Abstract
In this paper, we model the amplitude envelope of the broadband speech signal as a kinematic system and calculate its basic parameters, including displacement, velocity and acceleration. Such system captures the smoothed amplitude fluctuation pattern over time, illustrating how energy is distributed across the signal. Although the pulmonic air pressure is the primary energy source of speech, the amplitude modulation pattern is largely determined by articulatory behaviours, especially mandible and lip movements. Therefore, there should be a correspondence between signal envelope kinematics and articulator kinematics. Previous research showed that a tremendous amount of speaker idiosyncrasies in articulation existed. Such idiosyncrasies should therefore be reflected in the envelope kinematics as well. From the signal envelope kinematics, it may be possible to infer individual articulatory behaviours. This is particularly useful for forensic phoneticians who usually have no access to articulatory data, and clinical speech pathologists who usually find it difficult to make articulatory measurement in clinical consultations. Also in this paper, we illustrate a correspondence between the amplitude envelope kinematics and the lower lip kinematics (X-ray pellet history data) of one speaker reading one sentence. For future research, more speakers are needed to record both speech and articulatory signals to build a statistical model between the kinematics data of both domains.
He, Lei