SPAN: Speech Production and Articulation kNowledge Group
An understanding of how language is produced in space and time directly informs, and is informed by, our knowledge of the phonological representation of speech -- a representation that we take to be intrinsically articulatory and dynamic. Using innovative sensing and signal processing technologies, alongside advanced computational techniques, we investigate language production -- from its cognitive conception, to its biomechanical execution, to its signal properties.
The SPAN Group is an interdisciplinary research team, bringing together faculty and students from the Viterbi School of Engineering, and the Departments of Linguistics, and Computer Science. Our work involves collaborations with researchers in the Departments of Psychology and The Keck School of Medicine, and with other research institutions including Haskins Laboratories.
Our research makes use of real-time imaging of the vocal tract, and simultaneous synchronized acquisition of audio and video speech data. Our first movies illustrating these sensing techniques, described in our original publication can be viewed here. We have developed a method of automatic articulator movement tracking, and are making use of state-of-the-art signal analysis techniques, including those found in automatic speech recognition paradigms to interpret the phonetic data acquired using these technologies. Much of our work involves the development and application of novel machine learning approaches to speech analysis.