複音音樂的分析常常被廣泛的利用在音樂的特徵擷取上，而我們也可觀察到當音樂演奏時，每個被演奏的音通常都會隨著時間不斷變化，變化的資訊有音高、音色和音量大小等等。
在這篇論文中，我們針對擷取這些會隨著時間而改變的特徵來分析研究。利用非負矩陣分解的演算法，在沒有事先求得的音色模型的前提，以及未知所需要用來組成音樂訊號的基底數量時，我們提出一個動態建立這些基底的方法，並且可以自動的分析出需要的基底的個數。這樣一來，我們可以用最適當數量基底來表示音樂訊號的資訊。除此之外，為了更進一步的找尋每個局部時間點，不同的音高的變化，我們提出一個新的演算法，並與原本的非負矩陣分解法做比較。
利用我們所得到隨著時間不斷變化的這些資訊，可以在例如訊源分離、聲音的修正等應用上，提供很好的資訊。

Analysis of polyphonic music is often investigated for feature extraction. And the features of each note in the music are always time-varying, such as the pitches, timbre and volume.
In this thesis, we focus on the extraction of time-varying features. Based on the NMF algorithm, without a priori tone models and the necessary number of tone templates, we proposed a method to determine the number of templates and construct the templates adaptively and automatically. Thus, we can use the most appropriate templates to express the information of signals. Furthermore, in order to have the analyzed result more localized, a new algorithm is proposed. The new algorithm is favorable compared to traditional NMF methods.
With time-varying and localized information of music recordings, other applications such as source separation and sound modification are possible.

[1] P. Smaragdis, and J. Brown, “Non-negative matrix factorization for polyphonic music transcription,” in proc. IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, New Paltz, NY, pp. 177-180, 2003.
[2] D. Guillamet, and J. Vitria, “Non-negative matrix factorization for face recognition,” Topics in Artificial Intelligence, pp. 336-344, 2002.
[3] A. Cichocki, R. Zdunek, and S. Amari, "New algorithms for non-negative matrix factorization in applications to blind source separation." 2004.
[4] A. Cont, “Realtime audio to score alignment for polyphonic music instruments, using sparse non-negative constraints and hierarchical HMMs,” in proc. of International Conference on Acoustics, Speech and Signal Processing, Toulouse, 2006.
[5] F. Sha, and L. Saul, “Real-time pitch determination of one or more voices by nonnegative matrix factorization,” Advances in Neural Information Processing Systems, vol. 17, pp. 1233-1240, 2005.
[6] P. O. Hoyer, “Non-negative matrix factorization with sparseness constraints,” The Journal of Machine Learning Research, vol. 5, pp. 1457-1469, 2004.
[7] E. Vincent, N. Berlin, and R. Badeau, “Harmonic and inharmonic nonnegative matrix factorization for polyphonic pitch transcription,” in proc. of International Conference on Acoustics, Speech and Signal Processing, Las Vegas, Nevada, USA, pp. 109-112, 2008.
[8] T. Virtanen, “Monaural sound source separation by nonnegative matrix factorization with temporal continuity and sparseness criteria,” Audio, Speech, and Language Processing, IEEE Transactions on, vol. 15, no. 3, pp. 1066-1074, 2007.
[9] M. Nakano, J. Le Roux, H. Kameoka et al., “Nonnegative Matrix Factorization with Markov-Chained Bases for Modeling Time-Varying Patterns in Music Spectrograms,” Latent Variable Analysis and Signal Separation, pp. 149-156, 2010.
[10] C. Fevotte, N. Bertin, and J. L. Durrieu, “Nonnegative matrix factorization with the itakura-saito divergence: With application to music analysis,” Neural Computation, vol. 21, no. 3, pp. 793-830, 2009.
[11] Y. L. Chen, T. M. Wang, W. H. Liao et al., “Analysis and Trans-synthesis of Acoustic Bowed-String Instrument Recordings － A Case Study Using Bach Cello Suites,” in Proc. of the 14th Int. Conference on Digital Audio Effects (DAFx-11), Paris, France, 2011.
[12] R. Hennequin, R. Badeau, and B. David, “Time-Dependent Parametric and Harmonic Templates in Non-Negative Matrix Factorization,” in Proc. of the 13th Int. Conference on Digital Audio Effects (DAFx-10), Graz, Austria, September 6-10, 2010.
[13] D. Lee, and H. Seung, “Learning the parts of objects by non-negative matrix factorization,” Nature, vol. 401, no. 6755, pp. 788-791, 1999.
[14] D. Lee, and H. Seung, “Algorithms for Non-negative Matrix Factorization,” Advances in Neural Information Processing Systems (NIPS), vol. 13, pp. 556-562, 2001.
[15] N. Bertin, C. Fevotte, and R. Badeau, “A tempering approach for Itakura-Saito non-negative matrix factorization. With application to music transcription,” in In Proc. IEEE Intl. Conf. Acoust. Speech Signal Processing (ICASSP' 09), Washington, DC, USA, pp. 1545-1548, 2009.
[16] P. Smaragdis, “Non-negative matrix factor deconvolution; extraction of multiple sound sources from monophonic inputs,” Independent Component Analysis and Blind Signal Separation, pp. 494-499, 2004.
[17] W. S. Su, “Pitch and Partial Tracking of Polyphonic Musical Signals,” Tainan: National Cheng Kung University, Department of Computer Science and Information Engineering, 2009.
[18] W. Y. Su, “Restoration of Dynamic Image Sequences Resampling and High Resolution Reconstruction,” Electrical Engineering, New York: POLYTECHNIC UNIVERSITY, 1994.
[19] P. Fournier, “Bach: 6 Suiten für Violoncello solo,” Archiv Records, CD 1, Track 1, 1997. Barcode: 0028944971125
[20] S. Kuijken, “Bach: Sonatas & Partitas, BWV 1001-1006,” Deutsche Harmonia Mundi Records, CD 1, 1990. Barcode: 0035627704321
[21] V. Emiya, “Transcription automatique de la musique de piano,” Ph. D. dissertation, Institut TELECOM; TELECOM ParisTech, Paris, France, 2008.