Sample-Efficient Diffusion for Text-To-Speech Synthesis
This addresses the challenge of high data requirements for speech synthesis systems, making it more accessible in resource-limited settings, though it is incremental as it builds on existing diffusion and autoencoder methods.
This work tackles the problem of text-to-speech synthesis in low-data regimes by introducing Sample-Efficient Speech Diffusion (SESD), which uses a novel U-Audio Transformer architecture and latent diffusion to achieve more intelligible speech than the state-of-the-art VALL-E model while training on less than 1k hours of speech, specifically using less than 2% of the training data.
This work introduces Sample-Efficient Speech Diffusion (SESD), an algorithm for effective speech synthesis in modest data regimes through latent diffusion. It is based on a novel diffusion architecture, that we call U-Audio Transformer (U-AT), that efficiently scales to long sequences and operates in the latent space of a pre-trained audio autoencoder. Conditioned on character-aware language model representations, SESD achieves impressive results despite training on less than 1k hours of speech - far less than current state-of-the-art systems. In fact, it synthesizes more intelligible speech than the state-of-the-art auto-regressive model, VALL-E, while using less than 2% the training data.