AI Paper Atlas

Research map / Video / Video generation

Taming Text-to-Sounding Video Generation via Advanced Modality Condition and Interaction

Kaisi Guan, Xihua Wang, Zhengfeng Lai, Xin Cheng, Peng Zhang, Xiaojiang Liu, Ruihua Song, Meng Cao

ECCV 2026 · Video › Video generation · approach: Video-to-audio synthesis · 12 citations (OpenAlex)

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 changedWhereMeasuredEffect
add cross-modal interaction (DCA) to independent Wan+SDA towersWan+SDA · AVSync15AV-Align: 0.205 -> 0.275Improved
use disentangled CRR captions instead of a shared audio-video caption (ablation)BridgeDiT · AVSync15FVD: 1362.83 -> 765.74Improved
use disentangled CRR captions instead of video-only shared caption (ablation)BridgeDiT · AVSync15CLAP: 9.67 -> 35.95Improved
add CRR rewriting instead of raw LLM captions (ablation)BridgeDiT · AVSync15FAD: 19.42 -> 5.34Improved
add semantic anchors to the rewriter (ablation)CRR · AVSync15FAD: 15.74 -> 5.34Improved
add semantic anchors to the rewriter (ablation)CRR · AVSync15AV-Align: 0.224 -> 0.275Improved
add inference-time prompt expansion (ablation)BridgeDiT · AVSync15FVD: 1463.81 -> 765.74Improved
use dual cross-attention instead of full attention fusion (ablation)BridgeDiT · AVSync15AV-Align: 0.253 -> 0.275Improved
use bidirectional DCA instead of A2V cross-attention only (ablation)BridgeDiT · AVSync15AV-Align: 0.236 -> 0.275Improved
use bidirectional DCA instead of A2V cross-attention only (ablation)BridgeDiT · AVSync15FVD: 746.37 -> 765.74Hurt
use uniform early-bias block placement instead of late layers (ablation)BridgeDiT · AVSync15AV-Align: 0.1831 -> 0.2746Improved
add DCA to SANA+AudioLDM backbonesSANA+AudioLDM · AVSync15FAD: 17.09 -> 9.46Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy