How accurate is the BPM detection?

It works well for tracks with a steady, clear beat — pop, rock, electronic, hip-hop. It's less reliable for music with a loose or shifting tempo, since it's a statistical estimate of the most common beat interval, not a music-theory analysis.

What audio formats can I upload?

MP3, WAV, M4A/AAC, OGG, FLAC, Opus, and WebM files up to 100MB are supported. Everything is decoded and analyzed in your browser.

Why does the tool show three tempo candidates instead of one?

Beat-detection algorithms can't always tell a tempo from its double or half — a 90 BPM track can look like 180 BPM to the detector. Showing the top 3 candidates with confidence is more honest than presenting a single guess as fact.

Why did it detect half or double the real tempo?

This is called an octave error, and it's inherent to beat-tracking: the detector finds the strongest repeating interval in the percussion, but that interval can correspond to every beat, every other beat, or every second beat depending on the rhythm. If the top candidate looks off, check the other two — the real tempo is very often one of them.

What does the confidence percentage mean?

For tempo, it's the share of detected peak-to-peak intervals that matched that candidate's beat length — higher means a more consistent, clearly pulsed track. For key, it's a correlation score against known major/minor pitch profiles — higher means the song's notes line up more strongly with that key.

How accurate is the key detection?

Chroma-based key detection works reasonably well for tonally clear music with a stable key, but it can be thrown off by heavy distortion, atonal or ambient passages, key changes mid-song, or live/rubato performances where pitch and timing drift. Treat the result as a strong starting guess, not a certainty — the runner-up is often the relative major/minor of the correct key.