The Speech Production Research and Education Initiative was funded by the Graduate College of the University of Illinois, as part of its Focal Point program, for the purpose of creating a multi-college conversation on the subject of speech production science, technology, clinical practice, and education. This Focal Point Project is a forum that facilitates the study of speech production by connecting many labs and providing students with an integrated view of this topic.
Speech production is a complicated process that starts with cognition as a speaker makes a plan for speech, and follows with muscle commands in which a speaker controls breathing and muscle movements for articulation. In the process, the speaker continuously monitors her speech output and make adjustments. We speak easily and tend to take the system for granted not realizing its importance to health, science, society and the economy. The application of speech knowledge encompasses speech disorders, language preservation, effective language learning, and speech technologies for economic and security applications. Students who are currently being trained in these important research areas need to navigate cross-disciplinary topics to become a new generation of educators in the university, clinic and market-place.
SPREI is the initiator of two key activities in support of multi-college conversation:
- The course Speech Production Research Education Initiative gave its participating students hands-on experience with EMA, EPG, EGG, pressure and airflow transduction, and a variety of labial and tongue force transducers.
- The Speech Production Workshop brought researchers from many colleges and universities together in order to address problems of statistical and physiological data analysis arising from speech production research.
SPREI Graduate Course
As part of SPREI, five graduate students and five faculty from five colleges at the University of Illinois met biweekly through the 2010-11 school year to experiment with speech production measurement technologies. Readings, discussion, and hands-on experimental sessions investigated the use of:
- Electromagnetic articulography (EMA), using the Articulograph AG500
- Surface electromyography (sEMG)
- Labial and lingual contact transducers for maximal and sustained force capacity experiments
- Nasal pressure and airflow transduction
- Electropalatography (EPG)
- Electroglottograph (EGG)
- Microphone and video transducer arrays
- Torrey Loucks — Department of Speech and Hearing Science, University of Illinois at Urbana-Champaign
- Chilin Shih — Department of Linguistics, Department of East Asian Languages and Cultures, University of Illinois at Urbana-Champaign
- Ryan Shosted — Department of Linguistics, University of Illinois at Urbana-Champaign
- Mark Hasegawa-Johnson — Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign
SPREI Publicly Available Speech Production Data
One of the objectives of this program is to provide a common articulatory dataset that may be used for the study of many different types of speech production phenomena.
The Arctic Articulatory Database (Chilin Shih, Ryan Shosted, Torrey Loucks, Chris Carignan, and Panying Rong) contains Articulograph, audio, and manually-checked time-aligned phone transcriptions for a selection of TIMIT and Arctic sentences. Sentences are selected to guarantee uniphone coverage, and to maximize diphone coverage of English. In order to encourage the speaker to produce interesting types of pronunciation variability, each sentence is recorded three times, with a listener at close, normal, and far from the speaker. Uniphone sentences are recorded nine times each = (3 distances) x (3 speaking styles), where the three speaking styles are accented, clear, and fast.