EmoMix: Emotion Mixing via Diffusion Models for Emotional Speech Synthesis

TL;DR

EmoMix is a diffusion-based emotional TTS system that enables the synthesis of speech with mixed emotions and controllable intensity, overcoming limitations of previous methods in emotion diversity and intensity regulation.

Contribution

Introduces EmoMix, a novel diffusion model-based TTS framework capable of generating mixed emotions and controlling emotional intensity in speech synthesis.

Findings

Effective in synthesizing speech with mixed emotions.

Demonstrates precise control over emotional intensity.

Outperforms existing methods in emotion diversity and intensity accuracy.

Abstract

There has been significant progress in emotional Text-To-Speech (TTS) synthesis technology in recent years. However, existing methods primarily focus on the synthesis of a limited number of emotion types and have achieved unsatisfactory performance in intensity control. To address these limitations, we propose EmoMix, which can generate emotional speech with specified intensity or a mixture of emotions. Specifically, EmoMix is a controllable emotional TTS model based on a diffusion probabilistic model and a pre-trained speech emotion recognition (SER) model used to extract emotion embedding. Mixed emotion synthesis is achieved by combining the noises predicted by diffusion model conditioned on different emotions during only one sampling process at the run-time. We further apply the Neutral and specific primary emotion mixed in varying degrees to control intensity. Experimental results validate the effectiveness of EmoMix for synthesizing mixed emotion and intensity control.