Meta Audiobox Aesthetics: Unified Automatic Quality Assessment for Speech, Music, and Sound

TL;DR

This paper introduces a unified, automated approach for assessing audio aesthetics across speech, music, and sound, using new annotation guidelines and no-reference models that outperform existing methods and are openly available.

Contribution

It presents a novel annotation framework and no-reference prediction models for audio aesthetics, enabling consistent, automated quality assessment across diverse audio types.

Findings

Models achieve performance comparable or superior to human MOS scores.

The approach is applicable to speech, music, and sound, demonstrating versatility.

Open-source code and datasets support future research and benchmarking.

Abstract

The quantification of audio aesthetics remains a complex challenge in audio processing, primarily due to its subjective nature, which is influenced by human perception and cultural context. Traditional methods often depend on human listeners for evaluation, leading to inconsistencies and high resource demands. This paper addresses the growing need for automated systems capable of predicting audio aesthetics without human intervention. Such systems are crucial for applications like data filtering, pseudo-labeling large datasets, and evaluating generative audio models, especially as these models become more sophisticated. In this work, we introduce a novel approach to audio aesthetic evaluation by proposing new annotation guidelines that decompose human listening perspectives into four distinct axes. We develop and train no-reference, per-item prediction models that offer a more nuanced assessment of audio quality. Our models are evaluated against human mean opinion scores (MOS) and existing methods, demonstrating comparable or superior performance. This research not only advances the field of audio aesthetics but also provides open-source models and datasets to facilitate future work and benchmarking. We release our code and pre-trained model at: https://github.com/facebookresearch/audiobox-aesthetics