毛康 林勇



摘 要: 为解决目前作曲系统存在作曲耗时长、匹配度低等问题,提出基于音型数据库结构的多特征作曲系统设计。首先,通过训练模块、作曲模块、数据库模块和用户界面模块完成对系统的总体框架设计,然后,通过处理器、存储卡、显示器、计算机和音响完成对作曲系统的硬件设计,最后,通过提取数据,对音型结构进行识别,再利用Matlab平台的函数和Java语言程序实现乐曲数据转换和输出,从而实现对作曲系统的软件设计,至此完成基于音型数据库结构的多特征作曲系统设计。实验结果表明,与传统的作曲系统相比,提出的多特征作曲系统的作曲耗时明显减少,匹配度可达到90%以上,表明提出的作曲系统具有更高的有效性。
关键词: 音型数据库; 多特征作曲系统; 系统设计; 作曲耗时; 数据处理; 对比验证
中图分类号: TN919?34; TP311 文献标识码: A 文章编号: 1004?373X(2020)08?0060?03
Design of multi?feature composition system based on figure database structure
MAO Kang1, LIN Yong2
(1. HuaiYin Normal University, Huaian 223300, China; 2. Chonbuk National University, Chonju 54896, Korea)
Abstract: A design of multi?feature composition system based on figure database structure is proposed to reduce the composition time?consumption and improve matching degree of the current composition system. The overall framework design of the system is completed by means of training module, composition module, database module and user interface module. The hardware design of the composition system is completed with the processor, memory card, display, computer and stereo. The figure structure is recognized by extracting data. The conversion and output of the music data are completed by means of the functions of the Matlab platform and the Java language program, so as to realize the software design of the composition system. The multi?feature composition system based on the figure database structure is completed. The experimental results show that, in comparison with the traditional composition system, the time?consumption of the proposed multi?feature composition system is significantly reduced, and its matching degree can reach more than 90%, which indicates that the proposed composing system has higher effectiveness.
Keywords: multi?feature composition system; system design; figure database; composition time?consumption; data processing; comparison validation
0 引 言
近年来,随着人们的物质生活水平提高,对精神文化的重视程度不断增加,人们对音乐也投入了更多的关注,越来越多的人开始进行音乐创作。作曲作为音乐创作活动的基础,对于音乐创作具有十分重要的指导作用,将科技与作曲相结合,为作曲带来了新的活力[1]。
目前,常见的作曲技术有修正型粒子群算法[1]、递归神经网络算法[2]、音乐文法[3]和遗传算法[4]等。由于在传统的作曲系统中,仍然存在一些不足,如作曲耗时较长、匹配度较低等。因此,对作曲系统进行创新具有十分重要的意义。
为解决以上存在的问题提出基于音型数据库结构的多特征作曲系统设计。……