How Instrument Tuners Actually Detect Pitch
It's tempting to assume a tuner just listens for "the loudest frequency" and reports that back as the note being played. A good tuner does something more deliberate than that — and the specific technique it uses is exactly why it can tell 440Hz and 442Hz apart instead of blurring them into the same reading.
Why tuners don't just look for the loudest frequency
One obvious approach is to break the incoming audio down into frequency bands and report whichever one has the most energy. The problem is resolution: at the buffer sizes practical for real-time detection, that frequency-domain approach simply isn't precise enough to distinguish two pitches that are close together — the difference between being perfectly in tune and being noticeably sharp can be a couple of Hz, well below what that method can reliably separate.
Autocorrelation: reading the waveform's repeating pattern
Instead, an accurate tuner analyzes the actual shape of the sound wave over time and measures how closely it matches itself when shifted forward by different amounts — a technique called autocorrelation. A pitched sound repeats at a regular interval; autocorrelation finds that interval directly by testing how well the wave lines up with a delayed copy of itself, and the delay that produces the strongest match reveals the note's true fundamental frequency. This operates on the raw waveform rather than a frequency breakdown, which is what gives it the finer precision a simple loudest-frequency approach can't match.
Why background noise doesn't trigger a false reading
Before any pitch is calculated at all, the incoming signal has to clear a minimum loudness threshold. Room hum, a quiet background, or the moment just before you start playing all sit below that threshold and are treated as silence rather than analyzed for a pitch — without this check, a tuner would flicker to random, meaningless notes any time the room wasn't perfectly quiet, even with no instrument being played at all.
Why a tuner turns off echo cancellation and noise suppression
Most apps that use your microphone — voice chat, video calls, voice recorders — leave your browser's built-in echo cancellation, noise suppression, and automatic gain control turned on, since those features genuinely help spoken audio sound clearer. A tuner deliberately turns all three off instead. That processing is built and tuned for voice, and it can subtly reshape the waveform in ways that interfere with exactly the kind of precise, repeating pattern autocorrelation needs to read accurately. Getting the rawest possible signal from the microphone matters more here than it does for a phone call.
How cents and note names get calculated
Once a fundamental frequency is found, it's converted to the nearest musical note using the standard equal-temperament reference of A4 = 440Hz, the same reference virtually all modern tuning is built around. The difference between the detected frequency and the nearest in-tune note is expressed in cents — hundredths of a semitone — which is why a tuner can show you "how far off" you are rather than just which note is closest.
Our Online Tuner runs this exact autocorrelation-based detection live from your microphone — no account or software install needed, and nothing you play is ever recorded or uploaded.