MeanAudio: Fast and Faithful Text-to-Audio Generation with Mean Flows

TL;DR

MeanAudio introduces a novel, fast text-to-audio generation method that produces realistic sounds with only one function evaluation, significantly outperforming existing diffusion-based systems in speed and quality.

Contribution

The paper proposes MeanAudio, a new TTA system using MeanFlow and an enhanced transformer, enabling rapid, high-quality audio synthesis with minimal inference steps.

Findings

Achieves real-time factor of 0.013 on RTX 3090

100x speedup over state-of-the-art diffusion systems

Effective in both single-step and multi-step audio generation

Abstract

Recent years have witnessed remarkable progress in Text-to-Audio Generation (TTA), providing sound creators with powerful tools to transform inspirations into vivid audio. Yet despite these advances, current TTA systems often suffer from slow inference speed, which greatly hinders the efficiency and smoothness of audio creation. In this paper, we present MeanAudio, a fast and faithful text-to-audio generator capable of rendering realistic sound with only one function evaluation (1-NFE). MeanAudio leverages: (i) the MeanFlow objective with guided velocity target that significantly accelerates inference speed, (ii) an enhanced Flux-style transformer with dual text encoders for better semantic alignment and synthesis quality, and (iii) an efficient instantaneous-to-mean curriculum that speeds up convergence and enables training on consumer-grade GPUs. Through a comprehensive evaluation study, we demonstrate that MeanAudio achieves state-of-the-art performance in single-step audio generation. Specifically, it achieves a real-time factor (RTF) of 0.013 on a single NVIDIA RTX 3090, yielding a 100x speedup over SOTA diffusion-based TTA systems. Moreover, MeanAudio also shows strong performance in multi-step generation, enabling smooth transitions across successive synthesis steps.