You want to learn a song's guitar solo by ear, but it's too fast to follow note by note. If you just play the audio slower with no further processing, the vocals and instruments sound low and warped — the classic "record spinning too slowly" effect. Treating speed and pitch as independent parameters is what lets you slow things down without that side effect, or change the pitch without speeding up or slowing anything down.
Why speed and pitch are linked by default
When you play audio faster or slower with no additional processing, what you're really doing is changing the effective sample rate — and that moves the pitch in the same direction as the speed. Faster means higher-pitched (the "chipmunk" effect); slower means lower-pitched (the "warped record" effect). It's the natural physical behavior of a tape or vinyl record played at the wrong speed.
How speed and pitch get separated in digital processing
To change only the speed without affecting pitch, the algorithm (atempo in FFmpeg) rearranges audio samples in the time domain, stretching or compressing the signal while preserving the original frequency structure. It has a technical limit: each application of the filter only accepts factors between 0.5x and 2x — for more extreme changes (4x, say), the filter gets chained internally several times to reach the desired factor.
To change only the pitch without affecting speed, the reverse trick is used: the sample rate gets reinterpreted (which changes pitch and speed together, like a sped-up record), and then the speed gets compensated back to the original with atempo. The net result is only the pitch shifted, in exact semitones.
The three available modes
- Speed with pitch preserved (default): changes only playback speed, from 0.1x to 10x, without the voice sounding higher or lower.
- Pitch in semitones (independent of speed): raises or lowers pitch between -24 and +24 semitones (two octaves in each direction) while keeping the chosen speed.
- Record effect (pitch not preserved): turns off pitch preservation for the classic effect where speed and pitch rise or fall together, like a record sped up or slowed down by hand.
Real use cases
- Learning a difficult musical passage by ear: slow down to 50-75% with pitch preserved, to hear every note without it sounding like a different note.
- Listening to a lecture or podcast faster: speed up to 1.25x-1.5x with pitch preserved, to save time without the voice sounding high and artificial.
- Creative or joke effect: turn off pitch preservation for the classic "chipmunk voice" when speeding up, or "demon voice" when slowing down.
Frequently asked questions
Why does my voice sound like a chipmunk when I speed up audio? Because you're changing speed without preserving pitch — pitch rises along with speed. Enable the preserve-pitch option to avoid it.
What's the available speed range? From 0.1x to 10x.
How much can I raise or lower the pitch? Up to 24 semitones in each direction (two octaves), while keeping the chosen speed.
Is it processed in my browser? Yes, with FFmpeg compiled to WebAssembly, with nothing uploaded to any server.
Change the speed or pitch of any audio with audio speed and pitch, from 0.1x to 10x and ±24 semitones, 100% in your browser.