Audio Pitch vs Speed: What Happens When You Change Each?

Pitch and speed answer different questions

Pitch describes perceived highness or lowness; playback speed describes how quickly a recording advances. Playing recorded samples faster naturally raises pitch and shortens duration. A pitch-shifting or time-stretching algorithm can separate these effects approximately, with possible artifacts.

For a stable periodic tone, frequency is a useful numerical model of pitch. Speech and musical sounds contain harmonics, transients and changing spectra, so doubling a measured frequency is not the same as making every voice sound naturally like a different speaker. Pitch shifting is not voice cloning.

• MDN: playback rate and pitch correction

Semitones, cents and frequency ratios

In equal temperament, a shift of s semitones and c cents has ratio r = 2^((s + c/100)/12). One cent is one hundredth of a semitone. An octave is 12 semitones: r = 2. A minus-12 shift gives r = 0.5.

For a 440 Hz reference, +12 produces 880 Hz and -12 produces 220 Hz. +3 gives approximately 523.25 Hz; +100 cents is exactly the same interval as +1 semitone, about 466.16 Hz. These are calculated ideal frequencies, not measured guarantees for a complex recording.

Natural rate versus duration-preserving pitch

With a natural rate change, new duration = original duration / r. A +3-semitone shift makes a 60-second source about 50.45 seconds, unless tempo compensation is also applied. This is useful for a deliberately sped-up effect, but it will not remain aligned to an unchanged video timeline.

Duration-preserving mode raises or lowers pitch, then compensates tempo to keep approximately the same length. Tempo-only editing aims to change duration without intentionally changing pitch. Browser playback controls may apply pitch correction automatically, so do not assume a 2x player preview is the same process as a natural-rate export.

Tubelexity shares the filter construction between its pitch and speed tools. Exported audio is re-encoded; the compensation mode is not lossless and exact sample boundaries can differ slightly.

• Change tempo while choosing pitch behavior
• Change video timestamps and its audio together

Why voices can sound strange

Large shifts can change perceived vocal size and timbre. Standard pitch filters do not promise independent preservation of vocal formants. Sustained vowels may become metallic or phasey; sharp consonants and drum attacks may smear. These problems are different from clipping caused by excessive level.

Start with small changes and compare at similar listening levels. For speech, check consonants and breaths; for music, check cymbals, attacks and sustained notes. Listen across an edit boundary as well as the middle of a clip. A waveform that looks normal does not prove transparent sound.

Equalization changes the balance of frequency bands; it does not move a melody into another key. Use EQ for tonal balance after the pitch decision, and leave headroom if boosting bands.

• Adjust tone without changing musical intervals
• Use deliberately styled voice effects

Choose a workflow for the actual goal

To rehearse music in a different key at the same tempo, use duration-preserving pitch. To slow a lesson for comprehension, use pitch-preserving speed. To create a deliberately fast, high-pitched effect, use natural rate. To keep synchronized video, use the video-aware tool rather than changing only the soundtrack length.

Always retain the original and compare a short representative passage before processing a long file. Browser memory, first-use media-engine loading and codec support still apply; an educational article does not need that engine just to be read.