Effect of negative and positive emotions on the fundamental frequency and formants
Abdellah Agrima, Laila Elmazouzi, Ilham Mounir, Abdelmajid Farchi · 2017
The speech signal provides rich information about the speaker's emotional state. Therefore, this article provides an experimental study and examines the detection of negative emotion (Sadness) and positive emotion (Joy) with regard to the neutral emotional state. The data set is collected from speeches recorded in the Moroccan Arabic dialect. Our aim is first to study the effects of emotion on the selected acoustic characteristics, namely the first four formants F1, F2, F3, F4, the fundamental frequency F0, and then compare our results to previous works. We also study the influence of speaker gender on the relevance of these characteristics in the detection of emotion. The main tool is classification algorithms using the WEKA software. We found that F0 presents the best rates of recognition regardless speaker gender.