What it does
The pitch shifter raises or lowers every note in a sound by the same amount, without making it faster or slower. Singers use it to move a backing track into a key that suits their voice; players use it to match a recording that was tuned a little sharp or flat; teachers use it to make a piece easier to sing along with. It runs in your browser, so your file stays on your device.
Each semitone is one step on a piano keyboard; 12 semitones is an octave. Fine tune works in cents (hundredths of a semitone) for recordings that are slightly out from standard tuning. Check the tuning first with the instrument tuner, and use audio speed if you want to change the tempo instead.
How to use it
- Press Open a sound file.
- Set Semitones with the box or the slider: +2 moves a song in C up to D; −3 moves it from C down to A.
- Add Fine tune in cents if needed, for example −20 for a record that plays sharp.
- Press Play the result, then choose Save as and Download.
What each setting does
- Semitones: whole steps from −12 (one octave down) to +12 (one octave up). The frequency of every note is multiplied by 2^(n/12): +1 is × 1.0595, +7 is × 1.4983, +12 is × 2.
- Fine tune, cents: from −50 to +50; 100 cents is a semitone.
- Save as: MP3 or WAV, as in the audio converter.
How it works: the sound is first stretched in time without changing its pitch (WSOLA), then resampled back to its original length, which moves the pitch by the same ratio.
Worked example
A 440 Hz A shifted +4 semitones becomes 440 × 2^(4/12) = 554.4 Hz, a C♯, and the file stays exactly the same length. Shifted −12 it becomes 220 Hz. The checks for this page measure the result with the YIN pitch finder and expect it within 0.5 Hz.
FAQs
Why do voices sound odd after a big shift? Moving a voice up a lot also moves the resonances of the singer’s throat, which gives the “chipmunk” sound. Shifts of up to about 4 semitones sound natural.
Does it change the speed? No. The length is kept to the sample.
Is my file uploaded? No. Everything happens on this device.
Where the numbers come from
- Time-stretching by WSOLA (Verhelst and Roelands, 'An overlap-add technique based on waveform similarity', ICASSP 1993); resampling with a windowed-sinc filter. Code is fibub's own working, checked in tests/audio.mjs.
- Equal temperament: frequency ratio 2^(n/12) for n semitones. MP3 saving with LAME via lamejs (LGPL-3.0, credited in NOTICE).
Standard scientific values, such as constants, atomic weights and melting points, are facts found in every textbook; we name where ours were checked. If you are an author or publisher and want a value, a credit or a table changed or removed, tell us and we will act quickly.
Last updated 2026-11-16.