ST-AudioQA, a spatio-temporal audio QA dataset and benchmark built from first-order ambisonic (FOA) renderings of static and moving sound sources, and ST-Audio Encoder, which jointly learns event semantics and time-resolved source trajectories, and ST-AudioLM, which exposes these audio tokens to an LLM for spatio-tempo...
Text-guided long-term human motion generation is essential for applications such as continuously operating digital avatars. Despite recent progress on short clips, existing models struggle with multi-segment instructions that involve anaphora, where constraints established in earlier segments must be preserved over tim...
Shun Kato, Di-Xin Yang, Kotaro Amaya et al.· 0 citations
A pretrained text-to audiovisual generation model is adapted through source-conditioned feature modulation to jointly learn object addition and removal, and quantitative and qualitative evaluations demonstrate effective audiovisual object removal and addition.
Wei-Han Xu, Kan-Jen Cheng, Koichi Saito et al.· 0 citations
Spectral Alignment is proposed, a lightweight, guidance-based method that calibrates the power spectrum of intermediate predictions to a pre-computed prior and is complementary to Classifier-Free Guidance (CFG).
