Research map / Video / Video generation
Harmony: Harmonizing Audio and Video Generation through Cross-Task Synergy
Teng Hu, Zhentao Yu, Guozhen Zhang, Zihan Su, Zhengguang Zhou, Youliang Zhang, Yuan Zhou, Qinglin Lu, Ran Yi
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 global-local decoupled interaction module (ablation) | Harmony baseline · Harmony-Bench speech | Sync-C: 4.20 -> 4.29 | Improved |
| add RoPE alignment (ablation) | Harmony · Harmony-Bench speech | Sync-C: 4.29 -> 4.80 | Improved |
| add cross-task synergy training (ablation) | Harmony · Harmony-Bench speech | Sync-C: 4.80 -> 5.09 | Improved |
| add SyncCFG at inference (ablation) | Harmony · Harmony-Bench speech | Sync-C: 5.09 -> 6.51 | Improved |
| add SyncCFG at inference (ablation) | Harmony · Harmony-Bench speech | Sync-D: 10.16 -> 8.63 | Improved |
| use Harmony instead of Ovi | Harmony · Harmony-Bench average | Sync-C: 4.04 -> 5.61 | Improved |
| use Harmony instead of Ovi | Harmony · Harmony-Bench average | Sync-D: 9.62 -> 7.53 | Improved |
| use Harmony instead of Ovi | Harmony · Harmony-Bench average | IB: 0.18 -> 0.19 | No clear effect |
| use Harmony instead of Ovi | Harmony · Harmony-Bench average | Imaging Quality: 0.65 -> 0.65 | No clear effect |
| train with audio-driven task (clean audio) vs joint generation only (ablation) | Harmony · lip-sync convergence | Sync confidence | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- JavisDiT — JavisDiT: Joint Audio-Video Diffusion Transformer with Hierarchical Spatio-Temporal Prior Synchronization (ICLR 2026) · better on 8 of 8 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Native Audio-Visual Alignment for Generation (NeurIPS 2026 · similarity 0.90)
- MMAudio: Taming Multimodal Joint Training for High-Quality Video-to-Audio Synthesis (CVPR 2025 · similarity 0.89)
- PrismAudio: Decomposed Chain-of-Thought and Multi-dimensional Rewards for Video-to-Audio Generation (ICLR 2026 · similarity 0.88)
- Seeing and Hearing: Open-domain Visual-Audio Generation with Diffusion Latent Aligners (CVPR 2024 · similarity 0.88)
- Foley-Flow: Coordinated Video-to-Audio Generation with Masked Audio-Visual Alignment and Dynamic Conditional Flows (CVPR 2025 · similarity 0.88)
- Model-Guided Dual-Role Alignment for High-Fidelity Open-Domain Video-to-Audio Generation (NeurIPS 2025 · similarity 0.88)
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.
