Piano World Home Page Forums Our Most Popular Forums Piano Tuner-Technicians Forum So in tune that it sounds terrible

Yes, this is what happens when a piano is tuned only with respect to one partial alignment. It will sound clinical because indeed, everything has been clinically aligned. Thirds are usually perfectly progressive as machines create them, from analyzing the more stable 4th partial across the compass.

This does nothing to align the whole piano to its optimum point because other partials are being ignored. It will sound lifeless and dead, but will achieve a 100% on the exam.

Tunewerk, what is the alternative and best way/process/tuning method to achieve the highest quality and most musical tuning then, if a one partial alignment tuning leads to a clinical tuning?

Please can you also explain what you mean when you speak of a one partial aligning tuning?

Sure, I'll do my best to share my understanding. I agree alot with the late Virgil Smith on his whole tone approach, although synthetic and analytic tuning methods can both be used to achieve an optimum tuning.

A single partial alignment tuning is what most machines use. They look at a stable partial like the 4th across the spectrum, and extrapolate fundamental target frequencies from the ladder using the sample measured deviations. Units like the RCT and Sanderson Accutuner both use similar algorithms that just apply iH differences evenly between samples. They use scaling functions in areas like the bass and treble to apply statistical knowledge of what usually sounds best in these areas.

On a Kawai or Yamaha the partials fall pretty much into line. Here you can use a single partial alignment method and come out pretty good. I can't think of other examples right now, but those two kinds of pianos come to mind as having a pretty clinical spectrum, so clinical methods do okay. Machines and aural tuners who rely on 3rds can produce good results on these pianos.

On more musical pianos, different parts of the spectrum have different amplitudes and duration that really demand attention. Steinways are great for this - and I happen to think it is central to why they are so loved. They will have strong 2nd or 3rd partials across areas where the soundboard can create negative inharmonic effects, etc.

Using clinical methods on these pianos will create that effect where everything is technically in tune (listening to 8ves, 3rds, 10ths, etc.), but nothing is really in tune. The quality comes across - the best way I can think to describe at the moment - as a blender type background sound; a cacophony of beats that create a buzz-like background noise.

To create a piano that is both in-tune and full of vibrancy and life, all partials must be considered. So this means iterative methods must be used in the temperament to listen and adjust and correct. As notes are tuned, inharmonicity changes slightly.

If you look at the 4th partial then, at the end of a great tuning, it may not fall on a perfect curve. There might be deviations there to help the 3rd align in 5th intervals that demanded more priority. Looking at the curves won't help you find the ideal tuning.

Haye Hinrechsen did some research work that was the closest to this that I've seen. I think there was a thread here on his entropy reduction algorithms.

Tunewerk, your post is interesting and helpful.

Thank you,

what say tunewek is about a style of justness as obtained with ETD that make you follow one partial. (and tend to make somewhat dull tunings in some hands)

side effect of using them is that the ear (unconsciously) begin to focus only on a part of the tone

Then when unison are tuned a similar process arise and the tuner "stop listening" at some point.

("listening", I would say "experiencing the tone" be it with the hands that play or hold the lever, the ears and whatever other sense may exist )

in theory, even with a justness that is not really optimal, if the tuning is re worked at unison time, the result may not be obligatory dry and clinical. less consonant, and certainly not as coherent it could be, but an experienced tuner may be good enough to correct 'on the fly' when unison are tuned.
But that make a lot of operations, take time, the tuner may find it easier to look for a better ETD or to use it only "lightly", with some distance (and with tuning checks)

at some point the EDT get really not useful, or not for what it was intended first.It is still a measuring tool that can show drifts, but to take the control upon even a very good one is all but easy. only really experienced aural tuners may be able to do so.

Thanks Isaac

I would like to understand what mechanism is operating to "change" the inharmonicity of a string when tuning? I don't know any that would produce that outcome.

Perhaps this thread just goes to show that the growing popularity of UT's is a natural backlash due to the prevalence of sterile, perfect, machine generated tunings.

the one I know is more stress, but it goes for very minimal quantities of iH.
But I see that relatively conceiveable that a change in iH occur due to the way the strings couples are tuned, there may be some "absorption" of partials by the fundamental tone under some circumstances (?) .
The iH of the wound string is said to change depending of the hammer stroke force.

The resiliency of the couple tuning pin/front segment, have also a role in the spectra. Does it change iH I do not know but it can be less perceived (or more ?)

That is my hypothesis, and as long as unbalance stay unnoticed during the music, I can accept a somewhat older tone.

But very often, at some point something heard is less pleasing, disturbing to me, and I do not get waht it may add to the music.

Then other solutions have been find mostly based on a search for purer intervals, in a high iH context.
And that cause some impossibility, as at some point the iH gives a correction to the frequencies heard and the all intervals enlarge , in absence of any added stretch.

Naturally, the tuners try to have less motion in the top spectras, when tuning, hence the searches for iH based solution, so the spectra can be quiter in the instrument when intervals at the double and triple octave are played

I believe that iH play also a role in tone projection. the change of pitch generated create a simile travel motion for the ear. it allows the pitch to be heard better from some distance as it does not slope down in pitch as much with distance.

have no idea of the levels of attenuation, but the pitch of soprano section heard at the back of the concert hall is lower.

An admitted theory say that iH maintain the justness of piano and make the tone less boring. (with levels as 0.7cts for A49)

Then , using only a partial to determine the pitch of the fundamental in double and triple octaves cannot be precise enough (hence the better results obtaines with multi partial ETD's)
But even with thoses, the theory that was explained to me is that the piano spectra itself is too inconsistent to be used with similar optimal results than what trained tuners obtain.

A thorough "spectral analysis" of every note of the instrument may be conducted prior to the tuning computation, then a "less bad" solution is looked at.

By being trained to recognise consonance) the aural tuner goes the same direction, with flaws due to the impossibility to us all the tests availeable.

SO the pattern we use is based on one or 2 octaves and the rest is done "musically" (in my opinion)

Listening to beat comparaison between fast beating intervals at 3 octaves range gives no guarantee of the clarity or acoustical justness perceived by the musician. We just use there "recipes" that are known to be working.

Good question, Ed! I was wondering the same thing myself.

I would say that the more evident is the stiffening of backscale (and the string itself) as ih is influenced by tension and by the level of firmness its fixation.

May be what was told relates only with tension, in that case the the change is very little, for instance with 4 cts. Check it out on your favourite spreadsheet.

If the ih have less "obstacles" to fight, acoustically, (tuned couples of strings) may be it can be modified,

Ed Foote stated something important, which is that with a certain unison shape, the bridge is stiffened, (hence better energy transfer) , may be that are the strings that are stiffened, as a smooth road for the waves with minimal contrary interactions may change the elastic quality of the unison - the way it push back the hammer.

If there are for a time a lot of fight between waves certainly the wires can act erratically and they loose energy needed to send back tge hammer, as if they where less tight.

Before speculating on a mechanism it may be good to first check if the IH actually changes at all, which is easy to do: measure IH with, say, Tunelab five times, detune the string, and measure again 5 times. We repeat measurements to get an idea of the inaccuracy of the measurement.

Kees

Doel, what exactly do you mean by "detune the string" as far as how much? My understanding is that if a string is off more than 10-20 cents from its intended pitch, it is a given that the iH will change. If I come across a piano that is out more than this, I typically correct the notes before I sample them in RCT. The users manual suggests this also.

440>443 for instance.

OK, show the numbers then if you think it's "a given".

Kees

Since you suggested the test, why don't YOU run the test in accordance with your methodology and report back to us with your results. That way, you wlll be assured that the results have not been biased. And, even if you don't have an EDT, download it free for the purposes of this test.

You are right.

With Tunelabs model of IH I get for C3 .175 +/- 0.005.
When I tune the string 100 cents flat I get 0.200 +/- 0.005.

So the change in IH is .025 per whole tone. Withing Tunelabs inharmonicity model this amounts to for example a shift of 0.4 cents for the 4th partial. If the note was tuned 10 cents flat this would be a shift of 0.04 cent, for example.

Kees

WOW! - Op-Art in a tech forum.