EMORL-TTS: Reinforcement Learning for Fine-Grained Emotion Control in LLM-based TTS

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Haoxun, Liu, Yu, Sun, Yuqing, Shi, Hanlei, Qu, Leyuan, Li, Taihao
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917219979493376
author Li, Haoxun
Liu, Yu
Sun, Yuqing
Shi, Hanlei
Qu, Leyuan
Li, Taihao
author_facet Li, Haoxun
Liu, Yu
Sun, Yuqing
Shi, Hanlei
Qu, Leyuan
Li, Taihao
contents Recent LLM-based TTS systems achieve strong quality and zero-shot ability, but lack fine-grained emotional control due to their reliance on discrete speech tokens. Existing approaches either limit emotions to categorical labels or cannot generalize to LLM-based architectures. We propose EMORL-TTS (Fine-grained Emotion-controllable TTS with Reinforcement Learning), a framework that unifies global intensity control in the VAD space with local emphasis regulation. Our method combines supervised fine-tuning with reinforcement learning guided by task-specific rewards for emotion category, intensity, and emphasis. Moreover, we further investigate how emphasis placement modulates fine-grained emotion intensity. Experiments show that EMORL-TTS improves emotion accuracy, intensity differentiation, and emphasis clarity, while preserving synthesis quality comparable to strong LLM-based baselines.
format Preprint
id arxiv_https___arxiv_org_abs_2510_05758
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle EMORL-TTS: Reinforcement Learning for Fine-Grained Emotion Control in LLM-based TTS
Li, Haoxun
Liu, Yu
Sun, Yuqing
Shi, Hanlei
Qu, Leyuan
Li, Taihao
Sound
Recent LLM-based TTS systems achieve strong quality and zero-shot ability, but lack fine-grained emotional control due to their reliance on discrete speech tokens. Existing approaches either limit emotions to categorical labels or cannot generalize to LLM-based architectures. We propose EMORL-TTS (Fine-grained Emotion-controllable TTS with Reinforcement Learning), a framework that unifies global intensity control in the VAD space with local emphasis regulation. Our method combines supervised fine-tuning with reinforcement learning guided by task-specific rewards for emotion category, intensity, and emphasis. Moreover, we further investigate how emphasis placement modulates fine-grained emotion intensity. Experiments show that EMORL-TTS improves emotion accuracy, intensity differentiation, and emphasis clarity, while preserving synthesis quality comparable to strong LLM-based baselines.
title EMORL-TTS: Reinforcement Learning for Fine-Grained Emotion Control in LLM-based TTS
topic Sound
url https://arxiv.org/abs/2510.05758