Piano World Home Page Forums Our Most Popular Forums Piano Tuner-Technicians Forum Novel Piano Tuning App

Ron, thanks for testing the app! You are not alone with your request for a clearer marking of the zero. So I will look into it.

Non-equal and custom temperaments are the first features on my list once the app is stable and v1.0 has been released to the app store. Do you have any favorite temperaments you would want to see included as defaults?

If the unisons are in tune or only slightly out of tune, you do not need to mute them. In fact, the IH measurements you get from tuned unisons are better suited for calculating a tuning because they are an average of the involved strings and so the resulting tuning will not only respect the characteristics of a single string. But my algorithm is not able to derive this average if the unison is strongly out of tune so in this case you are fine with muting. I should make this clearer in the manual.

What do you mean by "precision" curve? Tuning curves for different tuning styles may be similar. This strongly depends on the instrument. Have you tried playing with the weights yourself? You can use a slider to increase and decrease weights for certain intervals and see how the deviation and tuning curves change. The deviation curves are more sensitive and interesting than the tuning curve. Check out the deviation curves for different tuning styles.

pianoscope always adjusts its tuning curve to an individual instrument. (In theory you could measure one instrument and use the tuning curve to tune a different instrument, but why should you?) However, there is the "Weighting" parameter in the settings of the inharmonicity view which lets you decide, how neatly pianoscope should follow jumps in the inharmonicity, like with poorly scaled pianos. https://www.pianoscope.app/manual/en/pianoscope.html#_curve_settings
If you use a small weighting value, a more idealized version of the IH curve will be used and the resulting tuning curve will become smoother. You can activate the checkbox to show the "Resulting Curve" to visualize what this parameter does.

:-) I have designed pianoscope for iPhones and iPads. If it is successful, I might one day port it to Android or other platforms. But I currently do not have any plans. If you are a lucky owner of a brand new Apple Macintosh with an M1 processor, you could get pianoscope to run on it. I have not tried it yet, so I am curious myself.

Jean, in build 369 when you tap the instrument name I am now showing an alert which can take you to the document browser in which you then can rename the document. Sadly, Apple currently does only allow renaming existing documents from within the document browser.

Frank, I use a mild well temperament of my own design that I scale up or down - my daily use one is an ET replacement has a maximum offset of only 1.3 cents, but that is just enough to feel a difference. A few teachers prefer a stronger version with a maximum offset from ET of 2.1 cents. They are graphed on the Rollingball.com site under the Koval variable temperaments.

So short answer - for me, I just need to be able to enter a couple of custom temperaments!

Thanks for the fine work, I haven't run into any major problems so far. It is more just getting accustomed to the file and naming protocols!

It might also be a good idea on the tuning screen to show which stretch is selected just as a confirmation to those switching back and forth from 12ths to balanced to octaves...

Instead of Hz under the note name for the curve, is there a way to display a cents deviation from an unstretched tuning? I suppose you'd have to pick a partial or just use the Hz to create an offset for the fundamental?

Ron Koval

So are a set number of periods of each of the partials derived by the fourier expansion compared to the elapsed time for them to accumulate?

Ron, I am glad you like it. It is a good idea to show also the tuning target as an offset from ET in cents. I will include it or add an option. And I will check if I can find a nice place for showing the selected tuning style.
And thanks for hinting me to rollingball.com. What a great web site. I did not know it before.

Ed, I am inclined to answer yes, but to my mind your question has a slight tautological touch. Perhaps I can roughly sketch out the principle in other words to clarify:

The analog/digital conversion of the audio circuits in the iPhone or iPad is the master reference clock for all measurements in pianoscope. It samples the air pressure change detected by the built-in microphone 44100 times a second. In every moment, pianoscope takes the most recent ~0.3 seconds of these samples. (The exact analysis duration varies with the tuned note, but this is not important for this argument.)

A Fourier analyis is now mathematically applied to them. It results in thousands of sinusoidal partials which when summed up again reconstruct the the original signal. The set of amplitudes of these thousands of partials is called the "spectrum" of the signal. The spectrum inherits its precision from the precision of the original A/D conversion. But not all of the spectrum amplitudes are relevant for tuning. They also contain effects from noise, background sounds and ambient reverb. But the actual musical partials of a piano tone show themselves as sharp peaks in the spectrum. pianoscope now analyzes the peak structure and applies a physical inharmonicity model to discern interesting piano tone peaks from background noise peaks. pianoscope might not pick up all the partials of the string vibration, because the soundboard might not amplify some of them or their frequency might be too low for the built-in microphone. Nevertheless, there is enough information in the partial pattern to reliably derive an inharmonicity and a base frequency from them.
The human hearing does a similar thing. You can listen to Wagner through a small radio speaker which does not reproduce any frequencies below 100Hz, but you can still musically discern the pitch of bass notes. They may sound thin, because the lower partials are not there, but your sense of pitch is still working even with a sparse partial spectrum.

Thank you for answering. This is what I suspected. You are using a "modeled" period average.

Can your system be set to only measure within a narrow frequency band with period average? To only measure a particular partial and to have the readout in digital cents from pure ET?

Internally, I am already having this information. I currently do not present it in the user interface this way. I combine the information from multiple partials into a single weighted average which drives both the scale indicator AND the strobe.
I could add a mode for focusing on a particular partial in a future version. TuneLab does a similar thing. What would be your exact use case for it?

Congratulations Frank. I'm impressed by the polish and attention to detail. I think you have avoided many of the mistakes that I made when I was first starting out. One thing I'm tempted to copy is your idea of having a free trial of full functionality. My biggest source of 1-star reviews is people who install the app thinking it's free and then realizing they have to pay money to tune their piano. I've tried everything I can think of to make it clear upfront the app isn't free short of changing it to a pay-up-front app (which is impossible after the app is published as free with in-app purchases) but people still end up feeling tricked. Your app softens that blow by actually giving them a free trial period.

I would like to add informaton about the new App to those featured on the Electronic Tuning page of my website https://www.davidboyce.co.uk/electronic-tuning.php
Would it be OK to do that, Mr. Illenberger?

I have an M1 Mac but I don't think you can download apps that require Testflight to run. I'll be happy to try it out later though, once it's officially on the AppStore.

Paul

Thank you, Anthony! We are all standing on the shoulders of giants and I took a lot of inspiration from your fine Pianometer.

I can feel your pain with the broken in-app purchase and review system in the Apple App Store. In my main job, I have been selling project management software for Mac and iOS for 17 years now. Apple is treating paid but free-to-download apps as if they were completely free. It would be easy for them to require a purchase before you are entitled to review an app. As a free-to-download app you even have to share the download charts with completely free apps which does not make any sense. Switching to a time-based trial won't completely alleviate these problems, as my experience shows. But there are other good reasons for time-based trials so you should give it a go. The best remedy we found to reduce silly 1-star reviews is a very short message in all caps at the beginning of the app description in the app store, like this:

*** FREE TRIAL FOR ONE WEEK WITH FOLLOWING IN-APP PURCHASE ***

And additionally you can encourage your happy customers to write real reviews. There is an API in iOS with which you can prompt the user three times per year to give a rating or to write a review (see SKStoreReviewController). It is up to you, when to call this API. You can choose a moment, when the likelihood is high that the user is in a good mood, like after he/she has completed the tenth tuning with high precision or so.

David, you are more than welcome!

Thanks Frank. I had put this message at the top of the app store description (2nd sentence) with little effect.

I put a similar message as a little popup the first time people open the app, but that seemed to make things worse.

I never put a review prompt, though I've thought of doing that. The main thing holding me back is that I don't like getting those myself. Some apps seem to ask you every week without giving you a way to stop it. But if it had 3 buttons with "Yes", "Maybe later", and "Don't ask again" that would be reasonable.

I think your App Store message is far too long to solve the problem. You could try a much shorter variant.

Apple is strongly limiting the frequency of the review prompts. You can call the API a hundred times, but Apple will only prompt the user a couple of times a year. And I think they also balance it across all apps to avoid annoyance.

I didn't "feel tricked", but I first took a note of Pianometer here and on your website, before going to the appstore. So I was informed. Maybe you could rename the free version to "Pianometer-trial" e.g.

Well, it is partly a philosophical interest. I am a professional who is capable of deriving the best partial/intervallic balance across the compass of a piano by aural techniques alone. I will never cede authority to the makers of tuning instruments for pianos about how to define an "in-tune" state.

The main advantage a tuning instrument offers me is if it is able to resolve fine increments of pitch rapidly, say 0.3 cents. Once I know where a note needs to be, giving it a test blow and then being ale to resolve if it changes more quickly than the aural test can reveal would be an advantage. Strobe displays resolve fine increments slowly because you must wait for the pattern to move.

There are plenty of beats to resolve rough pitch plenty quickly, it is the slight settling that takes time to measure with a strobe display. Plus the strobe display is too sensitive in the upper treble compared to the accuracy required.

Just visited the website. Terrific!

JG