Scholarlink

This study proposes a method for automatically generating piano arrangements of three-part female choral music for solo performance. The main melody is assigned to the right-hand part using phrase segmentation and multi-feature analysis, while the left-hand accompaniment is generated based on rhythmic patterns. A random forest classifier is used to assess the importance of each feature. Experimental results show that the proposed method achieves over 80% accuracy in melody estimation across most pieces. These results highlight the effectiveness of phrase segmentation and the importance of pitch-related features in choral melody extraction.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Automatic Piano Arrangement for Three-Part Choral of Female Voices

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Kana Yamada,
  • Aiko Uemura,
  • Norimasa Yoshida

摘要

This study proposes a method for automatically generating piano arrangements of three-part female choral music for solo performance. The main melody is assigned to the right-hand part using phrase segmentation and multi-feature analysis, while the left-hand accompaniment is generated based on rhythmic patterns. A random forest classifier is used to assess the importance of each feature. Experimental results show that the proposed method achieves over 80% accuracy in melody estimation across most pieces. These results highlight the effectiveness of phrase segmentation and the importance of pitch-related features in choral melody extraction.