The purpose of this study is to answer the need for efficient selection or composition/arrangement of cute music by clarifying the acoustic characteristics that contribute to the impression of cute music, including their contribution to the cute impression. Therefore, we first conducted a principal component analysis using Log-Mel spectrograms of three types of music, “cute music,” “energetic music,” and “sad music,” to roughly extract acoustic characteristics specific to “cute music”. The results indicated that cute music typically exhibits a relatively fast rhythm (about 120 bpm), the energy was concentrated within narrow frequency bands within a short period but the position of the bands is variable exhibited a broad frequency spread across entire compositions. These features were found in the visualized images of eigenvectors up to the third principal component of the cute music, with a cumulative contribution degree of 23.1% up to the third principal component. Furthermore, assuming that the above features can be represented by two indices, envelope value and spectral centroid, we evaluated these values and found that there was a significant difference in both indices for “cute music” compared to the other two types of music.

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

Acoustic Characteristics of Music that Contribute to Cute Impressions

  • Shino Narukawa,
  • Yumi Wakita

摘要

The purpose of this study is to answer the need for efficient selection or composition/arrangement of cute music by clarifying the acoustic characteristics that contribute to the impression of cute music, including their contribution to the cute impression. Therefore, we first conducted a principal component analysis using Log-Mel spectrograms of three types of music, “cute music,” “energetic music,” and “sad music,” to roughly extract acoustic characteristics specific to “cute music”. The results indicated that cute music typically exhibits a relatively fast rhythm (about 120 bpm), the energy was concentrated within narrow frequency bands within a short period but the position of the bands is variable exhibited a broad frequency spread across entire compositions. These features were found in the visualized images of eigenvectors up to the third principal component of the cute music, with a cumulative contribution degree of 23.1% up to the third principal component. Furthermore, assuming that the above features can be represented by two indices, envelope value and spectral centroid, we evaluated these values and found that there was a significant difference in both indices for “cute music” compared to the other two types of music.