Research map / Video / Video generation
Audio-Synchronized Visual Animation
Lin Zhang, Shentong Mo, Yijing Zhang, Pedro Morgado
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add audio conditioning to the image-to-video model (vs I2VD) (ablation) | Ours · AVSync15 test | AlignSync: 21.80 -> 22.25 | Improved |
| increase audio guidance factor from 1 to 4 (ablation) | Ours · AVSync15 test | FVD: 382.7 -> 349.1 | Improved |
| increase audio guidance factor from 1 to 4 (ablation) | Ours · Landscapes | RelSync: 49.96 -> 49.93 | No clear effect |
| raise guidance from 4 to 8 (ablation) | Ours · AVSync15 test | FVD: 349.1 -> 420.7 | Hurt |
| add first-frame temporal convolution (ablation) | Ours · AVSync15 test | FVD: 383.3 -> 347.4 | Improved |
| add first-frame spatial attention on top of FF-Conv (ablation) | Ours · AVSync15 test | FVD: 347.4 -> 325.6 | Improved |
| add first-frame spatial attention on top of FF-Conv (ablation) | Ours · AVSync15 test | FID: 11.6 -> 11.8 | No clear effect |
| train on automatically curated AVSync-AC instead of raw VGGSS (ablation) | Ours · AVSync15 test | FVD: 1307.9 -> 428.8 | Improved |
| add manual curation on top of automatic curation (ablation) | Ours · AVSync15 test | FVD: 428.8 -> 325.6 | Improved |
| increase audio amplitude 100x instead of audio guidance (ablation) | Ours · qualitative | control quality | Hurt |
| add class-name text conditioning | Ours · AVSync15 | quality | No clear effect |
Closest papers by meaning
- Syncphony: Synchronized Audio-to-Video Generation with Diffusion Transformers (ICLR 2026 · similarity 0.92)
- Audio-Sync Video Generation with Multi-Stream Temporal Control (NeurIPS 2025 · similarity 0.88)
- Read, Watch and Scream! Sound Generation from Text and Video (AAAI 2025 · similarity 0.87)
- Animate and Sound an Image (CVPR 2025 · similarity 0.87)
- SonicVisionLM: Playing Sound with Vision Language Models (CVPR 2024 · similarity 0.87)
- 🎧MOSPA: Human Motion Generation Driven by Spatial Audio (NeurIPS 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
