Emotional expressiveness of the singing voice: The role of macrostructural and microstructural modifications of spectra

Abstract

Singing is a conversation with the listener by means of the language of emotions. All timbral and emotional nuances are manifested, to a great extent, by various modifications of spectra. Research carried out previously showed the important role of macrostructures of spectra in expressing emotions, i.e. changes of the integral of amplitude and frequency of formant zones. Thus, for instance, in cases of joy and anger the amplitude of the level of the higher singer's formant increases. Sorrow and especially fear are characterised by a significant drop of the level of the higher singer's formant. Our recent studies carried out together with Yury M. Kouznetsov showed that the emotional expressiveness of the singing voice is also manifested by microstructural modifications of spectra, i.e. there are some significant (and statistically representative) deviations of frequencies of overtones from their ideal harmonic position on the frequency scale of the fundamental pitch. Minimum deviations take place in cases of emotionally neutral singing and in cases of sorrow (+2 – +3%), maximum deviations were registered for fear and anger (+7% and higher). It was noticed that the direction of deviations (rising or falling) may have an accidental nature or may depend on emotions expressed or on whether the singing is in a major or in minor key. The authors named this phenomenon “quasiharmony of overtones of the singing voice”. The paper contains an explanation of the phenomenon.