Research map / Image generation / Text-to-image diffusion
SounDiT: Geo-Contextual Soundscape-to-Landscape Generation
Junbo Wang, Haofeng Tan, Bowen Liao, Albert Jiang, Teng Fei, Qixing Huang, Bing Zhou, Zhengzhong Tu, Shan Ye, Yuhao Kang
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 SLRCM and S-AdaLN scene conditioning (ablation) | SounDiT · SoundingSVI (169K), component ablation (2 experts) | FID: 25.375 -> 19.195 | Improved |
| add SLRCM and S-AdaLN scene conditioning (ablation) | SounDiT · SoundingSVI (169K), component ablation (2 experts) | PSS Scene: 0.428 -> 0.734 | Improved |
| add SLRCM (ablation) | SounDiT · SoundingSVI (169K), component ablation (2 experts) | FID: 20.335 -> 19.195 | Improved |
| add SLRCM (ablation) | SounDiT · SoundingSVI (169K), component ablation (2 experts) | IIS: 0.728 -> 0.750 | Improved |
| add S-AdaLN (ablation) | SounDiT · SoundingSVI (169K), component ablation (2 experts) | FID: 23.435 -> 19.195 | Improved |
| add S-AdaLN (ablation) | SounDiT · SoundingSVI (169K), component ablation (2 experts) | PSS Scene: 0.572 -> 0.734 | Improved |
| increase MoE experts from 2 to 8 (ablation) | SounDiT · SoundingSVI (169K), expert scaling | FID: 19.195 -> 16.839 | Improved |
| increase MoE experts from 2 to 8 (ablation) | SounDiT · SoundingSVI (169K), expert scaling | PSS Scene: 0.734 -> 0.753 | Improved |
| replace PixArt + MHCA with SounDiT | PixArt + MHCA · SoundingSVI (169K) | FID: 34.108 -> 16.839 | Improved |
| replace PixArt + MHCA with SounDiT | PixArt + MHCA · SonicUrban (237K) | FID: 41.456 -> 11.553 | Improved |
| use SounDiT instead of pretrained CoDi on audio-image similarity | CoDi · SoundingSVI (169K) | AIS: 0.641 -> 0.538 | Hurt |
| use SounDiT instead of Sound2Scene on perception PSS | Sound2Scene · SoundingSVI (169K) | PSS Perception: 0.735 -> 0.729 | No clear effect |
Closest papers by meaning
- Audible World Models: Spatially Aware Sound Generation for 3D Worlds (NeurIPS 2026 · similarity 0.88)
- Both Ears Wide Open: Towards Language-Driven Spatial Audio Generation (ICLR 2025 · similarity 0.86)
- SoundVista: Novel-View Ambient Sound Synthesis via Visual-Acoustic Binding (CVPR 2025 · similarity 0.85)
- SoundBrush: Sound as a Brush for Visual Scene Editing (AAAI 2025 · similarity 0.85)
- MRSAudio: A Large-Scale Multimodal Recorded Spatial Audio Dataset with Refined Annotations (NeurIPS 2025 · similarity 0.84)
- NeRAF: 3D Scene Infused Neural Radiance and Acoustic Fields (ICLR 2025 · similarity 0.84)
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.
