Zero-Shot Audio-Visual Editing via
Cross-Modal Delta Denoising

Yan-Bo Lin1*
Kevin Lin2
Zhenyuan Yang2
Linjie Li2
Jianfeng Wang2
Chung-Ching Lin2
Xiaofei Wang2
Gedas Bertasius1
Lijuan Wang2
1 UNC Chapel Hill
2 Microsoft
* Work done during an internship at Microsoft
Paper Code Data
Key Challenges in Joint Audio-Video Editing. Existing methods primarily focus on zero-shot text-to-video or text-to-audio editing separately. Solely editing only video or only audio often leads to coherence and synchronization issues between two modalities. As highlighted in red circle, the motion or presence of sounding objects may not align with the corresponding audio. Additionally, edited content may exhibit audio artifacts along the temporal dimension (shown in the purple squares). These factors make the edited results feel less natural and cohesive. In contrast, our AVED jointly edits audio and video by leveraging cross-modal information as additional supervision to improve editing quality to alleviate synchronization issues.

All in one

Demo Videos

Source prompt:

machine gun

↓

Target prompt:

laser gun

Input Video

ControlVideo+ZEUS

TokenFlow+ZEUS

RAVE+ZEUS

AvED (Ours)

Source prompt:

a dog is howling

↓

Target prompt:

a lion is roaring

Source prompt:

a race car is driving on the road

↓

Target prompt:

a police car is driving on the road

Source prompt:

firework lit up in the night sky

↓

Target prompt:

a splash of water lit up in the night sky

BibTeX


        @inproceedings{lin2026aved,
          title={Zero-Shot Audio-Visual Editing via Cross-Modal Delta Denoising},
          author={Lin, Yan-Bo and Lin, Kevin and Yang, Zhengyuan and Li, Linjie and Wang, Jianfeng and Lin, Chung-Ching and Wang, Xiaofei and Bertasius, Gedas and Wang, Lijuan},
          booktitle={2026 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)},
          year={2026}
        }