用多模态联合训练，让视频生成高质量同步音频
MMAudio: Taming Multimodal Joint Training for High-Quality Video-to-Audio Synthesis
- 联合训练视频与文本-音频数据，提升生成质量
- 8秒音频生成仅需1.23秒，参数量仅157M
- 兼顾视频/文本到音频生成，适合多场景应用
我们提出MMAudio框架，通过新型多模态联合训练，实现给定视频和可选文本条件下的高质量、高同步性音频合成。不同于仅依赖有限视频数据的单模态训练，MMAudio联合使用大规模易获取的文本-音频数据，学习生成语义对齐的高质量音频样本。此外，引入条件同步模块，在帧级对齐视频条件与音频潜在表示，增强音画同步。基于流匹配目标训练，MMAudio在公开模型中达到视频到音频生成的新基准，涵盖音频质量、语义对齐与音画同步三项指标，推理时间仅1.23秒生成8秒音频，参数量为157M。同时在文本到音频生成任务上表现优异，表明联合训练未损害单模态性能。代码与演示见：https://hkchengrex.github.io/MMAudio
原文摘要 · Abstract (English)
We propose to synthesize high-quality and synchronized audio, given video and optional text conditions, using a novel multimodal joint training framework MMAudio. In contrast to single-modality training conditioned on (limited) video data only, MMAudio is jointly trained with larger-scale, readily available text-audio data to learn to generate semantically aligned high-quality audio samples. Additionally, we improve audio-visual synchrony with a conditional synchronization module that aligns video conditions with audio latents at the frame level. Trained with a flow matching objective, MMAudio achieves new video-to-audio state-of-the-art among public models in terms of audio quality, semantic alignment, and audio-visual synchronization, while having a low inference time (1.23s to generate an 8s clip) and just 157M parameters. MMAudio also achieves surprisingly competitive performance in text-to-audio generation, showing that joint training does not hinder single-modality performance. Code and demo are available at: https://hkchengrex.github.io/MMAudio
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。