Research map / Video / Video generation
Benchmarking Single-Factor Physical Video-to-Audio Generation
Tingle Li, Siddharth Gururani, Kevin J. Shih, Gantavya Bhatt, Sang-gil Lee, Zhifeng Kong, Arushi Goel, Gopala Anumanchipalli, Ming-Yu Liu
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 text captions at inference (ablation) | MMAudio · FlatSounds-Physics | physical Confidence: 0.221 -> 0.226 | Mixed |
| add text captions at inference (ablation) | MMAudio-Phys · FlatSounds-Physics | physical Confidence: 0.289 -> 0.306 | Improved |
| add text captions at inference (ablation) | FoleyCrafter · FlatSounds-Physics | physical Confidence: 0.217 -> 0.205 | Hurt |
| add text captions at inference (ablation) | ThinkSound · FlatSounds-Physics | physical Confidence: 0.219 -> 0.228 | Improved |
| add text captions at inference (ablation) | Hunyuan-V2A · FlatSounds-Physics | physical Confidence: 0.296 -> 0.305 | Improved |
| add text captions at inference (ablation) | Hunyuan-V2A · FlatSounds-Single | Hit Coverage (%): 68.55 -> 65.21 | Hurt |
| add text captions at inference (ablation) | MMAudio-Phys · FlatSounds-Single | Hit Coverage (%): 56.46 -> 50.69 | Hurt |
| add text captions at inference (ablation) | Hunyuan-V2A · FlatSounds-Single | Timing Error (ms): 44.34 -> 44.76 | Hurt |
| add text captions at inference (ablation) | Hunyuan-V2A · VGGSound | DeSync: 0.326 -> 0.340 | Hurt |
| add text captions at inference (ablation) | MMAudio-Phys · VGGSound | FAD-PASST: 66.31 -> 54.73 | Improved |
| add text captions at inference (ablation) | Hunyuan-V2A · VGGSound | Inception Score: 7.59 -> 15.29 | Improved |
| fine-tune MMAudio with physics-aware captions (MMAudio-Phys) | MMAudio · FlatSounds-Physics | physical Confidence: 0.226 -> 0.306 | Improved |
Closest papers by meaning
- PrismAudio: Decomposed Chain-of-Thought and Multi-dimensional Rewards for Video-to-Audio Generation (ICLR 2026 · similarity 0.88)
- AVGen-Bench: A Task-Driven Benchmark for Multi-Granular Evaluation of Text-to-Audio-Video Generation (ICML 2026 · similarity 0.86)
- Read, Watch and Scream! Sound Generation from Text and Video (AAAI 2025 · similarity 0.86)
- ThinkSound: Chain-of-Thought Reasoning in Multimodal LLMs for Audio Generation and Editing (NeurIPS 2025 · similarity 0.86)
- Seeing and Hearing: Open-domain Visual-Audio Generation with Diffusion Latent Aligners (CVPR 2024 · similarity 0.86)
- VinTAGe: Joint Video and Text Conditioning for Holistic Audio Generation (CVPR 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.
