The copyright of this material is retained by the Information Processing Society of Japan (IPSJ). The material has been made available on the website by the author(s) under the agreement with the IPSJ. Please be complied with Copyright Law of Japan and Code of Ethics of the IPSJ if any users wish to reproduce, distribute, or modify the whole or any part of the material.
,a),b) Significance of melodic information in section identification tasks of popular music Hibiki Miyazawa,a) Yuzuru Hiraga,b) Abstract: Popular music typically consist of sections called verse-a, verse-b, and chorus. These sections can be identified to a certain extent based on their inherent music properties alone. However, what features of the music (melody line, harmony/chord, accompaniment, lyrics, etc.) are dominant in such judgments are still unclear. In this study, we conducted a psychological experiment which tried to identify whether melody line information alone is sufficient for section identification. The participants were presented with melody lines extracted from unknown songs, and were asked to identify whether the segment was from verse-a, verse-b or chorus. The results show that the participants performed above chance level and sections were identified sufficiently, suggesting that melodic information include significant cues for section identification. Keywords: section, melody line, popular music, chorus, catchiness. [][2] Graduate School of Library, Information and Media Studies, University of Tsukuba a) s82634@s.tsukuba.ac.jp b) hiraga@slis.tsukuba.ac.jp 2 2 [3] 90% c 209 Information Processing Society of Japan
[4] RWC [5] 30 AIST Annotation for the RWC Music Database[6] 8 90% 8 80% Table maximum and minimum lengths of sections ( ) 39.45 26.3 39.8 ( ).29 0.00 2.90 2. 2. RWC [5] 30 AIST Annotation for the RWC Music Database[6] 30 3 = 90 [6] SMF MIDI No. GarageBand wav 2.2 20 25 8 9 9 22.3 2. 90 90 8 PC 500ms PC PC c 209 Information Processing Society of Japan 2
3. 3. F (2, 34) = 2.066, p >.05 p <.00 33.3% 3.2 2 F (2, 34) = 20.649, p <.00 Bonferroni (p <.05) (p <.0) (p <.00) Fig. percentage of correct answers 3.3 3 F (2, 34) = 57.964, p <.00 Bonferroni (p <.00) 3.4 4 5 6 2 3 4 0.8 3.5 7 8 9 30 A B C ID. chorus 5 22.6 22.6% Fig. 2 2 degree of confidence 3 Fig. 3 catchiness c 209 Information Processing Society of Japan 3
情報処理学会研究報告 図 4 散布図 Fig. 4 scatter plot (percentage of correct answers - degree of confidence) 図 7 誤答の内訳 Aメロ Fig. 7 incorrect answers (verse-a) 図 5 散布図 Fig. 5 scatter plot (percentage of correct answers - catchiness) 図 8 誤答の内訳 Bメロ Fig. 8 incorrect answers (verse-b) 図 6 散布図 Fig. 6 scatter plot (degree of confidence - catchiness) 表 2 相関係数 Aメロ Table 2 correlation coefficients (verse-a) Aメロ -0.27-0.48 0.64 表 3 相関係数 Bメロ Table 3 correlation coefficients (verse-b) Bメロ -0.09 0.2 0.79 図 9 表 5 表 4 相関係数 サビ Table 4 correlation coefficients (chorus) サビ 0.82 0.8 誤答の内訳 サビ Fig. 9 incorrect answers (chorus) 誤答率 Table 5 percentage of incorrect answers 回答 0.80 c 209 Information Processing Society of Japan 刺激 Aメロ Bメロ サビ Aメロ 22.6 25.0 Bメロ 42.0 0.0 サビ 23.3 7.6 4
3.6 0 melo abs F (, 34) = 98.240, p <.00 F (2, 68) =.855, p >.05 F (2, 68) =.53, p >.05 80% 70% 6 4. 4. 3. 3.6 Fig. 0 Fig. 0 percentage of correct answers compared to absolute evaluation 80 6 percentage of correct answers of the stimulation answered more than 80% in absolute evaluation Table 6 percentage of incorrect answers compared to absolute evaluation 22.6 25.0 42.0 0.0 23.3 7.6 4.6 7.8 5.9 6.9 0.4. 2. c 209 Information Processing Society of Japan 5
4.2 3.2 3. 4.3 3.3 4.4 3.4 0.8 3.5 4.5 3.5 3.4 5. RWC [] Goto, M.: SmartMusicKIOSK: Music Listening Station with Chorus-Search Function, Proc. 6th Annual ACM Symposium on User Interface Software and Technology, pp. 3 40 (2003). [2] Vol. 205-MUS-08, No. 7, pp. 9 (205). [3] Vol. 2000-MUS-35-, No. 49, pp. 6 (2000). [4] Vol. 208- MUS-8, No. 27, pp. 6 (208). [5] RWC Vol. 200- MUS-42-6, No. 03, pp. 35 42 (200). [6] Goto, M.: AIST Annotation for the RWC Music Database, Proc. 7th International Conference on Music Information Retrieval, pp. 359 360 (2006). c 209 Information Processing Society of Japan 6