面向抒情歌曲旋律的钢琴自动伴奏算法

发布时间：2018-03-15 02:29

本文选题：自动伴奏　切入点：抒情歌曲　出处：《河南师范大学学报(自然科学版)》2016年04期 　论文类型：期刊论文

【摘要】：目前钢琴自动伴奏中存在着一定的不和谐问题,其中一个重要原因是由于和弦处理的固定性往往不能满足旋律的不确定性要求.为此,提出一种面向拍号固定的抒情歌曲旋律的自适应伴奏算法.首先,根据拍号对旋律进行小节分割和乐句整合,在传统三度关系的三和弦基础上定义更泛化的三元和弦,自适应获取三元和弦的内部音程关系;然后,自适应计算旋律音对和弦的贡献分值,以获取小节中所有可能的三元和弦;自动选取和弦个数,将选出的和弦与知识库进行匹配,将匹配成功的和弦添加合适的伴奏音型,最终生成伴奏.实验结果表明,人们不容易将该算法生成的伴奏与人工的伴奏进行区分.
[Abstract]:At present, there are some problems of disharmony in piano automatic accompaniment, one of the important reasons is that the fixation of chord processing can not meet the requirements of melody uncertainty. In this paper, an adaptive accompaniment algorithm for the melody of lyric songs with fixed racquet is proposed. Firstly, a more generalized ternary chord is defined on the basis of the traditional three-degree relationship, which is based on the segmentation of the melody and the integration of the phrases according to the beat horn. The internal interval relation of ternary chords is acquired adaptively. Then, the contribution score of melodic sounds to chords is calculated adaptively to obtain all possible ternary chords in a section. The number of chords is automatically selected to match the selected chords with the knowledge base. It is difficult to distinguish the accompaniment generated by the algorithm from the artificial accompaniment by adding the appropriate accompaniment type to the matching chord and finally generating the accompaniment. The experimental results show that it is not easy to distinguish the accompaniment generated by the algorithm from the artificial accompaniment.
【作者单位】：河南师范大学计算机与信息工程学院;河南师范大学"智慧商务与物联网技术"河南省工程实验室;
【基金】：国家自然科学基金(61402153) 河南省重点科技攻关项目(122102210054) 河南省高校青年骨干教师项目(2011GGJS-061)
【分类号】：J624.1

【相似文献】