2016 INTERSPEECH INTERSPEECH 2016

Effects of Stress on Fricatives: Evidence from Standard Modern Greek

Abstract

This study investigates the effects of stress on the spectral properties of fricative noise in Standard Modern Greek (SMG). Twenty female speakers of SMG participated in the study. Fricatives were produced in stressed and unstressed positions in two vowel place positions: back and front vowels. Acoustic measurements were taken and the temporal and spectral properties of fricatives β€” using spectral moments β€” were calculated. Stressed fricatives are produced with increased duration, center of gravity, standard deviation, and normalized intensity. The machine learning and classification algorithm C5.0 has been employed to estimate the contribution of the temporal and spectral parameters for the classification of fricatives. Overall, duration and center of gravity contribute the most to the classification of stressed vs. unstressed fricatives.

πŸš€ Conference Pioneer β€” INTERSPEECH 2016
πŸŒ‰ Interdisciplinary Bridge β€” Interdisciplinary and Machine Learning and Speech & Audio
🧭 Keyword Pioneer β€” stress detection
🐣 Hot Topic Early Bird β€” machine learning
🐝 Cross-Pollinator β€” Artificial Intelligence, Computer Science, Computer Vision, Data Science & Analytics, Deep Learning, Healthcare & Medicine, Interdisciplinary, Knowledge & Reasoning, Machine Learning, Mathematics & Optimization, Natural Language Processing, Reinforcement Learning, Robotics, Security & Privacy, Speech & Audio