11 papers
FLOMSS: Generative Music Source Separation by Flow Matching
Shuyang Cui, Zhi Zhong, Keisuke Toyama +10
Music source separation (MSS) has been conventionally addressed by discriminative methods. Recently, flow matching has emerged as a strong generative paradigm for modeling complex…
AVIO: Learning to Add and Remove Sounding Objects in Audiovisual Scenes
Weihan Xu, Kan Jen Cheng, Koichi Saito +8
Adding or removing a sounding object requires coordinated changes to visual content and sound while preserving the surrounding scene. Yet paired supervision for localized non-speec…
Odoriko: A Shape-Aware Multimodal Diffusion Framework for Human Motion
Dongseok Shim, Julian Tanke, Kengo Uchida +5
Human motion generation has been widely studied across diverse input modalities, text, music, and video, and recent efforts have unified these into single multimodal frameworks. Ho…
Echoes Over Time: Unlocking Length Generalization in Video-to-Audio Generation Models
Christian Simon, Masato Ishii, Wei-Yao Wang +8
Scaling multimodal alignment between video and audio is challenging, particularly due to limited data and the mismatch between text descriptions and frame-level video information.…
Schrodinger Audio-Visual Editor: Object-Level Audiovisual Removal
Weihan Xu, Kan Jen Cheng, Koichi Saito +10
Joint editing of audio and visual content is crucial for precise and controllable content creation. This new task poses challenges due to the limitations of paired audio-visual dat…
FoleyBench: A Benchmark For Video-to-Audio Models
Satvik Dixit, Koichi Saito, Zhi Zhong +2
Video-to-audio generation (V2A) is of increasing importance in domains such as film post-production, AR/VR, and sound design, particularly for the creation of Foley sound effects s…
TalkCuts: A Large-Scale Dataset for Multi-Shot Human Speech Video Generation
Jiaben Chen, Zixin Wang, Ailing Zeng +8
In this work, we present TalkCuts, a large-scale dataset designed to facilitate the study of multi-shot human speech video generation. Unlike existing datasets that focus on single…
SoundReactor: Frame-level Online Video-to-Audio Generation
Koichi Saito, Julian Tanke, Christian Simon +7
Prevailing Video-to-Audio (V2A) generation models operate offline, assuming an entire video sequence or chunks of frames are available beforehand. This critically limits their use…
Dyadic Mamba: Long-term Dyadic Human Motion Synthesis
Julian Tanke, Takashi Shibuya, Kengo Uchida +2
Generating realistic dyadic human motion from text descriptions presents significant challenges, particularly for extended interactions that exceed typical training sequence length…
DisMix: Disentangling Mixtures of Musical Instruments for Source-level Pitch and Timbre Manipulation
Yin-Jyun Luo, Kin Wai Cheuk, Woosung Choi +8
Existing work on pitch and timbre disentanglement has been mostly focused on single-instrument music audio, excluding the cases where multiple instruments are presented. To fill th…
SpecMaskGIT: Masked Generative Modeling of Audio Spectrograms for Efficient Audio Synthesis and Beyond
Marco Comunità, Zhi Zhong, Akira Takahashi +7
Recent advances in generative models that iteratively synthesize audio clips sparked great success to text-to-audio synthesis (TTA), but with the cost of slow synthesis speed and h…