Sketching Audio Technologies using Vocalizations and Gestures

WP5 – Automatic imitation recognition

To technically integrate the developments of WP3 and the results of WP4 into a system that will feed WP6. Transforming vocal and gestural imitations into synthetic sounds requires a system to automatically identify the category of imitated sound source. Hence such a system should fulfill three functions: estimating the acoustical and gestural features of the imitation and type of articulatory mechanisms, segmenting the signal into sequences of meaningful elements, and predicting the category of the imitated sound source.