Choice of optimal measurement conditions for calculating the correlation between EMA sensor and video marker position coordinates in electromagnetic articulography

Abstract

Electromagnetic articulography (EMA) is a novel method of speech articulators assessment. By correlating signals from articulograph and video camera one can determine dependencies between internal and external speech articulators. This in turn is important for visual to articulatory speech inversion — the method allowing to estimate inner articulators movement without using aticulograph. This paper presents correlation analysis between video markers and EMA sensors placed on upper and lower lip. Optimal measurement conditions for calculating the correlation are also determined.