无需反演的文本引导音频编辑，速度提升64.5%且失真更小
DirectAudioEdit: Inversion-Free Text-Guided Audio Editing via Diffusion Prediction Contrast
- 直接利用扩散模型去噪路径构建音效修改路径
- 相比DDPM反演方法，FAD降低15.9%，KL降低15.8%
- 适合追求高效、无训练音频编辑的研究与应用
文本引导音频编辑旨在修改指定语言内容的同时保留无关源成分。现有免训练方法通常依赖反演编辑，而免反演编辑虽计算开销更低、重建误差更小，但在音频领域仍几乎未被探索。核心挑战在于通过扩散去噪动态构建从源到目标的编辑路径。本文提出DirectAudioEdit，首个免训练、免反演的音频编辑方法。在音乐与事件级基准上，基于两种骨干网络的实验表明，该方法相较DDPM反演，宏平均FAD与KL分别降低15.9%和15.8%，同时实现最高达64.5%的编辑加速。
原文摘要 · Abstract (English)
Text-guided audio editing aims to modify the language-specified acoustic content while preserving edit-irrelevant source components. Existing training-free methods typically rely on inversion-based editing. While inversion-free editing is appealing as it decreases computational overhead and reconstruction errors, it remains largely unexplored for audio editing. The key challenge is to construct a source-to-target editing path through diffusion denoising dynamics. In this paper, we introduce DirectAudioEdit, the first attempt to develop a training-free and inversion-free method for audio editing. Experiments on music and event-level benchmarks across two backbones show that DirectAudioEdit reduces macro-averaged FAD and KL by 15.9% and 15.8% compared with DDPM inversion, while achieving up to 64.5% editing speedup.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。