Vocal Pitch Detector: See Every Note You Sing
Sing or hum, and watch your voice draw itself: the note you're on, the exact pitch in hertz, and how many cents sharp or flat, live, as you sing.
Live until you stop it. Nothing is recorded: the mic light goes out the moment you tap stop.
Octave lock
Low voices confuse pitch detectors: the mic hears an overtone and reports the wrong octave. Tell us roughly where your voice lives and we stop guessing.
Free · No signup · Your voice never leaves your browser
Stop watching notes. Sing them back.
A pitch detector shows you where your voice is. The trainer gives it somewhere to go: a five-minute daily call-and-response. Hear a phrase, sing it back, and get told exactly which notes landed and why the others didn't. Free, no account, in the browser you're already in.
What note am I singing?
Press the button above, sing or hum, and the detector names the note instantly: A3, say, the A below middle C. You also get its exact frequency in hertz and how many cents (hundredths of a semitone) you're sitting sharp or flat of the note's center. It listens continuously, so you can slide around and watch the trail follow you. Knowing the note is step one; knowing whether you can land on a note someone else picks is the real question: test your pitch accuracy when you're curious.
Why does my tuner show the wrong octave?
Because it's often not hearing your fundamental at all. Every sung note is a stack of frequencies: the fundamental plus overtones at multiples of it. Phone and laptop microphones roll off sharply at the bottom of that stack. For low voices the mic may barely pick up the fundamental, so the detector locks onto the loudest overtone and confidently reports a note an octave up. Sopranos hit the mirror-image problem at the top.
That's what the octave-lock presets above are for: telling the detector roughly where your voice lives removes the wrong octaves from consideration entirely, which is the cheapest and most reliable fix there is. One honest caveat: at the very bottom of the bass range, a consumer mic physically captures so little fundamental that any software's reading gets approximate. We mark that territory as approximate rather than pretending otherwise.
How many cents is “in tune”?
A cent is a hundredth of a semitone: there are 1200 in an octave. Careful listeners can just barely distinguish pitches about 5–6 cents apart; below that, differences are essentially inaudible. Within about ±10 cents, almost everyone hears “in tune,” which is why this detector marks that zone green. Past ±15–20 cents, most ears notice something's off; a choir or an a-cappella blend is less forgiving than a solo over a band.
So don't chase 0¢: sustained singing that stays within about ten cents of center is in tune, and even professionals oscillate around the target (that's what vibrato is). The number to watch isn't the last digit; it's whether you're consistently parked on one side.
Why does the reading jump around?
Four usual suspects. Breathiness: the detector needs a clear tone, and it holds its tongue (shows nothing) rather than guess at a whisper. Room noise: hums from appliances and background chatter can momentarily out-shout your voice. Harmonics: an overtone can flicker louder than your fundamental for a frame, which reads as a wild octave spike. Vibrato: a healthy vibrato really does swing ±25–50 cents around the note, several times a second, so a raw readout would flail.
This detector deals with all four: it only reports frames it's genuinely confident about, filters out single-frame spikes, smooths the big readout just enough to read calmly without lagging your voice, and won't flip the note name until you've clearly crossed into the next semitone. The trail behind the readout stays raw on purpose: it shows what your voice actually did, wiggles and all, while the big numbers tell you what it means.
Voice tuner vs. guitar tuner: what's different for singers?
A guitar tuner is built for one job: six known target notes, a clean plucked string, and a needle answering “sharp or flat of the target?” Your voice breaks every one of those assumptions. It doesn't announce which note it's aiming for, it moves continuously between pitches, it carries breath and vibrato, and, if you're a bass or a soprano, it lives in ranges guitar tuners weren't tuned for.
A vocal pitch detector flips the design: it identifies whatever note you're on rather than measuring against a preset target, it draws a continuous trail instead of a needle (motion is the whole point of a voice), it expects vibrato instead of flagging it, and it covers the sung range generously at both ends. Use a guitar tuner for a guitar. For a voice, use something built to follow one.
Frequently asked questions
Does humming work, or do I have to sing?
Humming works beautifully, often better than singing, because a hum is steadier and has less breath noise for the detector to see through. Hum to find a note, then open up to an "ahh" on the same pitch and watch whether it holds. If the trail stays flat through the switch, your placement is consistent. That's a genuinely useful thing to know.
Is my voice being recorded?
No, and it can't be. Your audio is analysed on your device, frame by frame, and each frame is discarded the instant its pitch is measured. Nothing is stored, nothing is uploaded, and there's no server listening. When you tap stop, the microphone is released immediately: you'll see your browser's recording indicator go out.
How accurate is this pitch detector?
The detection itself is the McLeod Pitch Method, the same maths serious tuning software uses. On a clear held tone it's accurate to within a few cents, which is finer than any human judgment of "in tune." The honest limits are physical: a breathy tone gives the detector less to work with, and consumer microphones capture very little fundamental at the extremes of the bass range, so readings at the very bottom get approximate. Within the normal singing range, what you see is what you sang.
Why does it show nothing when I whisper?
Whispers don't have a pitch: they're shaped breath, with no vibrating fundamental for a detector to measure. Anything that reports a note for a whisper is guessing. Give it any gentle voiced sound (a hum is plenty) and the readout comes alive.
Can I use this to tune an instrument?
Technically it will read any clear sustained tone, but it's tuned for voices: smoothing that expects vibrato, note detection that expects sliding, and range presets named after voice parts. An instrument tuner is better for instruments. This page is for the instrument you carry around in your throat.