Jump to content

Tuning Precision & Accuracy


Recommended Posts

  • Replies 130
  • Created
  • Last Reply

Top Posters In This Topic

 

So now you have me going...

 

The idea that a concertina would be tuned to a different standard depending on the discriminational skills of the owner is in my opinion wrong. Why not strive to do the best you can +/- 0.5 on all the instruments, not just for the exception that can hear the difference. I would expect nothing less from a proffesional that I trusted to work on my instruments.....maybe thats why I have trusted Greg to do my work.

 

And you're prepared to wait while it's being done, and (I hope) to pay considerably more for all that extra work.

 

But many people aren't.

 

Yes I am willing to wait and pay of a job well done. In many ways those that are not willing to wait/pay get just what they paid for!

Why do you call it "extra" work?

Link to comment
Share on other sites

So now you have me going...

 

The idea that a concertina would be tuned to a different standard depending on the discriminational skills of the owner is in my opinion wrong. Why not strive to do the best you can +/- 0.5 on all the instruments, not just for the exception that can hear the difference. I would expect nothing less from a proffesional that I trusted to work on my instruments.....maybe thats why I have trusted Greg to do my work.

 

AND

 

The idea that the anglo(which is what I play) should be on the bottom of the pecking order and does not deserve to be tuned just as well as the english or duet is also wrong.

 

Very interesting ideas being tossed around.....just don't understand the logic.

 

Doug

I

 

 

Doug,

 

Please define 'professional' I think there are a number of highly competent repairers who are contributing towards a discussion which I would like to think will enable some form of approach amongst those of us who have to balance the capability of an instrument, the playing style of the customer, economic and service level considerations. I hope we can get to a consensus which might lead others, whether tuning for themselves generally, to not just 'rough tune' and walk away

 

To imply that I and others taking part in this debate are not professional in our respective approaches is both unreasonable and insulting.

 

You are concerned that some of us operate dual standards subject to the discernment of a player. We have to take into consideration the way a player hears his instrument, and how hard they play it, to they play predominantly chordally or more single voice. Some people do get more attention because they feel something might not sound right when they play specific chords or even intervals, as Stephen said:

 

1st stage the meter

2nd stage the tuner's ear

3rd stage the player's ear

 

with 3 being the final arbiter. At the end of the day the player has to be happy with his or her instrument.

 

There is of course the need to be at nominal to be able to play in sessions and ensembles.

 

Remember also, this is not like tuning a stringed instrument with a peg you can tighten & release, this is about removing metal that cannot be replaced, to make it even more entertaining, a reed can sound anything up to 15 cents (or more) different when sounded in free air on a rig, then when sounded in the actual concertina. As a result, all tuning is done by difference, it is difficult to determine a correction if the error cannot be measured, both in and outside the concertina.

 

I don't subscribe nor have I advocated treating Anglos with any less diligence than any other system. As I explained, Other systems have specific needs that Anglos do not. On the Other hand Anglos do have octaves, intervals and cross row duplication of notes which have to be matched. Duplicates only occur on an English system on the accidental enharmonics, but do also occur on duets in the overlap ranges. Perhaps a code of tuning practice should identify specific System's need?

 

 

Dave

Link to comment
Share on other sites

I've chosen to use 1/5th Comma Meantone ( Homogenous ) and whilst this is a halfway house compromise from the full 'perfect thirds' meantone (1/4 Comma) it still has notes that are quite widely spread away from Equal Temperament.... But I have yet to have another musician question the position of my notes, when playing in groups or sessions.

What I mean here is that if an instrument is 'in tune' with itself and is close to the Pitch standard ( in my case spread evenly around it) then other players will not think it out of tune.

 

Yes, I've found fifth comma to be the best - quarter comma sounds too "extreme" to me.

 

But let me explain, for the benefit of others reading this, that that entails tuning only one note in the scale to 0 and the rest are anything up to 30.5 cents apart, whilst both the enharmonic intervals G# to Ab and D# to Eb are tuned 28 cents apart, to avoid "the wolf" - and it all sounds much better in tune than the harshness of "regulation" equal temperament

 

So much for "zero"! :rolleyes:

 

I recall my days living in Co Clare. and the Knock on the door of someone needing a quick fix or tune up... on their way to a gig or before their Examination the following day at Limerick University!

 

You know the story, and the people. ;)

 

It's maybe Knock they should be going to, not you or me...

Edited by Stephen Chambers
Link to comment
Share on other sites

There is another facet to this that has not been alluded to, Lachenal produced his basic 'tutor' models with low grade and inefficient reeds, all intended to be affordable & 'cheerful'. No one would expect the same tuning precision as on say an edeophone. Therefore there was always some form of tuning tolerance & discretion given to the work bench. I think it is fair to say that what we have been discussing is as applied to a steel reeded instrument of middle to higher grade?

 

Dave

Link to comment
Share on other sites

A very important part of tuning is taking the readings. I believe the bellows pressure should be "moderate" and consistent with every note. Otherwise the results will be as varied as your bellows pressure when you took the readings. I learned this years ago when a very prominent player, an All-Ireland Champion, flew in from San Francisco for the expressed purpose of having me tune his Jeffries concertina before he was to make a recording. To save time, I asked him to play each note on his own concertina while I wrote down the pluses & minuses of each note. We were chatting about Irish music and musicians while this was being done. Next step was to correct the notes by flattening the sharp note and shapening the flat notes as per the previously done readings. It was a bit of a disaster. The instrument was no more in tune after one run though than when we started. I asked him to go for a walk while I did it again. Consistency of pressure is vital. He had been inconsistent with his pressure while playing the notes. The pitch of concertinas is pressure sensitive, especially with an instrument which had been brought down from high pitch to A-440 by filing the reeds flatter. (This was not me!) The amount of pressure sensitivity is also greater with the lower notes.

But this anecdote does not address David's original question. I aim for zero on the meter and take several passes to achieve this. I use a computer program for this. Then I play octaves and chords to confirm this. I then play the instrument for a while to check for inconsistencies of response, and finally a last tuning check. But there is no perfect tuning as there are no perfect instruments or perfect players. As Greg suggested, and I have discovered also, some players are brutal with their playing pressure, playing their instruments out of tune by their playing pressure, and soon putting their instruments out of tune, for the same reason. The amount of out-of-tuneness caused by excessive pressure depends on where the note is ---highness or lowness according to reed size, the number of times the reeds have been tuned, if the instrument was in one of the "high" pitched and has been brought down to A-440 by filing, and the quality of the reeds. Even a smaller amount of pressure differential can affect the tuning. As I stated earlier, I play at a consistent, moderate pressure when taking readings of the notes. My customers are usually hundreds, if not thousands of miles away. Their conception of their own playing style (hardness or softness) is very subjective and not necessarily the same as others' conception, or mine. Unless they are here for me to listen to, all I can do is tune to a "best practices" model and tune to what my experience has found to be a "good" playing pressure.

No, Dave, I wouldn't worry about that one customer, too much. I know all of us have been out to dinner with someone who always sends the meal back because it does not meet their imagined standard of perfection, and do so with a certain amount of rudeness to the poor wait person. Embarassing for those around him at the table, too. I suspect that noone could have made your concertina tuning customer happy.

Edited by Frank Edgley
Link to comment
Share on other sites

In many ways those that are not willing to wait/pay get just what they paid for!

 

You seem to be implying that I don't give them back a concertina that is in tune and playing well - but they (some of whom are very discriminating) don't agree with you.

 

 

Why do you call it "extra" work?

 

For the very simple reason that for me (with a "manual" meter, though not for Greg with an "auto" one), six passes take six times as long as one, or three times as long as two, all for the sake of a difference that isn't even audible.

 

Tell me, do you have any problems with the way pianos are tuned?

 

Edited after reading Geoff Wooff's comments below, my only experience of the Peterson 490 Auto Strobe being a personal demonstration on their booth at the Musikmesse in Frankfurt - though I'd still like to try using one myself...

Edited by Stephen Chambers
Link to comment
Share on other sites

 

 

I recall my days living in Co Clare. and the Knock on the door of someone needing a quick fix or tune up... on their way to a gig or before their Examination the following day at Limerick University!

 

You know the story, and the people. ;)

 

It's maybe Knock they should be going to, not you or me...

 

 

Yes indeed Steve!

Of all the stories I could tell the one 'happening' I enjoyed most was the time a very nice young lady from Miltown came to me with a sick Low F natural on her Jeffries. She said she needed that note for a Performance the following day, the culmination of her course work at the University... Well, that reed was cracked and unrepairable... So I went looking in my spare reed tin, a small collection housed in a 2 ounce tobacco tin... and there I found a Jeffries low F natural of exactly the right size... this fitted the slot without the need of any changes or packing and save for a few Cents of fine tuning was a perfect match. The likelihood of me having a correct replacement reed for a 100+ year old concertina was lost on our young friend... who just said " Oh thanks, how much do I owe you ?" I thought the moment was priceless so I just wished her well with her studies and did not charge her a penny. It made my day alright.!!

Link to comment
Share on other sites

Stephen,

I think there is no "extra" work. If you can tune a concertina in 3 passes and it take Greg 6 well so be it, as long as the outcome is the same, a well tuned concertina, understanding all the variations that come with tuning the instrument, the customer and cost of the process. I think it is sort of like practicing, you practice until you have mastered the part, not by how long it takes.

 

I wonder if one of our tech savvy members could set up a poll with a listening test. Maybe two tones played one after the other =/- 1.0 or +/- 0.5 apart and we can take the test to see which one is sharp or flat compared to the other, to see if this is really audible....might be fun and enlightining to all of us.

Link to comment
Share on other sites

http://tonometric.com/adaptivepitch/

 

I wonder if one of our tech savvy members could set up a poll with a listening test. Maybe two tones played one after the other =/- 1.0 or +/- 0.5 apart and we can take the test to see which one is sharp or flat compared to the other, to see if this is really audible....might be fun and enlightining to all of us.

Link to comment
Share on other sites

There's a clue in Dave's original post where he talks about the difference between precision and accuracy; for example see this article: (http://www.dspguide.com/ch2/7.htm). Unfortunately, I threw away my university surveying notes a few years ago otherwise I would have dug up the formulae for error correction which simplified calculation of the standard deviation and hence precision.

 

As Dave and others have pointed out, there has to be a tolerance - there is no such thing as "spot-on" on this planet; I've given up being offended by people who want an unachievable idea of perfection (and I'm talking generally here, not just about tuning concertinas), since it's usually pretty clear that when they ask for it, they have no idea of the effort involved in achieveng it, nor can they tell when they're presented with the outcome whether it's met the desired specification (since their own measuring faculties are mostly deficient).

 

As Rory Sutherland has also pointed out (look him up on TED Talks and check some of his other posts), all value is subjective. I can't find his direct quote just at the moment, but he's also said that since most people cannot discriminate at the very high end of any item's specification, hence most people don't want (and can't determine) the very best of anything, they simply want something which is very unlikely to be terrible (since they don't want to be laughed at down at the pub).

 

It could be that one tuner achieves exactly the same precision through taking six "passes" at tuning as another tuner might take through taking 3 "passes", simply because the precision is being improved through averaging of the measurements. This doesn't tell us anything about the accuracy, since there could be calibration differences between the instruments that the two tuners are using (for example - there may be a number of other differences as well - such as in bellows pressure, humidity, room acoustics).

 

You certainly shouldn't assume that one tuner is "better" because she can achieve in one "pass" what it takes someone else multiple "passes" to achieve, nor that the tuner who takes multiple "passes" must necessarily be achieving either greater precision or accuracy.

 

I put "pass" in quotation marks as it could be that we each mean different things by what a single "pass" is.

 

My own tuning practice (based on 13 concertinas and around 700 reeds) is as follows. I try very hard to get a precise reading at moderate bellows pressure of what the actual note is in the instrument in my room. I use either an analogue device or a strobe tuning app on the iPhone. I typically go through this process a number of times and write down the readings to verify that I have something like a consistent impression of the number of cents sharp or flat a particular note is. I then tune the reeds outside of the concertina, using the same tuning device as used to record the readings, recognising that the actual reed frequency is different outside the box to inside and thus trying to sharpen or flatten the reed by the appropriate level of cents. That is one "pass". If I've had to tune by more than +/- 10 cents, I'd call that rough tuning. I then re-assemble the instrument and check all the notes again, not just the ones I've corrected; again, I note down all the readings and at this stage I'd expect to be at least within 5 cents and hopefully within +/- 3 throughout the box. I then leave it for a day or so and check the readings again, but this time I'll only note down the notes which are outside -1/+2 cents. Then I'll go through the tuning process again, but only on those notes which need it. That's a second "pass" and is definitely fine tuning. My aim is to get to within +/- 1 cent by the end of a job and this can take 3 passes and sometimes more. I'm not so constrained by time as others may be and I'm almost never working to the sort of deadlines that a professional repairer like Stephen is, whose customers need the concertina for a gig or recording. It also takes me longer to meet "my" standard since I'm still at the beginner level of tuning and I'm very reluctant to take metal off to make dramatic changes so I proceed very cautiously.

 

In my experience (of many more instruments than those I've tuned), most instruments can have reeds which are out of tune to a considerable extent - up to 20 cents - and no-one notices, either because the other instruments adjust, or there are more serious issues which mask the out-of-tune-ness, or because they are close to a sweeter temperament or because they sound better in different rooms (or with an adjustment in playing style/dynamics), or, or, or.

 

I have come across one player who had stopped playing her instrument (a duet) because the push note sounded different than the draw. I couldn't tell the difference and nor could the other players present but the tuner showed a 1 or 2 cents difference. I showed the reed the file, put the box back together, checked the tuner again which showed I'd made some difference and the "customer" pronounced herself very happy. Some people have very acute ears but most don't. If your customer has those kind of ears, then they have to be prepared to pay for more time to be taken.

 

Alex West

Link to comment
Share on other sites

Hmmm. Sweet, bitter, sour...how about salt? :)

 

I have a 1/5 comma mean tuned instrument that Geoff Wolf sold to me and which Greg Jowaisis tweaked a bit when he did some minor adjusting for me. I concur that it sounds much better to my ears than the E.T. Edeophone I have. When I first started playing it I didn't think I'd notice the difference much but I really do like the sound much better now that I've had some time with it. And, I too have had not complaints from anyone I play with. The first time I took it to a rehearsal with a flute friend she was out of tune for the first couple of tunes, but then adapted and we could play lovely unisons. I never mentioned the tuning differences...until after the rehearsal.

Link to comment
Share on other sites

For the very simple reason that for me (with a "manual" meter, though not for Greg with an "auto" one), six passes take six times as long as one, or three times as long as two, all for the sake of a difference that isn't even audible.

By "auto" do you mean the type of meter that guesses what note is being played (with controls for reference pitch and temperament) and tells you the error?

 

FWIW the tuning work I've done so far is with a cheap but good Mac program called Katsura Strobe Tuner. It seems to work very well, though I think I'm now going to have to add the intervals for 1/5 comma Meantone to its temperament list! B)

 

The thing that for me would speed up the process the most is a way to note down the error value for each reed without having to put the concertina down and pick it up again between each note. I tried recording the process and speaking the errors out loud, then playing back the recording and transcribing them, but it wasn't really much faster overall. How do others deal with that issue? Recruit an assistant? ;)

Link to comment
Share on other sites

In my experience the Peterson 490 Auto Strobe takes just as long to use as the old Yamaha PT4 ( that I started with in 1976).... yes the automatic note recognition works fairly well but you still have to adjust the knobs to tune the thing into the exact pitch of the note you are trying to measure. In the end it works better when set to manual, especially for very high or low notes.

 

Stephen, I would stick with the PT4 , as I recall, it was quicker to use.

 

Many tuners now have programable settings for other Tempéraments but, as far as I am aware, these allow for only 12 notes... for the EC you need 14 and for some other types of music one might need many more than that.

Edited by Geoff Wooff
Link to comment
Share on other sites

on Precision v Accuracy.

 

in rifle shooting terms, a loose 5 shot group but with an average centering spot on the middle is accurate but imprecise, a tight group but centred in the 5 ring is precise but inaccurate, a five shot group through the same hole in the 10X ring is both accurate and precise

 

In concertina terms I have always discussed a tolerance around nominal, irrespective of temprement etc. You know what your target frequency or note is, and you are tuning your reed to that value. Stephen tunes a miniscule sharp, +1/ -0, Greg gets to around +/- 1 to 2 cents and then refines by ear, Johann tries to get as close to zero a practicable, I go +/- 1.5 cents and refine by ear, all tuners take cognizance of a players play style and playing pressures. Greg has a vacuum pup to eliminate pressure variation on the rig, we all use different metering systems, but work in and resolve to decimals of a cent, even if we use bands of say 0.5 cents for our tuning practice.

 

Error tends to follow a normal distribution, if the tolerance is say +/- 3 cents, a 6 cent spread, then 66% of the notes will be within +/- 1 cent; 98% will lie between +/- 2 cents, and 99.9999% will lie between +/- 3 cents.

 

Tol

+/- 66% +/- 98% +/- 99.99999% +/-

3 1 2 3

2 0.7 1.4 2

1.5 0.5 1.0 1.5

1 0.3 0.7 1

0.5 0.2 0.3 0.5

 

 

So an apparently loose tolerance, if centered on nominal can be surprisingly 'in tune'

 

However, and based on all of the above and earlier discussions: and subject to the condition of reeds plus the quality of the original build

 

AS A STARTING POINT:

 

1. Tuning should be accurate, centered within say a 0.5 cent of nominal through out it's compass

2. precision should be a further +/- 1 cents around the average.

 

all to be varied as customer and tuners ear dictates. However if not being tuned for a specific individual, then the tolerances and accuracy

shown above should be considered good practice???

 

In other words, if a player gets a concertina from a repairer, or dealer, and finds errors of 2 cents or more, he or she should feel empowered to take it back for correction? is this the advice we should be handing out, and be subjecting ourselves to?

 

Dave

Edited by d.elliott
Link to comment
Share on other sites

 

For the very simple reason that for me (with a "manual" meter, though not for Greg with an "auto" one), six passes take six times as long as one, or three times as long as two, all for the sake of a difference that isn't even audible.

By "auto" do you mean the type of meter that guesses what note is being played (with controls for reference pitch and temperament) and tells you the error?

 

FWIW the tuning work I've done so far is with a cheap but good Mac program called Katsura Strobe Tuner. It seems to work very well, though I think I'm now going to have to add the intervals for 1/5 comma Meantone to its temperament list! B)

 

The thing that for me would speed up the process the most is a way to note down the error value for each reed without having to put the concertina down and pick it up again between each note. I tried recording the process and speaking the errors out loud, then playing back the recording and transcribing them, but it wasn't really much faster overall. How do others deal with that issue? Recruit an assistant? ;)

 

 

This is one of the issues that lead to the development of Real Time Tuning Analysis (RTTA) for flutes. Others were more flute-centric - it's so easy to adjust flute pitch as you play that there is a real tendency to subconsciously do it when you should be trying to be brutally honest.

 

I don't think our RTTA systems (we now have two of them) will prove to be everything you need, but they might give you some food for thought. With flutes, our errors tend to be much greater than what can be achieved on the concertina (that's offset by the fact that the player can so easily correct them). We don't have strobe-tuner resolution - that would be a total laugh on the flute. But you might find an RTTA approach would serve for "disaster analysis", and for early tuning sweeps.

 

Further, we don't have the complication of alternate fingerings such as the Anglo has, or bidirectionality such as the English has. I guess you could find ways to get around those, eg play up the rows only on the Anglo, and do separate analyses on push and pull for the English.

 

Have a read at: http://www.mcgee-flutes.com/RTTA.htm. It outlines why I wanted an RTTA system, and then leads on to the two systems, the RTTA Polygraph, and Flutini. I'd suggest you download Flutini and give it a try - it will quickly illustrate the possibilities. The Polygraph is more complex, but is open source code based on an easy language, so it could be the basis of some further development. Scott, who developed Flutini, may be quite happy to make his code available too, I haven't discussed that. You might also want to go back to the Tartini package that is the basis for both our systems and work from there.

 

Happy to help in any way possible.

 

Terry

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×
×
  • Create New...