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 …

Deep generative models are a class of techniques that train deep neural networks to model
the distribution of training samples. Research has fragmented into various interconnected …

Foundation models, now powering most of the exciting applications in deep learning, are
almost universally based on the Transformer architecture and its core attention module …

Abstract Language models have been successfully used to model natural signals, such as
images, speech, and music. A key component of these models is a high quality neural …

We present SoundStream, a novel neural audio codec that can efficiently compress speech,
music and general audio at bitrates normally targeted by speech-tailored codecs …

Text to speech (TTS), or speech synthesis, which aims to synthesize intelligible and natural
speech given text, is a hot research topic in speech, language, and machine learning …

Developing architectures suitable for modeling raw audio is a challenging problem due to
the high sampling rates of audio waveforms. Standard sequence modeling approaches like …

In this work, we propose DiffWave, a versatile diffusion probabilistic model for conditional
and unconditional waveform generation. The model is non-autoregressive, and converts the …

This paper introduces WaveGrad, a conditional model for waveform generation which
estimates gradients of the data density. The model is built on prior work on score matching …

Despite recent progress in generative adversarial network (GAN)-based vocoders, where
the model generates raw waveform conditioned on acoustic features, it is challenging to …