Piano World Home Page Forums Our Most Popular Forums Piano Tuner-Technicians Forum Entropy Piano Tuner

Joe, are you indeed referring to major sevenths, or was the above a typo?

Dear Forum Members,

it is really a great experience to read your comments. We are very glad that you like the interface, but what surprises us most is that you write so positively about the resulting tune.

May be I should tell you how the whole story came up. Three years ago, when I published the idea about entropy-based tuning, I received reactions such as "Only a university of poor scribblers can produce such 'research'", and this led me to immediately stop all activities. The problem was that I had no means to verify the method. The situation changed last year, when I was approached by Michael Kohl, a piano student of at the local University of Music, who wanted to conduct a test of the method as part of his Bachelor thesis. The test was scheduled for April 2015. Recklessly I agreed, but this created an enormous pressure since I had to provide not only the theory and some code fragments but rather a functioning interface. Initially I tried it by myself and wrote a first Linux prototype of the EPT with a rudimentary synthesizer because this was the only way for me to actually hear the tuning result and to make sure that the whole test would not end up in a awful embarrassment.

In December Michael Kohl and me tried to tune the grand of a friend, the first real piano ever tuned by the method. We found it extremely difficult, it took us many many hours, but in the end the result was convincing enough to give green light for the large-scale test. In February my master student Christoph joined in. He convinced me to completely rewrite the code, reorganize everything in a modern professional way and to write a platform-independent application that can be used even on mobile phones. He is the one who created the graphical interface.

The test at the University of Music was carried out with Prof. Andreas Lehmann and Burkard Olbrich, who is the master piano technician at this institution. We were invited to his workshop and it turned out that he is also interested in the theory of tuning, compiling frequency tables and inharmonicity measurements. It was in fact a very nice cooperation. For the test Mr. Olbrich tuned two Steinway-C grands, the first one aurally and the other one with the EPT, and many of the difficulties he had in operating our application went into later improvements. The two grands were evaluated by 28 pianists, and a brief summary is reported on our webpages. The result is basically that the EPT creates an acceptable tune on a semi-professional level but of course it cannot compete with a professional aural tuning. However, for us it was already a big success to see that people do not run away when hearing the EPT and so we decided to publish the software.

This is basically the story. As you can see, our aim was primarily to prove that the idea published in the paper was not just nonsense, but we never intended to compete with existing ETD's. The whole development was completely independent, we do not own commercial ETD's or software packages, only few weeks ago we downloaded the free trial version of Tunelab on Android. Of course we saw the video about Dirks piano tuner and there are certainly similarities coming from the fact that all keys have to be recorded, but like you we have no idea how it works internally.

In the coming days I would like to ask you specific questions in different posts (please forgive me, I still have to learn the vocabulary, e.g. the two "diagonal sections" in my previous post are just the two bridges).

Best wishes,

Haye Hinrichsen

What would be more meanful is if more professional tuners do check the software.

ALfredo, you probably can do that assuming you have some time at hand, you certainly have at last a cell phone (I did use one, it works)

As I use an ETD is always with some distance, I mean I try to go the direction indicated by the software, which mean I tune almost as usual but try to agree with the software (I do not tune a bunch of notes then test, for instance)

WIth Entropy I could easily agree with the treble sectiopn, as I said, and the basses too where generally nice (some tones where not but assuming they are supposed to sound similarly as their neighbors the software can still be used)

I'd be curious to have comments from professional tuners

The last recordings from Grandpiano sound pretty well in tune at last from mediums to treble, as I could not listen on normal speakers, the basses are not at their best on my system
they seem to be in line, anyway

Regards

PS Even if you are a professional tuner, publishing something new on tuning create a lot of noise and rubbing.

Too bad that the sensationnalism of some journalist did some sabotage to the quietness of the reading.

Hi All,

Isaac, yes I do have an android and other items that I could use... Two other questions though, here in London I do not have a piano at hand,... I could use one in the showroom, though it would rise other issues.

And there is also the fact that, when I think about a test, I would like to use the best equipment, you know what I mean. Well,... forget all that. As you kindly suggest, I am going to download the android version, just hoping I get a good chance. Thank you.

GPM, It is so nice to "feel" your enthusiasm. Thank you for your latest recordings, to my ears they confirm what I have said. Let us know, now that you have heard a tuning where everything seems to "fit", what is it like with other kind of tuning?

Oh, pardon me, there is a misunderstanding. The intervals rec I asked for is meant to be (if possible) on a very fresh tuning with your best settings, and only middle strings (as much as possible). All together, playing and recording should not take you long, perhaps 20 minutes?

Regards, a.c.
.

Nope. No typo.

It's really not usable on all pianos, just very high quality ones, or the occasionally low quality one that somehow accidentally came out with an exceptionally good bass section.

I forget if you tune aurally. But if you do, try it sometime as a tertiary check when you are doing a very fine tuning.

Listening to them chromatically downward can help give that tiny bit of refinement to the low bass notes that may seem slightly ambiguous as to their exact placement. After all, an octave that's beatless still has a lot of wiggle room, especially in the bass where the frequencies are much slower. It's that extra 1% that makes all the difference.

But, I don't want to hijack this thread.

Hi Alfredo Nice if you give a try, I think that, as you are not used to tune with an etd, you will have the advantage of 'seeing' where the entropy want to go, and correct a little if you do not agree, in the end you probably will have then the best tuning possibme in the mood of the software.

that is how I use an etd, just for pointing a directoon. once we have the expected tone in the ear it is easy to continue.
tuning done with all usual tests so I do not give the control to the etd.

I think you will find that easier than ou think.

regards


I think unison sould be recorded with 3 stings or at last 2

pitch differences, beats differences, see what I mean

I will try these intervals etc. on this next tuning. The Ipad tuning has already drifted too much, lots of playing etc. My piano is notorious for not staying in tune etc. Several pros have said the same thing...it's a drifter.

I just started a new tuning file with the Laptop/Samson G Track mic. There are differences between the Ipad and laptop, so I will post the graphs later today, and start the re-tuning.

Alfredo, hard to put into words. It "feels" and sounds "right" to my ears. Pleasing, sweet, are words that come to mind. The piano as a whole sounds like everything "fits" in place. Even with the false beats that are now more prevalent, the hammers in need of voicing, and some sustain issues, the piano still sounds wonderful, and can you imagine what an expertly prepped pro piano would sound like with Entropy?

WT's etc have their place, so does ET, and for me, Entropy is going to be the one for ET.

This MIT article back in 2012, is a good one in explaining in laymen's terms, what Entropy is doing. The disbelievers at the end have been proved wrong..(they should listen to my piano!) I don't think it will spell the end for aural tuners, it's just another flavor in the cook's recipe box, however, done extremely well with a computer!

http://www.technologyreview.com/vie...-professional-musical-instrument-tuners/

Here are the Entropy graphs taken just now from the 2 day old Ipad/i436 mic tuning. It will be interesting to hear this, as there are quite a few differences.


1925 Mason & Hamlin BB 7ft Grand Piano

Entropy Calculation using "Infinite"

Inharmonicity Graph using Laptop/Samson G Track mic from the Ipad/i436 mic tuning




Recorded Graph using Laptop/Samson G Track mic from the Ipad/i436 mic tuning



Computed Graph using Laptop/Samson G Track mic from the Ipad/i436 mic tuning

I agree with Grandpianoman, namely that the entropy tuning sounds sweeter and better overall. For this reason I am not sure that going through individual intervals will do the tuning justice.

As an example, after tuning my piano with entropy I went through some checks and found a note that I believed was slightly off and thus I "corrected" it based on what a typical ET progression should yield. However, after playing some actual songs I found that I did not like the change that I made despite that change aligning more closely to a true ET.

In short, to my ears entropy makes the piano fit better together as a whole, not just as a series of individual intervals that fit a particular temperment.

Welcome to the forum, Staveoff.

Are you a technician? If so, it's considered good manners around here to identify yourself in your signature as a technician.

Quote:


Personally, I would like to hear how EPT handled these checks. I'm not suggesting that GPM retune the piano. But since GPM has a high quality piano, and he has so graciously volunteered to do some checks, I'm expressing interest.

If I'm the only one who's curious, then please GPM, don't worry about doing it.

Thanks,
-Joe

No problem Joe...Alfredo asked as well....I am curious how different the Laptop/Gtrack mic is going to sound, so I will re-tune and record the intervals, as well as the same 3 pieces.

Thank you, GrandPianoMan,

-Joe

Question regarding overpull:

Dear All,

reading your comments and searching the web I tried to understand what you call "overpulling". As far as I can see, one tunes a little bit too sharp in order to pre-compensate a future pitch drop. What is not yet clear to me is the origin of this phenomenon. There are several explanations around:

1) Intrinsic relaxation of a string after pitch raise due to the plasticity of steel under high tension. For example, a new violin string is known to decrease in pitch after installation. Is this also relevant for pianos?

2) Some kind of slip-stick relaxation of the tuning pin in the pin block, trying to equilibrate the residual torsion.

3) An overall deformation of the supporting structure (the frame) of the piano caused by the pitch raise.

To me the third explanation appears to be the most plausible one. However, this would require to know HOW the frame is being deformed during the pitch raise. In principle, if we knew for any pair of strings how much the increase of the pitch of the first one would decrease the pitch of the other, then it is in principle easy to set up an overpull theory in a linear approximation. I am going to write it up and post it here on the forum.

My question would be: Is there some empirical data available, telling us how much the pitch increase of one string induces a decrease of the other? Does it depend on the type of the piano and the frame structure? Is it okay to treat all strings within one bridge-section equally? If you have any ideas in this direction please let me know.

Best regards
Haye Hinrichsen

Haye

About strings as they are in theyr elasitic zone it could be that no plasticity occurs

but adding stress may have a little of that effect, I really do not know

The soundboard is may be more reacting than the plate in my opinion, (may depend of the plate thickness design and material)

it twist a little under stress, at the friction points(on the bridge) there is even with new strings a little residual tension that do not balance on both sides and the bridge tilt (it is very small but still can create trouble hence ths shapê used to stabilize the bridges and avoid excessive rolling motion

Basic pitch raise instructions, before sophisticated softwares but using ETDs where to compute 25% to 30% so the note is tuned 25%higher than its difference with original pitch

This simple process have limits, and can bring the wire above its elasticity range whenever the differenc is too high.

Pianos structure create slip in their middle mostly, a little less before the extremities, and almost no for their last octaves

mediums and octave 5 is where the pitch lower the most

When tuning aurally the principle is to pluck (new strings or long time neglected ^piano)
ALl A E, D, G, etc on the whole styructure , so the tension is added evenly and no "scissor effect" happens on the panel

(French name is "pincage" in engluish "chipping")
a high pitch is choosed , for instance 444 Hz with new strings

or 443 if I want to attain 442 (to simplify)

Then a rough tuning is done, fast but in a normal way


The ETD algorythm have the advantage they can record the nnext string to tune while the note had yet lost tension due to the tension added on the notes lower (usually the process is conducted from bottom up)

Pitch rise functions on ETD are computing a new pitch at each new note to be tuned, on the fly (but a basic tuning is indeed computed.


May be allowing to tune while sampling can be done, ? that would mean using the basic algorithm ? Or a basic "stretch" as a model

I have no idea of the feasability, but I understand this can be an interesting challenge

WHen sampling it would be enough to have a simple tuning display, (that would obviously make the recording less easy as the pitch moves instantly, but may be the saùpling process could be adapted to work on moving pitches and record only iH parameters ???

Best regards

Thanks so much for your involvment

Alfredo, let me suggest that you download Tunelab and process or verify the calibration of your cell phone.

Mine was about 3cts low, for instance.

I think of that because there is no calibration with EPT do you could have difficulties entering the wanted pitch in the tuning parameters.
Regards

Thanks for the warm greetings, Joe!

I'm not a technician. I'm just a guy who wanted to learn how to tune a piano so that I could enjoy playing on an instrument that is always in tune. I also wanted my young children to similarly enjoy listening and playing on a piano in tune all the time rather than waiting for the tuner to visit. A piano in tune is more likely to get played than one that is out of tune.

For me, playing a freshly-tuned piano is almost euphoric and even since I was a boy I would wait in anticipation for the tuner to finish so that I could sit down and play for hours enjoying the wonderful sounds coming from the instrument.
Unfortunately, there have been too many times when I had a tuner come tune my piano only to disappointed with the results. In short, I took matters into my own hands so to speak.

Let's just say I developed a great appreciation and respect for what tuners/technicians do. Plus, learning about the complexities of the piano and the physics of tuning has been fascinating.

I felt compelled to chime in about the Entropy tuner. I think that this program is fantastic and has the potential to revolutionize electronic tuning devices. The developers are also very accommodating and seem committed to making this program the best that it can be.

I will post intervals recorded. This on a recent Steinway A.

The original tuning 6month old) was similar to the computed one.

Interesting, there is a gain in coloration, beats are more lively, for instance 17ths exhibit a double beat that make the sound singing. It tend to hide too present fast beating too, which is an advantage.
Stretch is really not large but in extreme basses, but there I suspect the pitch just after the attack is not taken in account enough. Problem of having a display that works as the tuners ear, that point is not yet perfect, but the display can be used as a guide, while tuning aurally, or as a control once the note is tuned.

Here's another recording of a short piece by J. Brahms to add to the mix.
http://picosong.com/mh7Q

This was recorded on a Steinway B and tuned with the aid of the Entropy Piano Tuner.

Thanks Ed.

Would you mind to upload the tuning file to a site (Cloud, Google Drive, or other file sharing site?)

The iH data and the computed tuning is of some interest to me.

Is it a NY Steinway?

Olek,

Here's the tuning file of the Steinway B. Yes it's a New York Steinway.

http://www.fileconvoy.com/dfl.php?id=g0cecb163bb72a00e99969606442c42ce94a0b8539