项目名称: 自动作曲技术及其质量评估方法的研究
项目编号: No.61272238
项目类型: 面上项目
立项/批准年度: 2013
项目学科: 自动化技术、计算机技术
项目作者: 冯寅
作者单位: 厦门大学
项目金额: 36万元
中文摘要: 研究一种为汉语歌词自动谱曲,配钢琴伴奏的计算模型及其质量评估方法。该模型的作曲风格将通过一个基于统计的机器学习系统在一个样板乐谱集合上获取。研究在新近研发的钢琴自动伴奏系统的基础上进一步展开。研究内容涉及:1)汉语歌词的字数分布、韵律以及所涉及的情感或意境与各种谱曲技术对应关系的表达研究;2)创作风格的自动获取方法研究;3)机器音乐作品的美学及创新性评估方法的研究。成果以6-12篇论文形式发表并实现一个为歌词自动谱曲及配钢琴伴奏实验系统。通过本项目的研究,在理论上了解计算机模拟作曲家从事音乐创作这一特定思维的能力。同时可以探索如何利用机器寻求不同创作风格的各种美的形式。在应用上,预计所研发的实验系统在不断地增加样板乐谱的规模之后,将能实质性地提高音乐创作生产力。
中文关键词: 计算机音乐;自动作曲;智能系统;;
英文摘要: This project deals with the research on building a computational model on algorithmic composition & auto-harmonization, and the evaluation methodology on the machine-generated music, involving with the approach on a melody and piano accompaniment auto-generation for Chinese lyrics. A composing style will be achieved from a collected sample scores by means of a statistics-based machine learning technology. The project proposes a new research way on automatic composition and harmonization based on applicant's recent research issues on "meta-structure based technology" used in automatic piano accompaniment system. The research content will focus on three topics stated as follows: 1) how to build up and represent the relationships among the distribution of the number of Chinese character and the Chinese metrics, emotion and artistic conception described in the Chinese lyrics, and the composing techniques used in the selected sample-songs; 2) the technology on auto-composition and harmonization based on a given composition style; and 3) the evaluation methodology on machine-generated music in esthetics and creativity. The final achieving issues will be represented in 6-12 published technical papers. An experimental system on generating melody and piano accompaniment for Chinese lyrics will be implemented. This resear
英文关键词: Computer Music;Automatic Composition;Intelligence System;;