基于音高旋律元的柔和乐曲的自动作曲算法

提出了一种以柔和为导向、以音高旋律元 (Pitch melody unit, PMU)为基本操作单位的自动作曲算法.首先根据柔和乐曲的旋律特征构建了算法的 规则模型.然后定义了一阶音高旋律元和二阶音高旋律元,构建了音高旋律元转换表 (Pitch melody unit transition tables, PMUTTs)自动生成的模型,以表达相邻音高旋律元的链接概率,为 旋律计算提供基本数据.分别定义了用来自动生成新的音高旋律元、乐节、乐句、乐曲的4个旋律计算的算子,以自动生成乐曲.实验结果表明, 与以音符为基本单位的马尔可夫作曲方法相比,以音高旋律元为基本单位进行作曲可以提高旋律的逻辑严密性和美感度.

Automated Composition Algorithm for Gentle Music Based on Pitch Melody Unit

In this paper, an automated music composition algorithm for gentle music is proposed based on pitch melody unit (PMU) which is the minimal operation unit. Firstly, several algorithmic rule models are constructed according to the melody characteristic of gentle music. Secondly, the 1st-order and 2nd-order PMUs are defined, and a model is further constructed to automatically generate pitch melody unit transition tables (PMUTTs) which are the expression of link probability of two adjacent PMUs. Then, the basic data for melody computing are obtained. Thirdly, four operators of melody computing are defined to generate a new PMU, phrase, sentence, and music, respectively. Finally, the gentle music can be produced automatically. Experimental results demonstrate that this PMU-based approach can make the obtained melody more logical and aesthetic than the Markov model approach which is based on operating notes.