Challenge on Sound Scene Synthesis: Evaluating Text-to-Audio Generation

TL;DR

This paper introduces a comprehensive evaluation protocol for text-to-audio generation, combining objective metrics and perceptual assessments, revealing insights into model performance and guiding future research directions.

Contribution

It presents a new evaluation framework for sound scene synthesis that effectively combines objective and perceptual assessments, addressing controllability and evaluation challenges.

Findings

Larger models generally outperform lightweight approaches.

Objective metrics strongly correlate with human perceptual ratings.

Performance varies significantly across sound categories and architectures.

Abstract

Despite significant advancements in neural text-to-audio generation, challenges persist in controllability and evaluation. This paper addresses these issues through the Sound Scene Synthesis challenge held as part of the Detection and Classification of Acoustic Scenes and Events 2024. We present an evaluation protocol combining objective metric, namely Fr\'echet Audio Distance, with perceptual assessments, utilizing a structured prompt format to enable diverse captions and effective evaluation. Our analysis reveals varying performance across sound categories and model architectures, with larger models generally excelling but innovative lightweight approaches also showing promise. The strong correlation between objective metrics and human ratings validates our evaluation approach. We discuss outcomes in terms of audio quality, controllability, and architectural considerations for text-to-audio synthesizers, providing direction for future research.