Vocal Pitch Checker

Check if you are singing in tune with instant pass/fail feedback for every note. See your overall pitch accuracy percentage, a scrolling pitch trail against target notes, vibrato detection, and a voice range summary. Select a song key for guided practice. 100% browser-based, no downloads, no data uploaded.

Vocal Pitch Checker Tool

Understanding Your Results

Pass / Fail Badge

Each detected note gets an instant verdict: green check (PASS) if you are within ±20 cents of the nearest target note, or red X (FAIL) if you are further off. Twenty cents is roughly the threshold where most listeners perceive a note as "out of tune." When a song key is selected, the target is the nearest note in that key.

Pitch Accuracy Percentage

The large percentage is your running accuracy score — the fraction of all detected notes that passed the ±20 cent threshold. Above 80% is excellent, 60–80% is decent, and below 60% suggests the pitch needs work. The score updates in real time as you sing.

Pitch Trail

The scrolling canvas draws your detected pitch as a line moving left to right. Horizontal gridlines show the target notes (or notes in the selected key). When your pitch line aligns with a gridline, you are in tune. Deviations above or below the line show sharp and flat tendencies. The trail uses green for in-tune segments and red for off-pitch segments.

Vibrato Detection

The vibrato indicator monitors your pitch modulation rate. Natural singing vibrato oscillates at roughly 5–7 Hz. When the tool detects periodic pitch fluctuation in that range, the vibrato bar fills and the label reads "Vibrato detected." Slower fluctuation is labelled as pitch drift, and faster modulation as tremolo — both indicate areas to refine.

Voice Range Summary

The range bar tracks the lowest and highest notes you sing during the session. It shows the note names and the total span in semitones. This helps you understand where your comfortable range sits relative to typical voice classifications.

Session Stats

The stats grid shows total notes detected, percentage in tune, percentage sharp, and percentage flat. This breakdown helps you identify whether you tend to sing sharp or flat — the most common pitch issue for developing singers.

How Vocal Pitch Checking Works

YIN-Based Pitch Detection

The Vocal Pitch Checker uses the YIN algorithm, an advanced form of autocorrelation specifically designed for monophonic pitch estimation. Autocorrelation works by comparing a signal with delayed copies of itself — the delay at which the signal best matches itself corresponds to the fundamental period (and therefore frequency) of the sound. YIN improves on raw autocorrelation by applying a cumulative mean normalized difference (CMND) function that suppresses spurious peaks, dramatically reducing octave errors. An absolute threshold of 0.15 is used to accept only high-confidence pitch estimates, and parabolic interpolation between the three samples surrounding the best lag refines the estimate to sub-sample accuracy — typically ±1–2 cents for clean vocal tones.

Cents Offset Calculation

Once a fundamental frequency is detected, the tool calculates the offset from the nearest target note using the formula: cents = 1200 × log₂(detected / target). One cent is 1/100th of a semitone. The ±20 cent pass/fail threshold was chosen because psychoacoustic research shows that most listeners perceive pitch deviations beyond 20 cents as "out of tune." Professional singers typically stay within ±5–10 cents. When a song key is selected, the target frequency is the nearest note that belongs to the selected major or minor scale, rather than the nearest of all 12 chromatic notes.

Vibrato Detection via Pitch Modulation Rate

Vibrato is detected by analyzing the rate of pitch modulation over a sliding window. The tool tracks the instantaneous cents offset over the last 0.5 seconds and performs a zero-crossing analysis on the detrended signal. A zero-crossing rate corresponding to 5–7 Hz with a modulation depth of 20–80 cents is classified as natural singing vibrato. Rates below 4 Hz are flagged as pitch drift (unintentional wavering), while rates above 8 Hz suggest tremolo. This approach is computationally lightweight and runs in real time without requiring an FFT of the pitch contour.