Pitch Correction & Autotune: Pitch Changer Guide
A pitch changer like Vocal Cut's Pitch Changer moves an entire recording up or down by one fixed amount, while pitch correction, the thing most people call autotune, moves individual sung notes onto the nearest note of a chosen scale. The first transposes a song and the second tunes a performance. People mix the two up, which is why so many go looking for a way to "autotune a finished song" and come away frustrated.
What pitch correction is
Pitch correction is processing that listens to a single voice or instrument, works out which note is being sung, and nudges it toward where that note should be. If a singer lands slightly flat on a held note, correction pulls just that note up. Notes that were already in tune are left alone, or close to it.
The size of the move is the first big difference from transposing. A semitone, the distance between two neighbouring piano keys, is divided into 100 cents, and corrections are usually a fraction of a semitone. Semitones and Cents: A Pitch Changer Guide covers both units.
How does autotune work?
Automatic pitch correction runs through four steps, over and over, as the vocal plays:
- Detect the pitch. The software measures how fast the waveform repeats, which gives the note currently being sung.
- Know the key and scale. You tell it which notes are allowed, for example the seven notes of A minor. If you don't know the song's key, a Key Finder can estimate it, and What Is Musical Key? A Key Finder Guide explains what the result means.
- Find the nearest allowed note. A pitch sitting a little under C is assumed to be a C that came out flat.
- Shift toward it. The audio is repitched by the gap between what was sung and the target, without changing its timing.
The catch is in step one. Detection expects monophonic input, meaning one voice or one instrument playing one note at a time. A lone vocal has a clear repeating pattern to measure. A chord, or a singer with a band behind them, does not.
Set the wrong key or scale and the tool will snap good notes onto wrong ones.
Transparent tuning vs the autotune effect
The same processor produces both the correction nobody notices and the robotic effect everybody recognises. Most of the difference comes from one control, retune speed, which sets how quickly a note is pulled to its target.
At slow settings the correction eases in and is hard to hear. At the fastest setting, with the right key and scale chosen, every note jumps to pitch immediately. Singers naturally slide between notes and add vibrato, a gentle wobble in pitch on held notes. Correct too fast and those slides and wobbles get flattened into hard steps. That stepped sound is the autotune effect.
Other controls exist to keep tuning natural. Humanize-style controls relax the correction on long sustained notes so they don't sound frozen. Antares' Flex-Tune applies correction only as a note gets close to a scale note, so bends and slides further from the target pass through untouched.
There are also two ways of working:
- Automatic, real-time correction tunes the audio as it passes through. It's fast to set up.
- Graphical or manual editing tracks the pitch first, draws each note on screen, and lets you adjust notes one by one. It takes longer, and you decide what gets fixed and what stays.
A short history
Auto-Tune is made by Antares Audio Technologies and was released in 1997. Its creator, Andy Hildebrand, was an engineer who had previously analysed seismic data for the oil industry using autocorrelation, a way of finding repeating patterns in a signal. The same maths turned out to suit pitch detection.
A year later, Cher's "Believe" became the first hit to use Auto-Tune as an obvious stylistic effect. T-Pain's 2005 album Rappa Ternt Sanga made it his signature sound, still known as the "T-Pain effect."
Celemony's Melodyne took the note-by-note graphical route. It was first shown in 2001 and originally handled a single voice only. Its DNA (Direct Note Access) technology for polyphonic audio, material with several notes sounding at once, was announced in 2008 and released in 2009.
Autotune vs pitch correction vs a pitch changer
"Auto-Tune" is Antares' brand name. In everyday speech, "autotune" has become the general word for pitch correction. So autotune vs pitch correction is mostly a naming question. Auto-Tune is a product and pitch correction is the category it belongs to.
A pitch changer does a different job. It performs pitch shifting, also called transposing, on the whole file.
| Pitch correction (autotune) | Pitch shifting / transposing (pitch changer) | |
|---|---|---|
| What it moves | Individual notes of one voice or instrument | Everything in the file at once |
| How far | Usually a fraction of a semitone, different for each note | The same amount throughout, typically whole semitones |
| Needs to know | The key and scale | Only how far to shift |
| Input it expects | A single isolated voice or instrument | Any audio, including a full mix |
| Fixes an off note? | Yes | No, every note moves together |
| Typical job | Tuning a vocal take | Changing a song's key |
Why you can't autotune a finished song
Automatic pitch correction doesn't work usefully on a finished, mixed song, for two reasons.
First, detection breaks down. With drums, bass, chords, and vocals all sounding together, there is no single "main" pitch to measure reliably, so the corrector's guesses jump around.
Second, even when it guesses right, the shift applies to the whole signal. Pulling a flat vocal note up drags the guitar, bass, and keys up with it for that moment, so instruments that were in tune go out.
The usual workaround is to get the vocal on its own first. If you only have the mixed file, the Vocal Remover can separate the vocal from the instrumental using an AI model that runs on your own device. It downloads the model the first time you use it and takes real processing time, and how clean the result is varies with the mix. You can then tune that vocal with a pitch-correction plugin in a DAW (digital audio workstation, the software used to record and mix music) and lay it back over the instrumental. Leftover traces of other instruments can confuse pitch detection, so expect some manual cleanup.
What an online pitch changer can and can't do
Vocal Cut has no pitch-correction or autotune tool. Its Pitch Changer is an online pitch changer that moves a whole file you already have by one fixed amount:
- A Pitch transpose slider from −12 to +12 semitones in whole-semitone steps, labelled "-1 Octave," "No Shift," and "+1 Octave."
- A separate Playback speed slider from 0.50x to 2.00x. The two are independent; How Pitch, Speed, and Duration Relate explains why that matters.
It cannot fix one flat note, snap notes to a scale, or adjust by cents, and it has no formant control. Formants are the resonances that give a voice its character, covered in What Are Formants? A Pitch Changer Guide.
What it's good at is transposing. It can move a backing track into your vocal range, and notes that sit comfortably are easier to sing in tune. Fit a Song to Your Voice With a Pitch Changer walks through finding the right shift. It can also match a sample's key to your track. Used as a vocal pitch changer on an isolated voice, it works best with small shifts; How to Pitch Shift Without Ruining Vocals covers the limits. Processing runs in your browser, and your audio never leaves your device.
Common myths
"Autotune always sounds robotic." Only at fast retune speeds. Slow, gentle correction is hard to hear, which is why you rarely notice it.
"A pitch changer is autotune." Both alter pitch, but a pitch changer shifts everything equally. A singer who is flat on one note is still flat on it after transposing, just in a different key.
For more definitions like these, see the Audio Explainers hub.
Frequently asked questions
Is autotune cheating?
That depends on what you count as cheating. Pitch correction is a studio tool in the same family as editing between takes, EQ, and compression: it polishes a recorded performance. Timing, tone, and delivery still come from the singer. Used as an obvious effect, as on Cher's "Believe" or T-Pain's records, it's a stylistic choice that listeners can plainly hear.
Can autotune make a bad singer sound good?
It can make a singer sound more in tune, which is not the same thing. Pitch correction moves notes onto the scale, but it does nothing for timing, tone, breath, or emotional delivery. Notes that are a long way off may also get snapped to the wrong target. It helps a decent take more than it rescues a poor one.
What is the difference between Auto-Tune and Melodyne?
Auto-Tune, from Antares, is best known for automatic correction that tunes a vocal in real time as audio passes through, though it also has a graph mode for editing notes by hand. Melodyne, from Celemony, is built around note-by-note graphical editing: it analyses the recording first, then shows each note so you can adjust it individually.
Can I autotune a song that is already mixed?
Not usefully. Automatic pitch correction expects one voice or instrument at a time, and on a full mix any shift it applies moves every instrument along with the singer. The practical route is to isolate the vocal first, tune it on its own in a DAW plugin, and then recombine it with the instrumental.
Does Vocal Cut have autotune?
No. Vocal Cut has no pitch-correction or autotune tool. Its Pitch Changer transposes a whole file by −12 to +12 semitones in whole-semitone steps, with a separate playback speed control from 0.50x to 2.00x. It can't fix individual notes or snap them to a scale, but it's useful for moving a song into your range or matching a sample's key.