摘要
针对在MIDI演奏系统中如何评价用户的演奏效果的问题,本文提出一种更完善的旋律匹配算法来计算两段旋律之间的相似度。该算法主要分两步:第一,基于音乐的基本元素:音高、节奏、音强这三个角度,提取出用户通过MIDI设备演奏的旋律和系统的标准旋律的特征序列;第二,分别计算这两个序列在音高、节奏、音强这三个方面的匹配系数,然后根据不同的系数加权综合得到一个总的匹配系数。该系数能够更客观地反应用户对所弹旋律的熟悉程度和在伴奏下的表演效果。
Point to evaluate users' performance effect in the MIDI performance system, a better matching algorithm is approached to compute the similarity of two pieces of melody. The approach is composed of two different stages : the first stage is dedicated to extracting feature sequence of melody that is played by users with MIDI instruments and the characteristic one of standard melody in terms of musical basic elements-pitch, rhythm and intensity. While in the second stage, the pitch, rhythm, intensity matching coefficients are computed in this two feature sequences, and then get the total matching coefficient by weighted synthesizing this three matching coefficients. This total matching coefficient can objectively reflect how familiar the users are with a melody and performance effect in the background of the accompaniment.
出处
《计算机与现代化》
2009年第6期151-154,157,共5页
Computer and Modernization
关键词
相似度
音高
节奏
音强
匹配系数
加权综合
similarity
pitch
rhythm
intensity
matching coefficient
weighted synthesis