time-stretching

Time-stretching IS a specific kind of signal processing technique — it manipulates the time-domain representation of a signal while preserving certain properties. A competent speaker would call time-stretching 'a signal processing method.' Files against the nearest established kind (signal-processing, since phase-vocoder INSTANCE_OF signal-processing exists but time-stretching → phase-vocoder is DEPENDS_ON which is not transitive for INSTANCE_OF).

Time-stretching (phase-vocoder method) operates on the STFT: modify the phase progression to change duration while preserving pitch. Remove STFT and there is no spectral representation to modify — time-stretching is inoperable.

Phase vocoder is designed and maintained for the sake of time-stretching: it manipulates spectral phase to change duration without altering pitch. The servant (phase-vocoder) points at the master (time-stretching). Law 8d: for-whose-sake relationship.

Phase vocoder is a specific technique for time-stretching audio: it changes duration without altering pitch by manipulating spectral phase. A competent speaker would call it a kind of time-stretching method.

Granular synthesis is designed and used for time-stretching: by independently controlling grain onset times, one can stretch or compress audio duration without affecting perceived pitch. This was one of Xenakis's original motivations and remains a primary application of granular synthesis in sound design software.

Pitch-synchronous overlap-add is built and maintained for the sake of time-stretching audio while preserving pitch; its sustained practice is for time-scale modification.