SymMV
收藏资源简介:
SymMV是由北京航空航天大学创建的第一个包含丰富音乐注释的视频与符号音乐数据集,旨在解决视频背景音乐自动生成的问题。该数据集包含1140个视频-音乐对,涵盖超过10种音乐风格,总时长76.5小时。SymMV不仅提供高质量的音乐,还包括和弦、旋律、伴奏等详细注释,支持模型分阶段控制音乐生成,适用于视频背景音乐生成及其他音乐生成任务。
SymMV is the first video and symbolic music dataset with rich musical annotations developed by Beihang University, which aims to address the problem of automatic background music generation for videos. This dataset contains 1140 video-music pairs, covering more than 10 music genres, with a total duration of 76.5 hours. Besides high-quality music tracks, SymMV also provides detailed annotations such as chords, melodies, accompaniments and other relevant information, enabling staged control of music generation by models, and is applicable to video background music generation and other music generation tasks.




