Boosting Large Language Model for Speech Synthesis: An Empirical Study

TL;DR

This paper empirically investigates methods to enhance large language models with speech synthesis capabilities by combining them with a text-to-speech model, finding that coupling LLMs as text encoders yields the best results.

Contribution

It systematically compares three integration methods for augmenting LLMs with speech synthesis, highlighting the effectiveness of coupled approaches over direct fine-tuning.

Findings

Superposing LLMs and VALL-E improves speech quality.

Coupled LLMs and VALL-E outperform other methods.

Achieved 10.9% reduction in word error rate.

Abstract

Large language models (LLMs) have made significant advancements in natural language processing and are concurrently extending the language ability to other modalities, such as speech and vision. Nevertheless, most of the previous work focuses on prompting LLMs with perception abilities like auditory comprehension, and the effective approach for augmenting LLMs with speech synthesis capabilities remains ambiguous. In this paper, we conduct a comprehensive empirical exploration of boosting LLMs with the ability to generate speech, by combining pre-trained LLM LLaMA/OPT and text-to-speech synthesis model VALL-E. We compare three integration methods between LLMs and speech synthesis models, including directly fine-tuned LLMs, superposed layers of LLMs and VALL-E, and coupled LLMs and VALL-E using LLMs as a powerful text encoder. Experimental results show that, using LoRA method to fine-tune LLMs directly to boost the speech synthesis capability does not work well, and superposed LLMs and VALL-E can improve the quality of generated speech both in speaker similarity and word error rate (WER). Among these three methods, coupled methods leveraging LLMs as the text encoder can achieve the best performance, making it outperform original speech synthesis models with a consistently better speaker similarity and a significant (10.9%) WER reduction.