After 50 years of being a professional classical musician, I purchased a new M&H BB (wood action) and decided to tune it myself. I have spent decades tuning harpsichords and clavichords and even the odd organ in numerous temperaments, but never a piano. After six months of aural tuning, I found that the piano was extraordinarily stable in tuning, but I was not happy with the stretch. The BB seems particularly rich in overtones throughout its range and I couldn't, not being an experienced ET tuner, find the right partial sets to stretch the octaves. I purchased Dirk's Piano Tuner, recorded every note, calculated the stretch, and tuned the piano using ear protection and tuning every string, including the unisons, by referring only to the software, not to my ear. Never, ever, in my life have I heard a tuning as clean, pure, and as my wife says, happy. The piano is in a 24 hour a day 71 degree temperature and 42 to 48 percent RH. It has been a week of 3 to 4 hours a day of heavy practice, and the unisons are still completely pure. I know this is setting the fox among the chickens, but I have always been overly sensitive to tuning errors and unpleasant stretches, and find that, for me, with one piano, this is a perfect solution.

Congratulations!!!

I just sent you a personal message. See the blinking envelope at "My Stuff"?

Welcome to Piano World!

Congratulation Mwm!
Do you have any commercial affiliation with Dirk's Piano Tuner?

None whatsoever. I found that aural tuning of my own piano was something that should be left to a professional tuner with decades of experience. I have always wanted to tune my own instrument however, and, while I realize that I am using a crutch, it allows me to spend more time practicing and playing.

As a aside, when the instrument first arrived, I decided to try an EBVT, which I really liked. However, my wife, a voice teacher, was not pleased. The problem being that much of the voice repertoire has been transposed, and a piece that was originally in D major, does not respond well to being moved to F Sharp. Too much shimmering, and not enough purity in the intervals, so I had to go back to ET.

With reference to my comment above regarding EBVT, it was wonderful to play the Debussy Suite Bergamasque, and the Préludes, many of which tend to haunt the keys of D flat and G flat.

Interesting!!!...

Interesting It is. She sings with a professional Baroque Choir and Chamber Orchestra. They use different temperaments for each concert, depending on the repertiore. As well, the orchestra tuning varies from 392Hz for French Baroque through 432Hz for Beethoven. What is really interesting though, is that the upper strings tune A and D to the continuo keyboard temperament (organ and/or harpsichord) and the continuo instruments tune all strings to the keyboard. The choir, however, sings in just (as much as it is possible) intonation. They did a recent Charpentier concert, and determined that within a single movement Charpentier used two keyboards, one with all sharps and one with all flats. There was no other way to play the movement with went from keys in four sharps to five flats. Not possible in a single French Baroque temperament keyboard.

Thanks for the witnessing. I noticed that the tuning "model" used by Dirk's software is probably the one that allows the most natural consonance, I have no precise idea why .

There is something in regard of the comptuation of justness , based on an iterative process toward the "less bad" interval with probably a preference set for partial match that is well organized. (priorities)

Also the choice of the partial used to drive the display (one partial only) is based on the stability analysed when sampling.

I suggest then, that the same string must be used for the samples and as the first string tuned.

Too bad you did not listen while tuning as you may have missed part of the expêrience, as if you listen you will understand how octaves and othe rintervals are preferred, and you could tune in phase with the software more easily.
That should be a little better in the end, it is evidentto me

I actually did listen, though with a 27db reduction in sound level, and I did miss part of the experience, but I have found that the volume level, when tuning, particularly in the A4 - C6 range, to be painfully loud. I very much enjoy listening to the partials in the lower registers.

I have a question regarding stretch and temperaments. I read a paper on using Shannon Entropy to determine, though an iterive process, the optimal stretch (compromise) which leads to a "pleasent" result. I don't think Dirk's tuner uses this approach, although he and his college, who worked with him to develop the software, clearly use some form of iteration. The question is - Does the particular stretch for a partuclar piano remain the same for all temperaments, or does it change with the temperament? My thought is that it must change, since the optimal compromise between, for example, a 12th and 17th stretch for a particular note would change if the partials were pure to begin with, as in some of the fifths in Kirnberger III.

I should correct my last sentence in my reply above. I meant to say that if you are tuning pure fifths, the upper partials would not be the same as when tuning narrow ET fifths. In a perfect ET with all intervals equal, the stretch would be used to nearly match one or more sets of partials, and in a UT, the partial sets would be wildly different from note to note, and should, therefore, require a different stretch.

I dont really know how this have to be dealt with.

To me as soon as your consonance is good you ar ein the good "stretch"

SO with "pure intervals" the answer should be easy, then with beating 5ths I dont know, in any case you are "obliged" to stick with consonant octaves, that serve as a skeleton for notes above, and are part of skeleton for notes under.

You have some beating in the octaves in the end, but the partials are then more near each others.

Stretch is just a mean to have more partials together with the fundamentals, hence a purer tone in the end. so it keep consonance and reinforce it, with more or less high partials tendency (differnt levels)

That iH diffence we have at each not is what allow us to cheat, you have nothing like that in harpsichords (but they can be tuned with a "modern ET temperament style and beating double octaves I heard that yet in a re"cording studio).

I understand the tone is too strong and tiring, then may be you cna try to listen to some part of it, or paly more softly, or simply use earplugs as you did. (I was supecting you did listen as it is in the end impossible to tune "visually")

Also the display is fast , so you are oblmiged to grasp on some consonance and tune therein my opinion.

Try to "listen with the playing hand" (focus on the hand/hammer stroke) it helps to quieten the ear ...

Dirks told me he uses some iteracy indeed.

Regards

Tuning aurally to achieve good consonance in the octaves makes sense whether or not you are tuning an ET. Perhaps the stretch does remain the same.

I don't see why it should. Both the approaches you mention work out the best set of intervals they can find by varying the frequencies of the notes by iteration. The Shannon Entropy paper said the result was an ET. In principle you could apply extra constraints to produce unequal temperaments but this would probably only result in local variations within the octaves.

Stretch is an unfortunate concept really because the basic aim of tuning should be to bring out the best from a piano.

I agree wholeheartedly that bringing out the best from a piano is the goal. Most Professional tuners stake their reputation on that premise. As a person interested in both the science and math behind the piano as well as the music (although I think it sometimes gets in the way of just enjoying the music) I want to be able to get the best from my piano myself, which in my case, was facilitated by an ETD.