Neutral to happy emotion conversion by blending prosody and laughter

Abstract

In this paper, we propose a method to convert synthesised neutral speech to happy emotion speech by prosody modification and synthesizing laughter sequence. Excitation source features of verbal (spoken) and non-verbal (laughter) segments of elicited happy emotion are separately analysed and incorporated in the neutral speech. Pitch contours of verbal segments of the happy emotion are divided into three components and analysed separately for incorporation. Laughter sequence is synthesised by incorporating the appropriate pitch period and strength of excitation contours on a synthesised vowel utterance. Modulated frication noise is added to the laughter segments to enhance the quality further. The effectiveness of the synthesized happy speech is confirmed by the waveforms, spectrograms and subjective evaluations.