arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1902.01023cs.IR
收藏

通过特征级相似性融合增强分层音乐结构注释

Enhanced Hierarchical Music Structure Annotations via Feature Level Similarity Fusion

Christopher J. Tralie, Brian McFee

摘要:

我们描述了一种新的管道,自动发现音乐音频中重复部分的层次结构。该方法使用相似性网络融合(SNF)将不同的帧级特征组合成干净的亲和矩阵,然后将其用作谱聚类的输入。虽然以前的谱聚类方法的音乐结构分析有预处理的亲和矩阵与专门为此任务设计的算法,我们表明,SNF的方法直接产生的分割同意更好地与人类注释,作为衡量的“L-测量”度量层次注释。此外,SNF方法立即支持任意多个输入特征,允许我们同时发现以音色,和声和节奏表示编码的结构,而无需对基本算法进行任何更改。

英文摘要:

We describe a novel pipeline to automatically discover hierarchies of repeated sections in musical audio. The proposed method uses similarity network fusion (SNF) to combine different frame-level features into clean affinity matrices, which are then used as input to spectral clustering. While prior spectral clustering approaches to music structure analysis have pre-processed affinity matrices with heuristics specifically designed for this task, we show that the SNF approach directly yields segmentations which agree better with human annotators, as measured by the ``L-measure'' metric for hierarchical annotations. Furthermore, the SNF approach immediately supports arbitrarily many input features, allowing us to simultaneously discover structure encoded in timbral, harmonic, and rhythmic representations without any changes to the base algorithm.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑