Research map / Video / Video generation
EchoFoley: Event-Centric Hierarchical Control for Video Grounded Creative Sound Generation
Bingxuan Li, Yiming Cui, Yicheng He, Yiwei Wang, Shu Zhang, Longyin Wen, Yulei Niu
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 EchoVidia agentic framework instead of strongest VT2A baseline | HunyuanVideo-Foley-xxl · EchoFoley-6k | temporal controllability: 0.43 -> 0.72 | Improved |
| use EchoVidia instead of strongest baseline | HunyuanVideo-Foley-xxl · EchoFoley-6k | timbre controllability: 0.48 -> 0.78 | Improved |
| use EchoVidia instead of strongest baseline | HunyuanVideo-Foley · EchoFoley-6k | human instruction adherence: 2.60 -> 3.80 | Improved |
| use EchoVidia instead of strongest baseline | HunyuanVideo-Foley-xl · EchoFoley-6k | human audio-visual coherence: 4.20 -> 3.93 | Hurt |
| add Slow-Fast thinking (ablation) | Gemini-2.5 Pro · sounding event detection | F1: 0.59 -> 0.74 | Improved |
| add Slow-Fast thinking (ablation) | Qwen3-VL-30B-Thinking · sounding event detection | F1: 0.53 -> 0.71 | Improved |
| add Slow-Fast thinking (ablation) | Gemini-2.5 Pro · sounding event localization | IoU: 0.510 -> 0.842 | Improved |
| add Slow-Fast thinking (ablation) | Qwen3-VL-30B · sounding event localization | IoU: 0.484 -> 0.650 | Improved |
| increase control granularity from video to group to instance level | existing VT2A models · EchoFoley-6k | controllability | Hurt |
| use omni-modal models instead of vision-only VideoLLMs | sounding event detection · EchoFoley-6k | detection recall/F1: Qwen3-Omni-30B-Thinking F1 0.56 vs Qwen3-VL-30B-Instruct 0.49 | Improved |
Closest papers by meaning
- Read, Watch and Scream! Sound Generation from Text and Video (AAAI 2025 · similarity 0.88)
- Video-Guided Foley Sound Generation with Multimodal Controls (CVPR 2025 · similarity 0.88)
- ThinkSound: Chain-of-Thought Reasoning in Multimodal LLMs for Audio Generation and Editing (NeurIPS 2025 · similarity 0.87)
- Audio-Sync Video Generation with Multi-Stream Temporal Control (NeurIPS 2025 · similarity 0.86)
- AVGen-Bench: A Task-Driven Benchmark for Multi-Granular Evaluation of Text-to-Audio-Video Generation (ICML 2026 · 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.
