A Mobile Melody Creator Based on Genetic Algorithms and Motif-Driven Music Form Theory

• Main Authors: Yue-Zong Gu, 古岳宗
• Other Authors: Jian-Hung Chen
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/42668351515223941715

碩士 === 中華大學 === 資訊工程學系碩士班 === 100 === To create a generally accepted music often requires a lot of time and inspirations/motif of composers. Although anything of daily life may become a source of inspiration/motif for composers, while composers are inspired by any treasure inspiration/motif, they may not have a handy creation tool to turn inspiration into any music forms. Therefore, lacking a handy creation tool has already wasted many opportunities for composers. In recent decades, many automatic music composers in the field of musical research can produce interesting melodies. However, they suffer the following criticizes. First, the portability of those automatic composers is very limited; most of them are not portable for composers. Thus, once a composer has an inspiration/motif at anywhere in any moment, one cannot produce it into music in the real-time manner. Immediate inspiration (motif) cannot take into account in music creation. Second, the music created by those automatic music composers may be biased and subjective, and are often lack of art creation and variation, because the music creation of those composers is based on its own music library. Third, some interactive composers require iterative evaluations of user opinions, thus users may feel boredom after a long-term usage of composers. As a result, these automatic music composers are not suitable for entry users, and even composers and musicians. In this thesis, an automatic music creator based genetic algorithms and motif-driven music form theory is developed in Android platform. Once inspiration/motif occurs, users can use their mobile device to turn inspiration/motif into music, at anywhere in anytime. Users can pick up their mobile phone and click/slide simple actions on their phone as the inputs of inspiration/motif. In such way, portability is no longer an issue for composers. Furthermore, two fitness functions based on motif-driven music theory are designed and combined with a genetic algorithm in our music creator, so that high music variation and wonderful music innovation could be achieved automatically without interactions with users. In conse-quences, the developed music creator in this thesis can easily generate nice music melodies in MIDI format in the real-time manners. It is a handy computer-aided music creator for composers, musicians, and entry users. Anyone without music background can use this creator to create music and enjoy the interests of music compositions.