Why Slowing Audio Doesn't Have to Sound Like a Chipmunk

Slow a record down on a turntable and everything drops into a low groan. Speed a tape up and voices turn into chipmunks. For most of audio history, speed and pitch were two ends of the same lever — you couldn't move one without the other. Modern audio software broke that link, and understanding how explains why a tool like Capo can slow a song to a crawl while it still sounds completely natural.

Why speed and pitch were always linked

Sound is a wave. Its pitch is how often that wave repeats per second — its frequency. The naïve way to slow audio is to play its samples back more slowly, which stretches every wave out in time. A stretched wave repeats less often, so its frequency drops, and with it the pitch. Play it faster and the opposite happens. That's not a bug; it's just geometry. Slowing the playback is lowering the pitch, when you do it the simple way.

This linked behaviour has a name: varispeed — the classic tape-machine effect.

How time-stretching breaks the link

To change speed without changing pitch, software uses time-stretching. Instead of stretching the waves themselves, it chops the audio into very short grains and intelligently repeats or removes them — making the sound last longer or shorter while each individual wave keeps its original frequency. The pitch stays exactly where it was; only the timeline changes.

The reverse — pitch-shifting — changes the frequency while keeping the duration the same, which is what lets Capo move a song up or down in pitch and key without it speeding up.

Formants: why a shifted voice still sounds human

There's a subtler detail. A voice isn't a single frequency; it has formants — fixed resonances, set by the size and shape of someone's throat and mouth, that give it its character. Shift pitch carelessly and you drag the formants along too, which is what makes a pitched-up voice sound cartoonish. Quality pitch-shifting treats pitch and formants separately, so a transposed singer still sounds like a person, just higher or lower.

What this means in practice

Because Capo keeps speed and pitch as independent controls, you get the useful half of varispeed with none of the side effects:

• Slow a solo to 0.25× to learn it, and it stays in tune.
• Speed a lecture to 2.5×, and the voice stays natural.
• Transpose a song into your key, and the tempo doesn't budge.

The old lever is now two separate dials — which is exactly what you want when you're trying to understand a recording rather than warp it.