Going Retro: Astonishingly Simple Yet Effective Rule-based Prosody Modelling for Speech Synthesis Simulating Emotion Dimensions

TL;DR

This paper presents two simple rule-based prosody models for speech synthesis that effectively modulate emotion dimensions like arousal and valence, compatible with any commercial synthesizer and validated against human annotations.

Contribution

The paper introduces novel, straightforward rule-based models for emotion modulation in speech synthesis using SSML, demonstrating effective simulation of arousal and valence.

Findings

Arousal classification accuracy of 0.76 UAR

Valence classification accuracy of 0.43 UAR

Simple rule-based approach effectively modulates emotion dimensions

Abstract

We introduce two rule-based models to modify the prosody of speech synthesis in order to modulate the emotion to be expressed. The prosody modulation is based on speech synthesis markup language (SSML) and can be used with any commercial speech synthesizer. The models as well as the optimization result are evaluated against human emotion annotations. Results indicate that with a very simple method both dimensions arousal (.76 UAR) and valence (.43 UAR) can be simulated.