Determining the minimum sample length for the estimation of speaking rate

• Grant number: 16/12646-0
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: December 01, 2016
• End date: November 30, 2017
• Field of knowledge: Linguistics, Literature and Arts - Linguistics - Linguistic Theory and Analysis
• Principal Investigator: Pablo Arantes
• Grantee: Verônica Gomes Lima
• Host Institution: Centro de Educação e Ciências Humanas (CECH). Universidade Federal de São Carlos (UFSCAR). São Carlos , SP, Brazil

Abstract

We put forward the use of a methodology to determine the minimum speech sample length necessary to estimate speech rate (SR) and articulation rate (AR) from time series composed of SR and AR values calculated cumulatively along speech samples. Minimum sample length is determined by applying change point analysis, that finds the time point where the time series' underlying variance significantly changes. A preliminary study showed that it takes an average of 9.2 and 8.7 seconds for SR and AR stabilize and that rate level (slow, normal, fast) has a significant effect on stabilization time. In the present project we intend to extend the previous results by investigating the possible effect of speaking style (spontaneous and read speech) and linguistic unit used as reference frame (phone, syllable or word) on SR and AR stabilization time. (AU)