MRSAudio/MRSAudio
收藏Hugging Face2025-11-01 更新2025-11-01 收录
下载链接：
https://hf-mirror.com/datasets/MRSAudio/MRSAudio
下载链接
链接失效反馈官方服务：
资源简介：
MRSAudio是一个大规模的多模态录音空间音频数据集，包含500小时的多模态空间音频，整合了高保真的空间录音与同步视频、3D姿态追踪和丰富的语义注释，能够对真实世界的听觉场景进行综合建模。数据集包括四个子集：MRSLife（日常生活活动），MRSSpeech（演讲对话），MRSSing（独唱表演）和MRSMusic（乐器演奏），分别针对不同的任务和场景。
MRSAudio is a large-scale multimodal recorded spatial audio dataset containing 500 hours of multimodal spatial audio. It integrates high-fidelity spatial recordings with synchronized video, 3D pose tracking, and rich semantic annotations, enabling comprehensive modeling of real-world auditory scenes. The dataset includes four subsets: MRSLife (daily life activities), MRSSpeech (speech conversations), MRSSing (solo singing performances), and MRSMusic (instrumental music), each targeting different tasks and scenarios.
提供机构：
MRSAudio
