Relative Pitch Test

Play a C and then the E above it. Almost everyone can hear that the second note is higher. Rather fewer can say how much higher, and that is the whole of this test.

The distance between two notes is called an interval, and it has a name. Four semitones up is a major third. Seven is a perfect fifth. Twelve is an octave, the point where the note sounds like the same note again.

Those names are what the keypad offers, with the size in semitones printed under each one. You are not asked to know any theory. You are asked whether the gap you just heard is the same gap you have heard before under that name.

The perfect pitch test asks the opposite question. There a note arrives alone and has to carry its own name. Here the starting note is thrown away, and the distance is all that is left.

Every pair starts somewhere new, drawn from two octaves. That is not decoration. It is the one thing that keeps the test honest.

Fix the starting note and the task quietly changes. After a few pairs you would know that note, and naming the second one absolutely would answer every trial without hearing the gap at all. A test with a fixed root measures a mix of two skills and cannot tell you which one you used.

A moving root also blocks the other shortcut, which is holding a reference in your head from a song you know. You can still hum your way to an answer, and plenty of people do. That is genuinely relative pitch, so this page counts it.

The perfect pitch test gives you five seconds and no more. Given longer you can start from a note you know, walk up in steps and count, which is a different skill wearing that test's name.

Counting up in steps is what this test is asking for. Ruling it out would leave almost nothing behind. So there is no deadline here, and the time you take is reported rather than judged.

That second number is worth watching across runs. Two people can name the same share of intervals, one in two seconds and one in ten. The fast answer is recognition. The slow one is arithmetic, and it stops working the moment a tune moves faster than you can count.

The headline is the share you named exactly right. With all twelve names on the keypad, guessing scores about 8 per cent, so anything near that means the notes did no work. On the major scale setting there are seven names and guessing scores about 14 per cent.

The second figure counts every answer that landed within one semitone of the truth. It matters because the common confusions are neighbours: a major third for a perfect fourth, a minor seventh for a major sixth. A run that is often right and never far wrong is a different result from one that is scattered.

Answers are not marked as you go. Feedback would let you calibrate against the first few pairs and climb. The score would then measure how quickly you learned rather than what you arrived with.

For what a properly measured number looks like, a good digit span has one. Thousands of people have been tested one way, so a percentile there means something. Interval naming has no equivalent, which is why this page shows your figure and does not rank it.

One part of this is unusually clean. The notes are generated digitally, so the distance between them is exact whatever they come out of. Cheap earbuds do not stretch an interval, and Bluetooth delay, which ruins the auditory reaction time test, does nothing here at all.

What the page cannot control is the room, the volume and the noise around you. It also cannot control the setting you picked. Two notes played together are harder to separate than two played in sequence, so the two settings are two different tests.

It cannot control experience either. Anyone who has done ear training will score far above someone who has not, and that gap is practice rather than hearing.

This is not a measure of musical ability. Interval naming is one narrow skill inside music, and plenty of good musicians play by ear and have never named an interval in their lives.

It is not a hearing test either. If sounds are muffled or one ear seems quieter than the other, that is a question for a doctor and not for a web page.

Repeated runs will move this number, and it is worth being plain about what that means. You are getting better at naming intervals on a piano sound in a browser. That is a real skill and it is the only thing improving here. A good place to see the same argument made about a much older test is the piece on reaction time.