An analysis of Malay language emotional speech corpus for emotion recognition system

Abstract

Human speech is known as the information carrier because its signals can convey people's emotions, age, gender, and ethnic. Speech and emotions are interrelated where from speech, people can express their feelings. Emotional speech corpus in many languages such as English, German, Japanese, Dutch, and French are readily accessible for researchers. However, emotional speech corpus in the Malay language, specifically natural emotion speeches is not available. Thus, the focus of this paper is to develop a spontaneous (natural) emotional speech corpus and perform a prosody analysis of the speech to gather a preliminary understanding of speech-based emotions by age variance. The speech samples consist of natural emotion speeches with three types of emotions (i.e. “angry,” “happy” and “sad”). This paper compares the mean and standard deviations of pitch and intensity of 30 speakers comprising children of age 6–12 years, young adults aged 18–35 years and middle-aged adults of 36–55 years old. Prosody analysis of natural or spontaneous speech in Malay language showed that children can express more emotions followed by middle-aged adults and young adults.