harmonictonoise

Harmonictonoise is a term used in audio analysis to describe the relative balance between harmonic content and noise in a signal. It is not a formal standard metric, but it appears in discussions of timbre and signal quality.

Measurement involves analyzing the spectrum with a known or estimated fundamental frequency, f0. Identify the harmonic series at k·f0 for integers k = 1, 2, 3, and so on. Compute energy in the harmonic components by summing spectral energy around each harmonic bin within a tolerance. Compute energy in the residual components treated as noise. The ratio is defined as HN = E_harm / E_noise, and can be expressed in linear form or converted to decibels for convenience.

Applications and interpretation: A higher harmonictonoise ratio indicates a clearer, more pitched tone with stronger perceptible harmonic structure; a lower ratio signals a noisier or more inharmonic timbre. The metric can be useful in sound synthesis, voice processing, recording quality assessment, and instrument design, where a quantitative measure of timbre purity is desired.

Relation to other metrics: Harmonictonoise is related to concepts such as harmonicity, spectral flatness, and signal-to-noise ratio (SNR). It is not universally standardized; some discussions refer to harmonics-to-noise ratio (HNR) in voice analysis, which may overlap with harmonictonoise but use different definitions or emphasis. Users should be clear about the chosen definition when comparing results.

Limitations: The measure depends on accurate f0 estimation, windowing, and analysis bandwidth. It can be biased for noisy, transient, or highly inharmonic sounds. As with other descriptive features, harmonictonoise is most informative when used alongside complementary metrics.