The field of speech processing has undergone a transformative shift with the advent of deep
learning. The use of multiple processing layers has enabled the creation of models capable …

We introduce a language modeling approach for text to speech synthesis (TTS). Specifically,
we train a neural codec language model (called Vall-E) using discrete codes derived from …

Large-scale generative models such as GPT and DALL-E have revolutionized the research
community. These models not only generate high fidelity outputs, but are also generalists …

Recent advancements in generative speech models based on audio-text prompts have
enabled remarkable innovations like high-quality zero-shot text-to-speech. However …

Scaling text-to-speech to a large and wild dataset has been proven to be highly effective in
achieving timbre and speech style generalization, particularly in zero-shot TTS. However …

While recent large-scale neural codec language models have shown significant
improvement in zero-shot TTS by training on thousands of hours of data, they suffer from …

Generative models have gained more and more attention in recent years for their
remarkable success in tasks that required estimating and sampling data distribution to …

In this work, we present DiffVoice, a novel text-to-speech model based on latent diffusion.
We propose to first encode speech signals into a phoneme-rate latent representation with a …

Improving text representation has attracted much attention to achieve expressive text-to-
speech (TTS). However, existing works only implicitly learn the prosody with masked token …

We introduce VoiceCraft, a token infilling neural codec language model, that achieves state-
of-the-art performance on both speech editing and zero-shot text-to-speech (TTS) on …