Objective and Subjective Evaluation of an Expressive Speech Corpus

This paper presents the validation of the expressive content of an acted corpus produced for its use in speech synthesis. Firstly, objective techniques have been carried out by means of automatic identification of emotions using different data mining techniques applied to statistical features computed over the prosodic parameters of speech. Secondly, a listening test has been performed with a subset of utterances. The relationship between both objective and subjective evaluations is analysed and the obtained conclusions can be useful for the following steps related to expressive speech synthesis.

• Type: Conference
• Year: 2007
• Where: NOLISP
• Authors: Ignasi Iriondo Sanz, Santiago Planet, Joan Claudi Socoró, Francesc Alías
• Keywords: Data Mining Techniques, Expressive Content, NOLISP 2007, Signal Processing, Speech Synthesis