Tone deaf test, the kind that actually listens to you.

This one plays a note, then listens while you sing it back, and tells you how far off you were in cents. It takes about a minute. It runs in this tab, and nothing you sing leaves your machine.

Then find your range

Two sweeps, two minutes. It names your lowest and highest note and the voice type you sit closest to.

What this measures, and what nearly every other test measures

Search for a tone deaf test and you will get a quiz that plays two notes and asks whether they were the same. That is a test of perception, and the overwhelming majority of people who believe they are tone deaf pass it comfortably. Passing it and still being told you sing badly is a confusing place to end up, and it is where most people end up.

The reason is that singing in tune is two skills, not one. Hearing that a note is wrong is the first. Getting your voice to land on a note you can already hear is the second, and it is motor control rather than hearing. They fail independently, and the second one fails far more often.

This page tests the second one. It plays a target, listens to what you actually produce, and reports the distance in cents. A semitone is 100 cents. A trained singer holds a note inside about 10 of them.

Singing the right note in the wrong octave is not a mistake

If a target is played high and you sing it an octave lower, you have matched it. That is the same note. A man given a note in a soprano's range will usually and correctly drop it an octave, and a test that marks him wrong for it is broken.

So this page folds your answer onto the nearest octave before scoring it, and then tells you separately how many times it did that. The count is information about your voice, not a penalty.

Being consistently flat is a different problem from being scattered

Two people can average the same error and need completely different work. Somebody who lands 30 cents under every single target has one habit to correct, and it usually goes within a few sessions once they know. Somebody who is 30 cents out in random directions is not missing a correction, they are missing the feedback loop.

The result below separates them, because the average on its own hides the distinction that decides what you should practise.

What this page cannot tell you

It cannot diagnose amusia, the genuine perceptual condition that the phrase tone deaf properly refers to. Amusia is a problem with hearing pitch relationships, it affects a small percentage of people, and finding it requires a perception test rather than a microphone. If you score badly here and you cannot hear the difference when somebody else sings out of tune, that combination is worth taking to a professional.

It also cannot see your technique. Breath, larynx position and vowel all move pitch around, and a number in cents will not tell you which of them is responsible.

Where the number comes from

The detector is the same one behind the live monitor, the range test and the trainer. It used to read every note sharp, by 2.6 cents at a concert A and by nearly 30 at the bottom of a bass range, and the failure and its fix are written up with the measurements in my pitch detector read every note sharp. Reporting a cents figure from a detector with a known bias would have been worse than reporting nothing.

Each attempt is scored on the median of the confident frames, with the first part of every attempt discarded, because the start of a sung note is almost always a slide onto it rather than the note itself.

Nothing is uploaded

The microphone stream is analysed inside the page. There is no server, no account, no upload and no cookie for this test. Close the tab and it is gone.