TaDiCodec: Text-aware Diffusion Speech Tokenizer for Speech Language Modeling

TL;DR

TaDiCodec introduces a novel, end-to-end diffusion-based speech tokenizer that integrates text guidance, achieving high compression efficiency and superior speech reconstruction quality without auxiliary models.

Contribution

It presents the first single-stage, end-to-end diffusion autoencoder for speech tokenization with text guidance, eliminating the need for complex multi-stage training or pre-trained models.

Findings

Achieves 6.25 Hz frame rate and 0.0875 kbps bitrate for 24 kHz speech.

Maintains superior speech quality and accuracy metrics like WER, SIM, and UTMOS.

Demonstrates effectiveness in zero-shot text-to-speech with autoregressive and masked models.

Abstract

Speech tokenizers serve as foundational components for speech language models, yet current designs exhibit several limitations, including: 1) dependence on multi-layer residual vector quantization structures or high frame rates, 2) reliance on auxiliary pre-trained models for semantic distillation, and 3) requirements for complex two-stage training processes. In this work, we introduce the Text-aware Diffusion Transformer Speech Codec (TaDiCodec), a novel approach designed to overcome these challenges. TaDiCodec employs end-to-end optimization for quantization and reconstruction through a diffusion autoencoder, while integrating text guidance into the diffusion decoder to enhance reconstruction quality and achieve optimal compression. TaDiCodec achieves an extremely low frame rate of 6.25 Hz and a corresponding bitrate of 0.0875 kbps with a single-layer codebook for 24 kHz speech, while maintaining superior performance on critical speech generation evaluation metrics such as Word Error Rate (WER), speaker similarity (SIM), and speech quality (UTMOS). Notably, TaDiCodec employs a single-stage, end-to-end training paradigm, and obviating the need for auxiliary pre-trained models. We also validate the compatibility of TaDiCodec in language model based zero-shot text-to-speech with both autoregressive modeling and masked generative modeling, demonstrating its effectiveness and efficiency for speech language modeling, as well as a significantly small reconstruction-generation gap. We will open source our code and model checkpoints. Audio samples are are available at https:/tadicodec.github.io/. We release code and model checkpoints at https:/github.com/HeCheng0625/Diffusion-Speech-Tokenizer.