pitchdomain

Pitchdomain is a conceptual term used in audio signal processing and music information retrieval to denote the set of representations and operations that pertain to pitch. In practice, pitchdomain encompasses multiple coordinate systems, including linear frequency in Hz, MIDI note numbers, and chroma or pitch-class profiles that ignore octave. It also includes perceptual scales such as musical pitch and pitch intervals, often represented in log-frequency space (semitones).

The term is used when describing algorithms that detect, track, or modify pitch, such as pitch detection (YIN, autocorrelation, cepstrum), pitch estimation, pitch shifting, and melody transcription. Applications include automatic transcription, tuning, instrument analysis, and music synthesis.

Challenges include octave ambiguity, ambiguity in non-standard tunings, microtonal systems, and rapid pitch movement. Effective pitchdomain modeling requires choosing an appropriate representation for the task and robust mapping between representations. There is no universal standard for pitchdomain; researchers adapt representations to their domain, trading off resolution, robustness, and computational cost.

The term is used variably; in some contexts it emphasizes chroma-based representations, in others, precise frequency tracking. Overall, pitchdomain provides a framework for selecting and comparing the various ways pitch can be represented and manipulated within audio technologies.