PitchShifting

Pitch shifting is a digital signal processing operation that transposes the pitch of an audio signal by a specified factor, typically without substantial change to its duration. It is widely used to transpose melodies, harmonize vocals, and create vocal or instrument effects. A shift factor greater than one raises pitch; less than one lowers it. In many tools, pitch shifting can also alter duration if time-stretching is not applied, whereas many algorithms aim to preserve timing when shifting pitch and preserve pitch when adjusting duration.

Techniques used to achieve pitch shifting include time-domain methods such as PSOLA (Pitch-Synchronous Overlap-Add) and SOLA/WSOLA, which manipulate short segments of the waveform to change pitch while keeping the overall duration nearly constant. Granular synthesis can re-synthesize the signal from grains pitched up or down. Frequency-domain approaches, such as the phase vocoder, analyze and resynthesize spectra to change pitch, which can introduce phase-related artifacts and may require formant correction to preserve natural vocal qualities. Modern implementations often blend approaches to reduce artifacts and preserve transients.

Considerations for effective pitch shifting include potential artifacts like chirp, metallic or watery timbres, and altered transients if the processing is aggressive. Formant preservation helps maintain natural vocal quality but is not universal. Real-time processing adds latency, and higher-quality results are often achieved offline. Monophonic sources generally yield cleaner shifts than complex polyphonic textures.

Applications span music production, vocal processing, karaoke, sound design, and instrument tuning. Pitch shifting is widely available as plugins (VST, AU, AAX), standalone editors, and hardware units.