However, state-of-the-art speech assistants fail to comprehend the requests made by young users if they deviate from the predictable set of requests.
In order to build efficient speech systems for kids, the first step would be to understand the variation in acoustics and articulation of kids' speech as a function of age.
Since most of the studies have been cross-sectional, this project primarily focuses on the study of child acoustics using longitudinal data.
The major tasks involved are, study of:
(1) Development of vowel acoustics (Formant frequencies and pitch) with age.
(2) Development Subglottal Resonances (SGRs) with age.
(3) Effect of articulator physiology on acoustics.
We hope that a better understanding of children’s speech acoustics and the science behind speech development and production, would help build better speech systems for children.
Children's Speech Acoustics
Prof. Abeer Alwan, SPAPL, UCLA
Vishwas Shetty, SPAPL, UCLA
Back to SPAPL Home Page.
Abeer Alwan (alwan@ee.ucla.edu)