Research map / Video / Video generation
Towards Scene-Aware Video-to-Spatial Audio Generation
Jaeyeon Kim, Heeseung Yun, Gunhee Kim
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 |
|---|---|---|---|
| use end-to-end ViSAGe (directional) instead of Diff-Foley + Audio Spatial. two-stage | ViSAGe · YT-Ambigen test | CC (all): 0.577 -> 0.721 | Improved |
| add visual guidance to directional guidance (ablation) | ViSAGe · YT-Ambigen test | FAD_dec: 5.56 -> 3.86 | Improved |
| add visual guidance to directional guidance (ablation) | ViSAGe · YT-Ambigen test | CC (all): 0.721 -> 0.635 | Hurt |
| add direction embedding (ablation) | ViSAGe (PT) · YT-Ambigen test | CC (all): 0.430 -> 0.524 | Improved |
| add rotation augmentation (ablation) | ViSAGe · YT-Ambigen test | CC (all): 0.531 -> 0.635 | Improved |
| add patchwise energy map (ablation) | ViSAGe (PT+DIR) · YT-Ambigen test | CC (all): 0.524 vs 0.531 | Mixed |
| add VGGSound pretraining (ablation) | ViSAGe (DIR+PE) · YT-Ambigen test | FAD_dec: 4.44 -> 4.01 | Improved |
| use proposed code pattern instead of sequential delay (ablation) | ViSAGe · YT-Ambigen test | FAD_dec: 10.75 -> 4.44 | Improved |
| use proposed pattern instead of residual only (ablation) | ViSAGe · YT-Ambigen test | CC (all): 0.480 -> 0.544 | Improved |
| use proposed pattern instead of spatial only (ablation) | ViSAGe · YT-Ambigen test | FAD_dec: 5.89 -> 4.44 | Improved |
| joint directional+visual guidance instead of dual 7.5 & 2.5 (ablation) | ViSAGe · YT-Ambigen test | CC (all): 0.596 -> 0.635 | Improved |
| train on YT-Ambigen instead of YT360 | video-to-audio model · dataset comparison | FAD: 15.91 -> 3.95 | Improved |
Closest papers by meaning
- ViSAGe: Video-to-Spatial Audio Generation (ICLR 2025 · similarity 0.99)
- OmniAudio: Generating Spatial Audio from 360-Degree Video (ICML 2025 · similarity 0.89)
- Both Ears Wide Open: Towards Language-Driven Spatial Audio Generation (ICLR 2025 · similarity 0.88)
- MRSAudio: A Large-Scale Multimodal Recorded Spatial Audio Dataset with Refined Annotations (NeurIPS 2025 · similarity 0.86)
- Read, Watch and Scream! Sound Generation from Text and Video (AAAI 2025 · similarity 0.86)
- SonicVisionLM: Playing Sound with Vision Language Models (CVPR 2024 · 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.
