Breathy phonation has a high open quotient compared to modal phonation which results in greater influence of the subglottal cavity on the estimated short-time spectrum. This is reflected as an increase in spectral density at frequencies below the first resonance of the vocal tract around the glottal formant. On the contrary, pressed phonation has lesser influence of the subglottal cavity, and hence has a relatively lesser spectral density at low frequencies. In this paper, the use of low frequency spectral density (LFSD) as a feature for analysis and classification of breathy, modal and pressed phonation types is investigated.
Bibliographic reference. Gowda, Dhananjaya / Kurimo, Mikko (2013): "Analysis of breathy, modal and pressed phonation based on low frequency spectral density", In INTERSPEECH-2013, 3206-3210.