How to Test Tuner App Accuracy with FakePiano
You can test any tuner app’s accuracy in just a few minutes by using FakePiano to provide a known-exact pitch reference. While many tuner apps claim high precision, verifying those claims on a physical instrument is nearly impossible because strings drift, wood breathes, and human technique varies. FakePiano is a “ground truth” source that doesn’t change: a digital oracle that plays exactly what it says it is playing.
In this guide, we will show you how to use FakePiano (available for free at applicaudia.se/fakepiano) to benchmark any tuning app or tuning gear.
The Problem with Physical Instruments
If you try to test a tuner using a real guitar or piano, you introduce dozens of variables that have nothing to do with the tuner’s implementation. For example, a guitar string begins to go flat the moment you pluck it as its tension decreases.
A physical instrument is simply too “noisy” to verify that claim. You cannot use a moving target to measure a high-precision tool. You need a source that is mathematically perfect and stays steady for as long as you need to observe the reading.
Introducing FakePiano: The Tuner’s Ground Truth
FakePiano is a free web-based tool; a deterministic test oracle designed specifically for tuning analysis and DSP (Digital Signal Processing) verification. FakePiano simulates the complex, inharmonic behavior of real acoustic piano strings.
You can find it at applicaudia.se/fakepiano.
FakePiano is monophonic (it plays one note at a time) to ensure the cleanest possible signal for your tuner to analyze. It is not meant for playing music… it is meant for testing tuner apps and tuner gear.
It lets you select between different “piano profiles” (detailed in a later section). Because every tone generated by FakePiano is calculated with mathematical precision using established physical models, you know exactly what frequency is produced. If FakePiano says it is playing an A4 at 440.00 Hz with a specific inharmonicity coefficient, it is doing exactly that.
Key Features of FakePiano
- Monophonic Output: to provide a clear, single-frequency path for testing.
- 8 Piano Profiles: switch between levels of inharmonicity to see how your tuner handles real-world complexity.
- Cents Jogging: a high-precision control that lets you shift the pitch by as little as 0.1 cents.
How to Test Your Tuner App (Step-by-Step)
To run a valid accuracy test, you will need two devices. Device A will run FakePiano to generate the sound, and Device B will run the tuner app you are testing. Using two devices prevents internal audio routing issues that can sometimes occur when running a generator and a tuner on the same hardware.
Step 1: Set Up the Environment
Open a web browser on Device A (a laptop, tablet, or phone, or just another browser window) and go to applicaudia.se/fakepiano. Ensure your volume is at a moderate level, loud enough for a microphone to pick up, but not so loud that the speakers begin to “clip” or distort the waveform.
Step 2: Enable Audio and Select a Profile
Tap anywhere on the FakePiano interface or press a key to enable the audio engine. By default, it will load the “Upright” piano profile. This profile includes a realistic amount of inharmonicity. For your first test, you might want to switch to the “Ideal” profile.
Ideal profile removes all inharmonicity and provides pure mathematical harmonics (2x, 3x, 4x, etc.). This is the easiest “win” for a tuner app and serves as a good baseline.
Step 3: Prepare Your Tuner
Open any tuner app or device. Hold the microphone of Device B close to the speaker of Device A. If you are in a noisy room, try to move to a quieter space to ensure the tuner only hears the FakePiano signal.
Step 4: Play a Reference Note
Tap a key on the FakePiano keyboard (for example, C4, middle C). Observe the readout on your tuner app. A high-quality tuner should identify the note almost instantly and show it as perfectly in tune (0 cents offset). If the tuner takes several seconds to “settle” on a stable reading, it may have a high latency or a slow FFT (Fast Fourier Transform) window.
Step 5: Test the Cents Jog Buttons
On FakePiano, use the Cents Jog Buttons to dial in a specific offset, such as +5.01 cents.
Your tuner app should reflect this change immediately. If the tuner shows +3 cents or +7 cents when the source is exactly +5.1 cents, you have found a significant accuracy error. Even more telling is if you move the pitch by 0.1 cents (the smallest increment) and the tuner does not react at all.
Step 6: Test Across the Range
Real-world tuning isn’t just about the middle of the keyboard. Play a very low note (like A0, the lowest note on a standard piano) and a very high note (like C8). Many FFT-based tuner apps struggle at these extremes because of the resolution-vs-latency tradeoff in their analysis window (see our 0.1 cent explainer). A strobe tuner like StroboPro is designed to hold its 0.1 cent reading across this entire range.
Understanding the Piano Profiles
FakePiano includes 8 named presets. These profiles change the “B-coefficient,” which determines how much the overtones of the string are stretched sharp. This is a critical test because a tuner that can’t handle inharmonicity will give you a “correct” reading for the fundamental frequency that actually sounds out of tune when compared to the rest of the instrument.
| Profile | Inharmonicity Level | Typical Instrument |
|---|---|---|
| Concert Grand | Lowest | 9-foot Stage Piano (Long strings) |
| Studio Grand | Low | 6-foot Parlor Grand |
| Baby Grand | Medium | Small Home Grand |
| Upright | Medium-High | Standard Vertical Piano (Default) |
| Console | High | Compact Vertical Piano (Short strings) |
| Spinet | Highest | Very Small Vertical Piano (Highest stretch) |
| Other | Custom | Extra Preset for Testing |
| Ideal | None (B=0) | Pure Sine/Harmonic Tones |
The “Ideal” profile is perfect for verifying the raw frequency detection of an app. The “Spinet” profile is the hardest test for a tuner’s ability to track pitch through extreme inharmonicity. If an app stays accurate on a Spinet C1 note, its DSP engine is robust.
What Good vs. Bad Looks Like
When testing your apps, keep these benchmarks in mind. Most “needle” style tuners have an internal resolution of around one cent, sacrificing precision for speed. If you move the FakePiano jog wheel by half a cent, a typical needle tuner won’t react at all, or will just jitter in place.
StroboPro, however, is built on a high-definition strobe engine. Our product specification is an accuracy of less than 0.1 cents, and we have achieved less than 0.01 cents in a lab environment. When you use FakePiano to feed StroboPro a signal that is shifted by exactly 0.1 cents, you should see the strobe pattern on the screen begin to drift slowly to the right (if sharp) or left (if flat). This visual movement confirms that the engine picks up that microscopic change in pitch that a needle tuner would miss.
The One Test That Separates Real Strobes From Strobe-Look Displays
FakePiano can also answer a question most tuner marketing never addresses: is the strobe display on screen actually measuring your harmonics, or just animating them? Many tuner apps estimate a single pitch per instant, typically with an autocorrelation-style algorithm such as YIN, and a growing number of apps dress that single number up as rotating rings. If the rings all come from the same single estimate, they can only rotate in lockstep, no matter what the overtones are doing. It looks like a multi-harmonic strobe; it isn’t one.
Here’s the test, and it takes about a minute:
- Play a note on the “Ideal” profile (zero inharmonicity). Any decent tuner looks good here; that’s your baseline, not your verdict.
- Switch to the “Spinet” profile and play a mid-range note. Its overtones are deliberately stretched away from exact integer multiples of the fundamental (real string physics).
- Watch the display closely.
On a genuine multi-harmonic strobe tuner like StroboPro, each harmonic ring is phase-compared against its own reference, so the rings respond to the stretched overtones individually: they drift at different speeds, some in opposite directions. That disagreement is the measurement: it is exactly what piano technicians use to set stretch. On a strobe-look display driven by a single pitch estimate, the rings stay locked together, or the app simply shows one averaged reading. Rings that cannot disagree with each other are not independently tracking anything.
This matters beyond piano work: intonation setup, unisons, and diagnosing odd-sounding low notes all rely on seeing what the overtones, not just the fundamental, are doing. For the underlying mechanism, see our posts on independent harmonic rings and what a strobe tuner really is.
Why Accurate Bass and Treble Matter
Many tuners claim high accuracy but fail in the “extremes.” At 30 Hz (the low end of a piano or a 5-string bass), the distance between notes is only a few Hertz. For example, the difference between A0 (27.5 Hz) and Bb0 (29.14 Hz) is only 1.64 Hz. At these low frequencies, a small error of 0.5 Hz in the app’s math can translate to roughly 30 cents, which is why many bassists find that their tuners “don’t work” on the low B string.
Similarly, in the high treble, the frequencies are so high that many phone microphones and app algorithms “roll off,” losing precision. The harmonics of a C8 note can reach well above 10,000 Hz. If an app isn’t sampling the audio correctly or isn’t designed to handle these high-frequency harmonics, its accuracy will plummet.
Using FakePiano, you can verify if your tuner holds its ground at the very top and bottom of the scale. If it tracks a low A0 steadily and accurately, without jitter or drift, you have a professional-grade tool in your pocket. (StroboPro is specified to 0.1 cent accuracy even at these low frequencies.)
Conclusion
Don’t take a developer’s word for it: test it yourself. The tools to verify audio claims are now available to everyone. By using the monophonic, deterministic tones of FakePiano, you can remove the guesswork from your tuning setup.
Whether you are checking a $500 hardware tuner or a $5 app, the methodology stays the same: compare the device against a known ground truth. FakePiano provides that truth with 0.1 cent precision and realistic piano physics.
Head over to applicaudia.se/fakepiano today. Open your favorite tuner app, dial in a 0.01 cent offset, and see if your gear can actually hear the difference. If it can’t, it might be time to switch to StroboPro – try it free at strobopro.se, or download the iOS app or Android app.
Leave a Reply