MARS6: A Small and Robust Hierarchical-Codec Text-to-Speech Model
This addresses the challenge of robust and expressive TTS for applications requiring efficient, high-quality voice synthesis, though it appears incremental by combining existing techniques.
The paper tackles the problem of expressive text-to-speech (TTS) with zero-shot voice cloning, presenting MARS6, a 70M-parameter model that achieves similar performance to larger models by processing speech tokens at 12 Hz and improving output stability.
Codec-based text-to-speech (TTS) models have shown impressive quality with zero-shot voice cloning abilities. However, they often struggle with more expressive references or complex text inputs. We present MARS6, a robust encoder-decoder transformer for rapid, expressive TTS. MARS6 is built on recent improvements in spoken language modelling. Utilizing a hierarchical setup for its decoder, new speech tokens are processed at a rate of only 12 Hz, enabling efficient modelling of long-form text while retaining reconstruction quality. We combine several recent training and inference techniques to reduce repetitive generation and improve output stability and quality. This enables the 70M-parameter MARS6 to achieve similar performance to models many times larger. We show this in objective and subjective evaluations, comparing TTS output quality and reference speaker cloning ability. Project page: https://camb-ai.github.io/mars6-turbo/