AMULET: automatic MUltisensor speech labelling and event tracking: study of the spatio-temporal correlations in voiceless plosive production.