How to Find the BPM of a Song

1. Tap along with the beat

The most reliable method for most music, and the fastest. Tap a key or button in time with the track and let the tool average the intervals between your taps.

Tap the kick or snare rather than the hi-hat. Hi-hats usually subdivide the beat, so tapping them gives you double the real tempo. Keep going for at least eight taps — two bars in 4/4 — and watch the number settle. The stability figure next to the result tells you when the average has stopped moving.

Accurate for: almost anything with a steady pulse.

Fails on: rubato, ambient material, heavily swung rhythms.

2. Count beats against a stopwatch

Older, slower, and still exact. Count the beats in fifteen seconds and multiply by four. Count for a full minute and you have the tempo directly, with no arithmetic.

BPM = beats counted in 15 seconds × 4

The weakness is that any miscount is multiplied by four. Counting 31 beats instead of 32 puts you out by 4 BPM. Longer counts reduce the error, which is why a full minute is the accurate version of this method.

3. Let software analyse the audio

DAWs, DJ software and dedicated analysers detect tempo by finding onsets in the waveform and looking for periodicity. This is accurate on grid-locked electronic music and much less so on anything played by hand.

The characteristic failure is half or double the true tempo, because an algorithm cannot tell which layer of the rhythm you consider the beat. Check any automatic reading against the typical range for the genre.

4. Read it from the source

If the track exists as a project file, a MIDI file, or a stem pack, the tempo is stored in it. DJ pools and sample libraries list BPM in their metadata. This is the only method that gives you the tempo the producer actually used, including any tempo changes across the arrangement.

Why readings come out at half or double

This is the most common problem, and it is not an error so much as an ambiguity. A drum and bass track at 174 BPM has its snare where an 87 BPM track would put it. Both numbers describe the same music. Which one is "correct" depends on what the genre convention counts.

If your result looks wrong, double it or halve it and compare against the genre ranges. That resolves nearly every case.

A worked example

Say you tap along to a house track and get 124.3 BPM across twelve taps, with a stability figure of about 6 ms. That reading has settled: house sits in the 120–130 range, the stability is tight, and twelve taps is enough to average out one mistimed hit.

At 124.3 BPM a beat lasts 482.7 ms, a bar of 4/4 lasts 1930.81 ms, and an eighth-note delay would be set to 241.35 ms.

In practice you would round to 124, because a producer almost certainly set a whole number. A tapped reading landing within a few tenths of a round number is good evidence that the round number is the real tempo.