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

There is a free Google Play app called "Entropy Piano Tuner."

https://play.google.com/store/apps/details?id=org.entropytuner.app

It appears to be FOSS.. The description makes it seem similar to Dirk's Piano Tuning software.

From the web-site:


I downloaded it to my smart phone. But, I haven't had time to play with it yet.

Just out of curiosity, has any here tried this yet?

Whoa! I just found there's a GNU/Linux version on the University of Würzburg web-site:
http://www.entropy-tuner.org/

Yes, the software is GPL'ed.

It's available for

iPad

Android

Windows

Mac OS-X

Linux

and the source is there, too.

It takes too long to measure each string!

About this Entropy Piano Tuner software (EPT), see also the thread Octaves are UNISONS, i.e. the post of Withindale, June 13, 2015, and further comments.

That's very true, and is why the other software
packages like Veritune, Cybertuner, or Tunelab would
be faster when tuning out in the field.

However, for your home piano, or for a piano which you
anticipate you will tune more than once, it could very
well be worth the extra time to measure all the notes,
because you only have to do it once. And then, if the
discussion on the previous thread is correct, you will
have a tuning that is closer to an actual good aural
tuning, if we assume the entropy algorithms do indeed
mimic what the human ear does.

I'm going to try this software soon.....

if you have three hours time to tune a piano, maybe....

this is the paper
http://arxiv.org/abs/1203.5101

What I would love to know is how this entropy tuner does with a real piano...As soon as I can, I will tune my 7ft M&H BB with it, and post some piano pieces.

Its good news available in all major platforms. Hope it can tune historical temperament.

I downloaded it and tuned a couple of grands (1 small, 1 medium length) with it! Other than the extra time to measure each string, I was pleasantly surprised at the end result!

Tried it last night. I really like the interface (graphs and recording quality meters for each key, etc.). That would be great if it could also offer custom temperament values and maybe more choices for intervals/partial ratios.

Very well done, congratulations to their team!

The math underlying the entropy-based tuner assumes that every interval on the piano (all possible combinations of 88 notes, including 88 notes played simultaneously) is equally important, an assumption that only occurs in ET.

The authors attempted to change the weighting by working in subsets of octaves, fifths, and fourths.

"Instead of adding up the spectra of all piano keys,
we tried to work with subsets of octaves, fifths
and fourths, imitating the practice of aural tuners.
This destabilizes the method, probably driving
the pitches out of equal-tempered into just
intonation. Apparently the summation over all
keys allows the system as a whole to stay in equal
temperament."

A different mathematical approach would have to be developed to create an entropy-based UT tuner.

prout

Very cool.

I've been using an older free version of tunelab to try to tune a vertical of mine which has a very short scale, and so it's very difficult to get satisfactory results in the tenor and bass octaves.

This, on the other hand, sounds pretty stellar, at least in concept. I'm doing some action work on said piano at the moment, but I'm very eager to give the program a shot as soon as that work is completed.

Also, FOSS?? No more pauses every 10th note for the tunelab demo, "PLEASE BUY OUR SOFTWARE NOW" timer, even though you're completely broke?? Yes please.

I still haven't had a chance to put the Entropy Tuner through all its paces.

There are two things I did notice, though...

There's no way to calibrate the software. However, there is a workaround.

When you set the pitch, you can set it to the calibration for any other software you are running.

For instance, I have the Verituner software installed on m y notebook. The offset is -0.13. I checked and the equivalent frequency for -0.13 and it came out to A=439.9, I "think". (I'm not sure if that's exactly what it worked out to; the netbook is out in the car right now and I'm too lazy to walk out and get it.)
The point is, you can set the frequency to compensate for the soundcard.

2)The other thing that seems to be missing are the overpull targets. For that, it will probably be necessary to guesstimate the overpull.

One more thing...
Since this software requires a higher quality mike that what's already built into most PCs, maybe it would be possible to do quality unison tuning one string at a time matched to the display???

Before I started to sample the notes of my M&H BB yesterday, I noticed there was an upgrade for Entropy...I clicked on it, but it failed to install, subsequently, Entropy was not loading. The fix is that you must un-install it first, then do the upgrade.

I will have more time next week to do a full tuning of the M&H and will post a few recordings.

Interesting, I thought the upgrade was to version 1.1.4, but checking now. it's still ver 1.1.0.

Hi GPM,

For one reason I believe that this software might prove to be of value, and that is because "lower entropy" can be understood as "higher order".

When ever you are ready to record your tunings, may I dare to ask you something? May I suggest a procedure?

First, use your usual/favorite ETD and tune your M&H; make a record, so that aural tuners (like me) can hear the result: from C3 to A4, octaves, thirds, fourths, fifths, 10ths and 12ths; from A4 to C7, octaves, fifths, 10ths and 12ths. Play those chords chromatically with the same energy, record 3-4 seconds each.

Then tune your M&H with Entropy and make the same recording as above.

Oh,... It is only a wish :-)

My best regards,

Alfredo
.

hello all; the last version have an improved graph display, and ability to see all the data in excel

ih, recorded pitches, computed pitches, then re recorded pitches when you have tuned the piano, with diff in cts

Dear Alfredo, as the algorithm compute every time a slightly different tuning (for an unexpected reason)

I have yet asked them if a possibility to choose the "fundamental tuning " used to drive the computation into a direction, or to put limits, could be modified from 6 3 tuning to a 2 1 or 3 1 or whatever scheme would allow different final computation;

hoping this can be investigated

at that point there is a lot of strenght given to ih in the tuning , that sound nice, but I am unsure that a strong , consistent harmonic scheme is part of the final tuning ; probably no

best regards

ps basses and treble sound very clean

Hello Alfredo,

Will try to accommodate your request...certainly can do that after it's tuned with Entropy, but not sure I will have time to re-tune the EBVT III that is on there now.

Hello Isaac,

What do you mean by "but I am unsure that a strong , consistent harmonic scheme is part of the final tuning ; probably no" ?

That mean I am not sure the final pitches correspond to a same "consonant node" as when tuning with a balance of partial matches, as with Chas, or a direct match as pure twelve, 5ths or a 6:3 4:2 balance.

I think that each note is sounding clear and pure, normally all chords and intervals should also have an identical balance of sounds, but I am not sure it really happens, may be because of the iH inconsistencies, or may be only because iH progression is often strong.

I have no idea how much the algorithm can be "driven", borned, kept in a range of possibles.
The idea would be to understand how much the first theoretical computation is influencing the result, and if it is significant, allow different models choices at that stage.

The view of the data in Excel is cool and allow experiences