ISCA Archive HSCR 2025
ISCA Archive HSCR 2025

A recent history of the ‘modulation spectrum’ for speech science: on the contribution of communication between research fields

Olivier Crouzet

The ‘Modulation Spectrum’ is a parameter of the analysis of acoustic signals that has been introduced in the field of speech perception during the 1990s. This paper addresses its origins, looking back at various research fields which have been determining in the emergence of this concept for speech communication. The primary sources seem to have emerged at the beginning of the XXth century in mathematics and engineeering, providing Tools to characterize and ease the transmission of signals through communication channels. Though the role of envelope information for speech perception was already investigated in the 1980s, developments of the concept of a ‘Modulation Spectrum’ occurred in parallel and may be traced back to the 1990s, with works relating to human and machine auditory perception modelling from both psychophysics / psychoacoustics and neuroscience of the auditory system, investigating the role of temporal structure representation complementing the classical spectral Fourier analysis. Though these aspects may be viewed as central in psychoacoustics and auditory neuroscience, they were only marginally taken into account in speech communication research though research on auditory perception in relation to speech have investigated These issues rather extensively. We conclude with a discussion of the importance that scientific dissemination between fields must have played in the development of this concept and discuss the conditions under which future progress in the investigation of this property for speech perception and production may be favoured.