Harshness (also called raucousness), in
music information retrieval, is a Non-Contextual Low-Level Audio Descriptors (NLDs) that represents one dimension of the multi-dimensional
psychoacoustic feature called as musical
timbre.
Classical timbre’ NLDs are
roughness,
spectral centroid The spectral centroid is a measure used in digital signal processing to characterise a spectrum. It indicates where the center of mass of the spectrum is located. Perceptually, it has a robust connection with the impression of brightness of a soun ...
, and
spectral flux. While harmonicity and inharmonicity can also be considered NLDs, harshness differs from them, as well as from roughness, once it reckons for a distinguished perceptual audio feature expressed by the summary spectral periodicity. This feature is especially clear in single-
pitch, single-
note, musical audio, where the timbre of two different musical instruments can greatly differ in levels of harshness (e.g., the difference in harshness between a flute and a saxophone is evident). As it is supposed to be, harshness is independent of all others NLDs.
Musicology
Music technology
Music information retrieval