tempoas

Tempoas are a theoretical construct in music analysis and audio processing used to denote discrete tempo states or tempo trajectories within a piece. A tempoa encodes tempo values, typically in beats per minute, and their evolution over time, enabling analysis of tempo fluctuations independent of pitch or timbre.

The term is not standardized and appears mainly in niche literature and software documentation. It blends tempo with a pluralizing suffix and is often used interchangeably with "tempo trajectory" or "tempo curve."

Representations vary: tempoas may be a time series of BPM estimates, a sequence of piecewise-constant segments, or a continuous tempo function. They capture accelerations, ritardandos, rubato-like timing, and other tempo changes, with granularity chosen per application.

Applications include beat tracking, alignment of audio to scores, tempo normalization for cross-performance comparison, and tempo-aware synchronization in multimedia and dance analysis.

Inference of tempoas commonly uses statistical or machine learning methods such as hidden Markov models, Kalman filters, Bayesian change-point detection, or neural networks trained on tempo annotations. Evaluation often relies on ground-truth tempo tracks or alignment error metrics.

See also: tempo, beat tracking, tempo curve, tempo mapping, rhythm analysis.