1,a) 2,b) 1,c) LPMCC MFCC Fluctuation Pattern (LDA) Songle Pitman-Yor (VPYLM) 3278 1. (MIR: Music Information Retrieval) [1 5] [6 8] 1 National Institute of Advanced Industrial Science and Technology (AIST) 2 Kyoto University a) t.nakano [at] aist.go.jp b) yoshii [at] i.kyoto-u.ac.jp c) m.goto [at] aist.go.jp *1 N *2 [9] *1 http://www.nicovideo.jp/ *2 cfl 2014 Information Processing Society of Japan 1
OngaCREST [10] 4 2. 1 1 A 3. Latent Dirichlet Allocation: LDA [11] Songle [12] Pitman-Yor (VPYLM) [13] [14,15] n n n 3.1 [16 18] 3.1.1 PreFEst [19] F 0 20 32 ms LPMCC 12 ΔF 0 1 10ms GMM [16] 15% 16kHz LPMCC LPC MFCC cfl 2014 Information Processing Society of Japan 2
LPC 25 15 ΔF 0 50ms GMM GMM RWC [20] 100 80 20 32 1.0 GMM 12 GMM 27 3.1.2 LDA [17,18] k-means RWC 100 k = 100 LDA 100 Gibbs [21] 1 0.1 3.1.3 LDA [11,21] 1 3.2 3.2.1 25 ms MFCC 12 ΔMFCC 12 Δ 1 10ms 15% 16kHz 0.97 MFCC 15 22 Δ 50ms 3.2.2 k =64 k-means RWC 100 LDA (3.1.3) 3.2.3 3.1.3LDA 3.3 [22, 23] 3.3.1 6 Fluctuation Pattern (FP) [22, 23] 1200 3 FP 2 RWC [20] 100 95% 79 FP 23.2 ms FFT 11.6 ms Bark 20 6 FFT 0 10Hz 60 1200 = 20 60 [22,23] 11.025kHz MATLAB MA (Music Analysis) toolbox [23] 3.3.2 k =64 k-means RWC 100 LDA (3.1.3) 3.3.3 3.1.3LDA cfl 2014 Information Processing Society of Japan 3
2 Fluctuation patterns Fluctuation Pattern (FP) FP WSOLA FP 3.4 [12] 3.4.1 9 major, major 6th, major 7th, dominant 7th, minor, minor 7th, half-diminished, diminished, augmented major 5 /2, /3, /5, /b7, /7 14 (= 9 + 5) 12 168 (= 14 12) [12] HMM 3 major, natural minor, harmonic minor HMM [24] HMM HMM Viterbi 3.4.2 C 8 major, major 6th, major 7th, dominant 7th, minor, minor 7th, diminished, augmented 12 97 (= 8 12 + 1) 3.4.3 VPYLM tri-gram n =3 VPYLM tri-gram 1.0 10 5 VPYLM 10 5 4. *3 2000 2008 20 3278 A B 2 A 20 1 463 B RWC [20] 4.1 A: 3 6 10% 46 *3 http://www.oricon.co.jp/ cfl 2014 Information Processing Society of Japan 4
1 A 20 A 33 B B z 28 C 28 D 27 E 25 F BoA 24 G EXILE 24 H L Arc en Ciel 24 I 24 J w-inds. 23 K SOPHIA 22 L 22 M CHEMISTRY 21 N Gackt 21 O GARNET CROW 20 P TOKIO 20 Q 20 R 20 S Every Little Thing 19 T GLAY 19 11 9 463 5 6 10% 10% 3 10% 4 10% 5 6 4.2 B: 7 RWC 100 8 9 10 2 No.60, 70, 20 11 RWC 100 3 5 3 cfl 2014 Information Processing Society of Japan 5
7 0.02 0.02 0.02 3728 No.45-3.82 No.20 No.42-4.66 8-3.98-4.33 9 No.60 No.70 RWC 100 3728 No.15 No.55 No.90 No.73 No.99 RWC 100 3728 C FGCAm F G C 5. -3.86 No.6 No.8 No.29 No.60 No.81-7.51 10 RWC 100 3728-1 -5 No.56 No.41 No.54 No.82 No.84 11 RWC 100 3728 Songle [12] JST CREST OngaCREST RWC cfl 2014 Information Processing Society of Japan 6
2 B 5 () No. (1) 60 (2) 70 (3) 45 (4) 20 (5) 42 (1) 15 (2) 90 (3) 99 (4) 55 (5) 73 (1) 6 (2) 81 (3) 29 (4) 8 (2 ) (5) 60 M&Y (2 ) (1) 56 (2) 82 (3) 41 (4) 84 (5) 54 3 5 B... No. 56... F:maj C:maj G:maj F:maj C:maj G:maj... 82... G:maj C:maj F:maj G:maj C:maj F:maj...... E:maj A:min F:maj G:maj C:maj F:maj... 41 F:maj C:maj F:maj C:maj F:maj... 84... G:maj C:maj F:maj G:maj C:maj F:maj... 54 G:maj F:maj G:maj F:maj G:maj... [1] Vol. 60, No. 11, pp. 675 681 (2004). [2] Pardo, B.(ed.): Special issue: Music information retrieval, Communications of the ACM, Vol. 49, No. 8, pp. 28 58 (2006). [3] Casey, M., Veltkamp, R., Goto, M., Leman, M., Rhodes, C. and Slaney, M.: Content-Based Music Information Retrieval: Current Directions and Future Challenges, Proceedings of the IEEE, Vol. 96, No. 4, pp. 668 696 (2008). [4] Downie, J. S.: The music information retrieval evaluation exchange (2005 2007): A window into music information retrieval research, Acoust.Sci.&Tech., Vol. 29, pp. 247 255 (2008). [5] Downie, J. S., Byrd, D. and Crawford, T.: Ten Years of ISMIR: Reflections on Challenges and Opportunities, Proc. ISMIR 2009 (2009). [6] pp. 751 755 (2009). [7] Song, Y., Dixon, S. and Pearce, M.: Survey of Music Recommendation Systems and Future Perspectives, Proc. CMMR 2012, pp. 395 410 (2012). [8] Knees, P. and Schedl, M.: A Survey of Music Similarity and Recommendation from Music Context Data, ACM Trans. on Multimedia Computing, Communications and Applications, Vol. 10, No. 1, pp. 1 21 (2013). [9] Hamasaki, M., Goto, M. and Nakano, T.: Songrium: A Music Browsing Assistance Service with Interactive Visualization and Exploration of a Web of Music, Proc. WWW 2014 (2014). [10] 2013-MUS-99, No. 33, pp. 1 9 (2013). [11] Blei, D. M., Ng, A. Y. and Jordan, M. I.: Latent Dirichlet Allocation, Journal of Machine Learning Research, Vol. 3, pp. 993 1022 (2003). [12] Mauch, M. Songle: Vol. 54, pp. 1363 1372 (2013). [13] Pitman-Yor n-gram Vol. 48, pp. 4023 4032 (2007). [14] 2011-MUS-91, pp. 1 10 (2013). [15] Yoshii, K. and Goto, M.: A Vocabulary-Free Infinity- Gram Model for Nonparametric Bayesian Chord Progression Analysis, Proc. ISMIR 2011, pp. 645 650 (2014). [16] Fujihara, H., Goto, M., Kitahara, T. and Okuno, H. G.: A Modeling of Singing Voice Robust to Accompaniment Sounds and Its Application to Singer Identification and Vocal-Timbre-SimilarityBased Music Information Retrieval, IEEE Trans. on ASLP, Vol. 18, No. 3, pp. 638 648 (2010). [17] 2013-MUS-100, pp. 1 7 (2013). [18] Nakano, T., Yoshii, K. and Goto, M.: Vocal Timbre Analysis Using Latent Dirichlet Allocation and Cross- Gender Vocal Timbre Similarity, Proc. ICASSP 2014 (2014). [19] Goto, M.: A Real-time Music Scene Description System: Predominant-F0 Estimation for Detecting Melody and Bass Lines in Real-world Audio Signals, Speech Communication, Vol. 43, No. 4, pp. 311 329 (2004). [20] RWC : Vol. 45, No. 3, pp. 728 738 (2004). [21] Griffiths, T. L. and Steyvers, M.: Finding scientific topics, Proc. of the National Academy of Sciences of the United States of America, Vol. 1, pp. 5228 5235 (2004). [22] Pampalk, E., Rauber, A. and Merkl, D.: Contentbased Organization and Visualization of Music Archives, Proc. ACMMM 02, pp. 570 579 (2002). [23] Pampalk, E.: Computational Models of Music Similarity and Their Application to Music Information Retrieval, Ph.D. Dissertation, Vienna Inst. of Tech. (2006). [24] Mauch, M. and Dixon, S.: Simultaneous Estimation of Chords and Musical Context from Audio, IEEE Trans. on ASLP, Vol. 18, pp. 1280 1289 (2010). cfl 2014 Information Processing Society of Japan 7