AudioLDM 2: Learning Holistic Audio Generation with Self-supervised Pretraining

TL;DR

AudioLDM 2 introduces a unified framework for diverse audio generation tasks using a shared representation called 'language of audio' (LOA), leveraging self-supervised pretraining and diffusion models to achieve state-of-the-art results.

Contribution

The paper proposes a novel unified framework that uses LOA and self-supervised pretraining for multiple audio generation tasks, enabling in-context learning and improved performance.

Findings

Achieves state-of-the-art results on text-to-audio, text-to-music, and text-to-speech benchmarks.

Introduces a universal audio representation called LOA based on AudioMAE.

Demonstrates effective cross-modal translation and generation capabilities.

Abstract

Although audio generation shares commonalities across different types of audio, such as speech, music, and sound effects, designing models for each type requires careful consideration of specific objectives and biases that can significantly differ from those of other types. To bring us closer to a unified perspective of audio generation, this paper proposes a framework that utilizes the same learning method for speech, music, and sound effect generation. Our framework introduces a general representation of audio, called "language of audio" (LOA). Any audio can be translated into LOA based on AudioMAE, a self-supervised pre-trained representation learning model. In the generation process, we translate any modalities into LOA by using a GPT-2 model, and we perform self-supervised audio generation learning with a latent diffusion model conditioned on LOA. The proposed framework naturally brings advantages such as in-context learning abilities and reusable self-supervised pretrained AudioMAE and latent diffusion models. Experiments on the major benchmarks of text-to-audio, text-to-music, and text-to-speech demonstrate state-of-the-art or competitive performance against previous approaches. Our code, pretrained model, and demo are available at https://audioldm.github.io/audioldm2.