Andy Birko [ Sun Sep 07, 2008 11:39 am ]
Post subject:  Re: Acceptable intonation accuracy

TonyKarol wrote:
I havent taken the time to figure out why, its just an abservation, but I find that now matter whose gutiar, or how things are done, any chord or cluster with 3rds and 6ths are always off to my ear (and most others who I show this to).


As others have mentioned, tuning is weird to begin with. Equal tempering results in a lot of "out of tune" chords. Things like orchestras or choirs can "tune up" by ear and achieve a beauty in tuning that we can't achieve with instruments locked in to a certain temperament.

Here's an interesting site: http://www.phy.mtu.edu/~suits/scales.html

Author:  Jeffrey L. Suits [ Sun Sep 07, 2008 12:26 pm ]
Post subject:  Re: Acceptable intonation accuracy

Howard Klepper wrote:
Decline this commission. It's no brainer. I'd advise you to decline it at half a cent; your client is asking for the impossible.


There's a BIG sign, saying: "Don't step in this !". Pay it heed...

Author:  Erik Hauri [ Mon Sep 08, 2008 7:23 am ]
Post subject:  Re: Acceptable intonation accuracy

#1: Make sure your "contract" is a CONTRACT.
#2: If the client insists on a guaranteed intonation spec, make sure it is something you can hit with your eyes closed...even (especially) if you're charging more to try and hit the 0.05-cent spec. And put a caveat in, like "that's the intonation spec as it leaves the shop...and not a nanosecond later". If he balks, tell him "thanks, but no thanks".

If he's still interested, charge him the +$2000 an Eventide Eclipse, deliver it with the instrument, and tell him to make sure he plays through the pitch correction patch every time he touches the guitar. laughing6-hehe

Todd is right on here....homogeneity of string elasticity alone may not be enough to get to 0.05 cent on every fret. You'd need an oscillosope to see that. And as Colin's wife would (I'm sure) agree, tone is in the fingers.

Author:  Michael Jin [ Mon Sep 08, 2008 9:12 am ]
Post subject:  Re: Acceptable intonation accuracy

I've always found guitar intonation to be a sticky issue. Unfortunately the guitar isn't an instrument that can be perfectly accurate in intonation for all keys at once and yes, the difference in fretting technique alone between players can throw off the intonation... Also different people seem to have different levels of tolerance when it comes to what they consider "acceptable" and "unacceptable" intonation. I'd say in most cases, make sure your customer understands this fact and get it as close as you reasonably can... The guitar is an instrument that is constantly changing with time, weather, temperature, different sets of strings, etc...

Author:  Michael Dale Payne [ Mon Sep 08, 2008 2:40 pm ]
Post subject:  Re: Acceptable intonation accuracy

Most mid range to mid high end digital electronic chromatic tuners on the market claim a tolerance of +/- 4cent. Most tested musicians with near perfect pitch can recognize a 1.5-.75 cent variant. Those referred to as having perfect pitch seem to recognize variants between .75-.50 or less but they are very rare. I am guessing but I would bet that the average pro musician has pitch variant recognition somewhere in the 2 cent range. And this is all based on generated pitch souned one pitch at a time, not notes played on a guitar in a lyical format which is subject to fretting technique and attack and environmental variables.

If you have a prospective client wanting to spec a .50 cent accuracy tell him you will have to charge top dollar the vacuum chambered enviro lab needed to insure the instrument can maintain such a tolerance.

Then comes the issue of how he will confirm the tolerance you held. You both would have to have your tuners calibrated to each other or neither could confirm the others readings.

This is an unreasonable specification.

Author:  David Collins [ Mon Sep 08, 2008 3:32 pm ]
Post subject:  Re: Acceptable intonation accuracy

The request demonstrates a fundamental misunderstanding of the nature of the guitar. A player who can play and release the same note several times with intent to maintain the same pitch (without watching the tuner) and keep a tolerance of +/-1¢ is doing pretty good. If they try to play the same note fretted in the context of different chord shapes (say in E, A, and C shaped bar chords), they are doing good if they land within +/-2¢. If the guitar were made of carbon fiber, kept in an environmental chamber, fretted and plucked by a programmed robotic arm, you may be able to maintain consistency of 0.05¢, but only at specific points through the duration of the note. Many areas on the guitar will change 3, 4, or more cents from the attack through the decay.

The request also demonstrates a fundamental misunderstanding of our equal tempered scale. The only notes that are perfectly tune even in theory are unison and octaves. Fifths and fourths are around 2¢ sharp and flat of perfect harmony, while thirds and sixths are around 14¢ and 16¢ out of tune. No matter how perfectly the frets and saddle are positioned, it's still the responsibility of a skilled player to bring those notes in to tune by pushing, pulling, and stretching as they play if they want good intonation. This makes demands of such tolerances especially absurd, as in and of itself it will not make any equal tempered instrument sound any more in tune. Responsibility for good intonation ultimately lies in the hands of the player.

Now it's always good to try to educate your customers, as many are entirely ignorant of any inherent flaws that exist within our scale to begin with. So it can help to explain the limitations to the customer to prepare them for real world results. But........

This request also waves red flags of a naturally malcontent client.

They are probably requesting this because they can't play in tune, regardless of how well an instrument is built. No matter how accurate your intonation in frets and saddle, they will probably not sound any more in tune than on any other well set up instrument.

If you were to take a commission from this client, I would make it very clear that you would be happy to build them an instrument, but that intonation would not be taken as any special issue whatsoever. The guitar will simply be as well intonated as any other well made guitar. Make it clear that they should study up on temperament and intonation, and no matter how perfect a guitar is, the system is still flawed and final intonation would be left in their hands as they play. If this is their primary motive in commissioning a custom instrument, simply refuse the job. They will not be happy with the results, as neither yourself nor anyone else who lives in our universe of mathematics and physics will be able to solve their problems.

From how it sounds so far, I would likely walk away.

Author:  Rod True [ Mon Sep 08, 2008 3:38 pm ]
Post subject:  Re: Acceptable intonation accuracy

Nice to see you back David. Great input as usual [:Y:]

Author:  Michael Jin [ Mon Sep 08, 2008 4:02 pm ]
Post subject:  Re: Acceptable intonation accuracy

David Collins wrote:
The request demonstrates a fundamental misunderstanding of the nature of the guitar. A player who can play and release the same note several times with intent to maintain the same pitch (without watching the tuner) and keep a tolerance of +/-1¢ is doing pretty good. If they try to play the same note fretted in the context of different chord shapes (say in E, A, and C shaped bar chords), they are doing good if they land within +/-2¢. If the guitar were made of carbon fiber, kept in an environmental chamber, fretted and plucked by a programmed robotic arm, you may be able to maintain consistency of 0.05¢, but only at specific points through the duration of the note. Many areas on the guitar will change 3, 4, or more cents from the attack through the decay.

The request also demonstrates a fundamental misunderstanding of our equal tempered scale. The only notes that are perfectly tune even in theory are unison and octaves. Fifths and fourths are around 2¢ sharp and flat of perfect harmony, while thirds and sixths are around 14¢ and 16¢ out of tune. No matter how perfectly the frets and saddle are positioned, it's still the responsibility of a skilled player to bring those notes in to tune by pushing, pulling, and stretching as they play if they want good intonation. This makes demands of such tolerances especially absurd, as in and of itself it will not make any equal tempered instrument sound any more in tune. Responsibility for good intonation ultimately lies in the hands of the player.

Now it's always good to try to educate your customers, as many are entirely ignorant of any inherent flaws that exist within our scale to begin with. So it can help to explain the limitations to the customer to prepare them for real world results. But........

This request also waves red flags of a naturally malcontent client.

They are probably requesting this because they can't play in tune, regardless of how well an instrument is built. No matter how accurate your intonation in frets and saddle, they will probably not sound any more in tune than on any other well set up instrument.

If you were to take a commission from this client, I would make it very clear that you would be happy to build them an instrument, but that intonation would not be taken as any special issue whatsoever. The guitar will simply be as well intonated as any other well made guitar. Make it clear that they should study up on temperament and intonation, and no matter how perfect a guitar is, the system is still flawed and final intonation would be left in their hands as they play. If this is their primary motive in commissioning a custom instrument, simply refuse the job. They will not be happy with the results, as neither yourself nor anyone else who lives in our universe of mathematics and physics will be able to solve their problems.

From how it sounds so far, I would likely walk away.


QFT

Author:  Philip Perdue [ Mon Sep 08, 2008 6:08 pm ]
Post subject:  Re: Acceptable intonation accuracy

I find this thread very interesting. Recently I was having a discussion with a well respected custom ukulele builder on the subject of intonation. My position is that you can have good or even great intonation but never perfect intonation. The other builder seemed to take offense at my comments but that is his issue not mine. It is my understanding that the smaller the scale length the harder it is to get good intonation. My guess is that since ukuleles don’t use metal strings the there may be more room for error. There are numerous brands and a composition of ukulele strings each having different characteristics that need to be adjusted for. I plan on building a jig to test string properties for different brands of strings and sizes of ukuleles. That will give me the compensation needed for each string and I can work the issues from there.

Nehemiah, it looks like you have some really good advice already. You now have good information when you and the client get back together. One of the other posters may have had it right when they said the prospective client may not know what he is asking for.

Best of Luck

Philip

Author:  ncovey [ Mon Sep 08, 2008 7:14 pm ]
Post subject:  Re: Acceptable intonation accuracy

Thanks for all the feedback everyone, and it's especially nice to see the unity in opinion.
I just spent a few hours with this individual, and I can tell now, after having him here and playing with me for a bit, he has only the slightest notion of what he is asking me to do. I also logged in and had him read a couple replies to this post.
He argued with some of it and eventually realized the importance of the factors we all stated (to him).
He still wants me to build him a rather nice "OM" ,african blackwood/port orford , in a 25.4 scale , 1-3/4" nut. with a cutaway. with much less emphasis on the intonating. He's played mine and can see how even the slightest change in temperature, neck position etc..changes the accuracy.
So I have to say to all of you . Thanks ,and once he signs I'll start and post pics as I can.

Author:  Terence Kennedy [ Mon Sep 08, 2008 7:37 pm ]
Post subject:  Re: Acceptable intonation accuracy

Rod True wrote:
Nice to see you back David. Great input as usual [:Y:]

Ditto! Don't be a stranger David.
Terry

Page 1 of 2 All times are UTC - 5 hours
Powered by phpBB® Forum Software © phpBB Group
http://www.phpbb.com/