MoE-TTS
Enhancing Out-of-Domain Text Understanding for Description-based TTS via Mixture-of-Experts
Paper Information
Abstract
Description-based text-to-speech (TTS) models exhibit strong performance on in-domain text descriptions, i.e., those encountered during training. However, in real-world applications, the diverse range of user-generated descriptions inevitably introduces numerous out-of-domain inputs that challenge the text understanding capabilities of these systems. To address this issue, we propose MoE-TTS, a description-based TTS model designed to enhance the understanding of out-of-domain text descriptions. MoE-TTS employs a modality-based mixture-of-experts (MoE) approach to augment a pre-trained textual large language model (LLM) with a set of specialized weights adapted to the speech modality while maintaining the original LLM forzen during training. This approach allows MoE-TTS to effectively leverage the pre-trained knowledge and text understanding abilities of textual LLMs. Our experimental results indicate that: first, even the most advanced closed-source commercial products can be challenged by carefully designed out-of-domain description test sets; second, MoE-TTS achieves superior performance in generating speech that more accurately reflects the descriptions.
Framework Architecture
Building on the core idea of enhancing description-based TTS through the pre-trained knowledge and text understanding capabilities of large language models, we propose MoE-TTS. Our approach employs a mixture-of-experts framework that utilizes a modality-based routing strategy and modality-aware Transformer components to bridge large language models and TTS models.
Out-of-domain Descriptions Showcases
Note: As of early August 2025, the ElevenLabs multilingual_v2 API was used, as the alpha v3 remained inaccessible.