论文部分内容阅读
目前数字音乐表现的内容愈加广泛并且表现形式愈加多样化,从传统意义上的通过标签等人工标注的方式很难满足用户的需求。当用户在听了一首歌的时候,可能会喜欢曲风或者旋律类似的歌曲,因此可以从歌曲本身的特征出发,找到类似的歌曲给用户做个性化推荐。然而常见的MP3格式的音乐并不适合直接分析处理,而MIDI作为特殊的格式,为特征提取提供了便利。但是作为非音乐专业的人员分析MIDI格式还是有一定难度的,本文介绍了利用当前流行的编程语言来自动化处理midi文件,抽取出旋律特征,这样可以为基于特征的推荐做准备。
At present, the content of digital music performance is more extensive and the forms of expression are more diversified. It is difficult to meet the needs of users from the traditional way of marking by hand, such as tags. When the user listens to a song, he or she may like songs of the genre or melody, so that similar songs can be found starting from the characteristics of the song and personalized to the user. However, the common MP3 format music is not suitable for direct analysis and processing, while MIDI as a special format provides convenience for feature extraction. However, as a non-musician, there is still some difficulty in analyzing the MIDI format. This article describes using the currently popular programming language to automate the processing of midi files, extracting melodic features, and preparing for feature-based recommendations.