Auto-Tune and pitch correction use physics and signal processing to adjust a recorded voice toward desired musical notes. A singer’s pitch is related to the frequency of vibration of the vocal folds, so higher notes have higher frequencies. In a recording, software can estimate that frequency moment by moment and compare it with a musical scale.
This matters because it helps engineers fix small tuning errors, create stylized vocal effects, and understand how sound becomes data.
Key Facts
- Pitch is mainly determined by frequency: higher frequency means higher pitch.
- One octave higher means double the frequency, such as 220 Hz to 440 Hz.
- Equal temperament uses f = 440 x 2^(n/12), where n is semitone distance from A4.
- Pitch correction estimates the fundamental frequency f0 of the vocal sound over time.
- Correction amount can be described as pitch error = detected pitch - target pitch.
- Fast correction time creates the robotic Auto-Tune effect, while slower correction sounds more natural.
Vocabulary
- Pitch
- Pitch is how high or low a sound seems to a listener, usually related to the sound’s frequency.
- Frequency
- Frequency is the number of wave cycles per second, measured in hertz.
- Fundamental frequency
- Fundamental frequency is the lowest main frequency of a sound and is often heard as the note being sung or played.
- Semitone
- A semitone is the smallest step between two neighboring notes in the standard Western chromatic scale.
- Pitch correction
- Pitch correction is a digital process that detects a note’s pitch and shifts it closer to a chosen target note.
Common Mistakes to Avoid
- Thinking Auto-Tune changes only volume is wrong because pitch correction changes frequency relationships, not loudness.
- Assuming every sound has one clear pitch is wrong because noisy, breathy, or percussive sounds may not have a stable fundamental frequency.
- Snapping every note to the nearest key note can sound wrong because expressive singing uses slides, vibrato, and intentional pitch bends.
- Using an extremely fast correction time for natural vocals is usually wrong because it removes smooth pitch motion and can create a robotic effect.
Practice Questions
- 1 A singer sings A4 at 440 Hz, then sings the same note one octave higher. What is the new frequency?
- 2 Using f = 440 x 2^(n/12), estimate the frequency of the note 12 semitones above A4. Show why this matches the octave rule.
- 3 A vocal note starts slightly flat, slides up into tune, and then uses vibrato. Explain why a slow pitch correction setting may sound more natural than an instant snap to the target note.