Am I Tone Deaf? Test Your Ear, Then Your Voice
Two minutes, two parts. First your ear: which way did the note move? Then your voice: sing it back and we'll measure how close you landed, in cents. Most people find out they're untrained, not tone deaf.
Part one is just listening. No mic needed yet.
Free · No signup · Your voice never leaves your browser
What does “tone deaf” actually mean?
“Tone deaf” gets used two very different ways. In everyday speech it means “sings off-key.” In science it means congenital amusia: a rare, inborn difference in how the brain processes pitch, where a person genuinely can't tell whether a note moved up or down, even when the jump is a full semitone or more.
Those are not the same thing, and the numbers prove it: the best large study (a direct count of about 20,000 people) puts true amusia at roughly 1.5%, about 1 in 50 to 1 in 65 people. (You'll still see 4% quoted; that figure comes from older, self-selected samples and is now considered an overestimate.) Meanwhile, around 10–15% of people sing more than a semitone off-key. The gap between those numbers is this page's whole point: most “tone deaf” singers have a working ear and an untrained voice.
Why this test has two parts
Hearing pitch and producing pitch are separate skills, and research shows they can fail independently: poor-pitch singers routinely score just as well as good singers on listening tests. Scientists call the common pattern “purely vocal tone deafness”: the ear is fine, the voice just hasn't learned to obey it.
A listening-only test can't see that. So part one checks your ear with the task amusia researchers actually use (pitch direction, not just “were they different”), and part two hands you a note and measures, in cents, where your voice lands. Only both halves together can tell you which kind of “can't sing” you have. One kind is rare and neurological. The other is common and trainable.
How the ear test works
Each round plays two notes and asks one question: did the second go higher or lower? The gaps shrink from a wide fifth down to half a semitone. The one-semitone rounds are the ones that matter: an untrained but typical ear gets those comfortably, while genuine amusia fails right there. We use rich, harmonic tones in a frequency band that phone and laptop speakers actually reproduce, because a test tone your speaker can't play isn't testing your ear.
There's no trick, no melody to memorize, and guessing is expected on the hard ones. If the subtle rounds come out a coin-flip, we don't flag anything. We offer a re-check, because nine times out of ten a coin-flip means a noisy room, not a struggling ear.
How the voice test works
You hum one comfortable note, and we build the test around it: no “select your voice type” dropdown, no notes you can't reach. Then a piano plays a target, goes quiet, and your mic opens: sing the note back, in whatever octave suits your voice. The same pitch engine that powers our trainer measures your note in cents (hundredths of a semitone) and tells you how far you landed and in which direction.
For calibration: professional singers land within about 20–25 cents of a target. Trained amateurs sit near 50. Most untrained voices land 90–130 cents out. That's the scale your result sits on, and every band of it is closer to “in tune” than people fear.
Because the microphone only opens after the reference note ends, the test never scores its own piano through your speakers, which is why speakers genuinely work here, not just “are tolerated.”
What your result can and can't tell you
This is a screening: a first look, not a verdict. Two minutes in a browser can tell you a lot: whether your ear separates semitones (the thing amusia takes away), and how close your voice lands to a target (the thing training changes). It cannot diagnose amusia, and it never will: that takes longer, controlled testing. If your listening result is rough, the first suspects are always speakers, background noise, and attention; headphones and a quiet room settle most of it. If it stays rough and you're curious, research-grade tests like the NIH Distorted Tunes Test go deeper.
One more thing, because it matters: your audio is analysed on your device, in your browser. It is never recorded, never uploaded, never stored. There is no server listening.
Frequently asked questions
Can tone deaf people learn to sing?
Most people who think they're tone deaf can, because they aren't tone deaf: their ear passes a listening test and their voice is simply untrained. Studies bear this out: the large majority of occasional singers land accurately once the conditions ease up (slower notes, less to remember), and pitch accuracy measurably improves with real-time feedback practice. For the rare person with true congenital amusia, singing accurately is genuinely harder, because the ear can't referee the voice. But that's about 1 in 50 to 1 in 65 people, and the two-part test above is designed to tell you which case you're in.
Is tone deafness genetic?
True tone deafness (congenital amusia) has a strong hereditary component: about 46% of first-degree relatives of someone with amusia are also affected. So if pitch direction is genuinely hard for you and it runs in your family, that's a real pattern, not your imagination. But keep the base rate in view: amusia affects roughly 1.5% of people. The far more common family inheritance is simply nobody ever practising, and that one isn't genetic.
Can tone deafness be cured?
The question hides two different conditions. Congenital amusia is a perceptual difference in the brain, not an illness. "Cure" is the wrong frame, and the honest answer is that evidence for training it away is thin. But the thing most people call tone deafness, singing off-key with a normal ear, isn't a condition at all. It's an untrained skill, like never having learned to throw. That one responds directly to practice with feedback, which is exactly what the trainer on this site does.
What's the difference between being tone deaf and just being a bad singer?
Tone deafness is an ear problem: you can't reliably hear which way a note moved, so you have no way to know you're off. Bad singing with a good ear is an aim problem: you hear the target fine, but the mapping from ear to voice hasn't been trained. Researchers call it "purely vocal tone deafness," and it's by far the more common case. The tell is exactly what this test measures: pass the listening half and miss on the singing half, and you're an untrained singer, not a tone-deaf one.
How accurate is this tone deaf test?
The voice half is measured in cents by the same engine that powers our trainer, and it's accurate to within a few cents, finer than any human judgment of "in tune." The ear half is honest about its limits: it's a screening, and a phone speaker in a noisy room can blur it, which is why we use tones speakers can actually reproduce, re-check borderline results, and never turn one bad run into a label. No two-minute browser test can diagnose amusia. Anyone who says otherwise is selling something.
Do I need headphones or a microphone?
Headphones sharpen the listening half but speakers work fine: the tones live in a range small speakers reproduce well. The microphone is only needed for the singing half, and only after the listening half is done; skip it and you still get your full ear result. Because the mic opens only after each reference note ends, the test never hears its own piano through your speakers.
Is there an age limit on this test?
A soft one: pitch recognition keeps maturing through the mid-teens, which is why the NIH's long-running listening test is restricted to ages 16 and up. Younger singers can take this test for fun, but treat results as noise until around 16.