敲弦樂器如鋼琴通常會有數根琴弦綁在同一個琴橋上的情形，而由於強烈的耦合現象(coupling phenomenon)，使得產生的樂音具有高度複雜的振幅調變情形。所以，如何決定一適當的合成模型及其參數以合成近似原音的合成音一直是個困難的問題。
本論文基於三個前人的工作：耦合弦模型(coupled string model)、交換式鋼琴合成法(commuted piano synthesis method)、無限脈衝響應模型合成法(IIR synthesis method)，提出一個多通道遞迴式類神經網路架構。我們期望在不具備樂器物理特性的知識的情況下，電腦能藉著此一類神經網路訓練演算法自動地分析樂音以擷取適當的合成參數。

Struck string instruments such as pianos usually have groups of strings terminated at some common bridges, respectively. Because of the strong coupling phenomenon, the produced tones exhibit highly complex amplitude modulation patterns. Therefore, it is difficult to adjust synthesis model parameters according to the recorded instruments such that the synthesized tones can match the measurements.
In this paper, a multi-channel synthesis model is proposed based on three previous works, the coupled string model, the commuted piano synthesis method and the IIR synthesis method. This work attempts to automatically extract the synthesis model parameters by using a neural-network training algorithm without knowing any physical properties of the instruments.

[1] Alvin W. Y. Su and S. F. Liang, “Dynamics Modeling of Musical Strings by Linear Scattering Recurrent Networks”, International Computer Symposium 1996.
[2] Alvin W. Y. Su and S. F. Liang, “Synthesis of Plucked-String Tones by Physical Modeling with Recurrent Neural Networks”, in Proceedings of the IEEE 1997 Workshop on Multimedia Signal Processing, pp.71-76. Princeton, NJ, 1997.
[3] Alvin W. Y. Su and S. F. Liang, “A Generalized Model-Based Analysis/Synthesis Method for Plucked-String Instruments by Using Recurrent Neural Networks”, the 106th AES Convention and Conference, Invited Paper, 5/1999.
[4] Alvin W. Y. Su, S. F. Liang and C. T. Lin, “Model-Based Synthesis of Plucked String Instruments by Using a Class of Scattering Recurrent Networks”, IEEE Trans. on Neural Networks, vol.11, no.1, pp.1-16, Jan. 2000.
[5] Alvin W. Y. Su and S. F. Liang, “A New Automatic IIR Analysis/Synthesis Technique For Plucked-String Instruments”, IEEE Trans. On Speech and Audio, vol.9, no.7, p.747, 10/2001.
[6] Eric D. Scheirer, “THE MPEG-4 STRUCTURED AUDIO STANDARD”, E15-401D MIT Media Laboratory Cambridge MA 02139.
[7] Erik Holsinger, “How Music and Computers Work”, Ziff-Davis Press, Calif., 1994.
[8] F. J. Pineda, “Recurrent Backpropagation and the Dynamical Approach to Adaptive Neural Computation,” Neural Computation, vol. 1, pp. 161-172, 1989.
[9] G. Weinreich, “Coupled Piano String ”, Journal of the Acoustical Society of America, vol. 62, no. 6, pp. 1474-1484, Dec. 1977.
[10] H. Szu, “Fast simulated annealing,” in J. S. Denker, ed., Neural Networks for Computing, pp. 420-425, American Institute of Physics, New York, 1986.
[11] ISO/IEC JTC 1/SC 29/WG 11 N2503-sec5, “Information technology – Coding of audio-visual objects, Part 3 : Audio, Section 5 : Structured Audio” ,1999-3-10.
[12] J. Chowning, "The Synthesis of Complex Audio Spectra by Means of Frequency Modulation," Journal of the Audio Engineering Society 21(7), 1973; reprinted in Computer Music Journal 1(2), 1977.
[13] J. O. Smith III, “Waveguide Filter Tutorial”, in Proceedings of the 1987 International Computer Music Conference, pp. 9-16, Champaign-Urbana. 1987.
[14] J. O. Smith III, “Physical Modeling using Digital Waveguides”, Computer Music Journal special issue on Physical Modeling of Music Instrument, Part I, vol. 16, no. 4, pp.74-91, Winter/1992.
[15] J. O. Smith III, “Efficient Synthesis of Stringed Musical Instruments”, in Proceedings of the 1993 International Computer Music Conference, pp. 64-71, Tokyo, Japan, September 1993.
[16] J. O. Smith III and S. A. Van Duyne, “Commuted Piano Synthesis”, in Proceedings of the 1995 International Computer Music Conference, pp. 319-326, Banff. 1995.
[17] J. O. Smith III, ”Physical Modeling Synthesis Update” Computer Music Journal, vol. 20, no. 2 , pp. 44-56, MIT Press, Summer 1996.
[18] “MPEG-4 Structured Audio Page of Machine Listening Group in MIT Media Laboratory“, http://sound.media.mit.edu/mpeg4/
[19] “MPEG-4 Structured Audio: Developer Tools Page of CS Division in UC Berkeley ”, http://www.cs.berkeley.edu/~lazzaro/sa/index.html
[20] M. Riedmiller and H. Braun, “A direct adaptive method for faster backpropagation learning: The RPROP algorithm,” in Proceedings of ICNN 93, pp. 586-591, San Francisco, CA, 1993.
[21] N. K. Treadgold and T. D. Gedeon, “Simulated Annealing and Weight Decay in Adaptive Learning: The SARPROP Algorithm”, IEEE Trans. on Neural Networks, vol. 9, no. 4, pp. 662-668, 1998.
[22] R. J. Williams, and Jing Peng, “An Efficient Gradient-Based Algorithm for On-Line Training of Recurrent Network Trajectories,” Neural Computation, vol. 2, pp. 490-501, 1990.
[23] Samuel Pellman, “An Introduction to the Creation of Electroacoustic Music”, Wadsworth Pub. Co., Calif., 1994.
[24] S. F. Liang, “Dynamics Modeling of Musical String by ANN”, Master thesis, Depart. Of Control Eng., Chiao-Tung University, Taiwan, 1996.
[25] S. F. Liang and Alvin W. Y. Su, “Dynamics Modeling of Musical String by linear Scattering Recurrent Network”, ICS’96 Proceedings on Artificial Intelligence, p263-270, Taiwan, 1996.
[26] S. F. Liang, Alvin W. Y. Su and C. T. Lin, “A New Recurrent-Network Based Music Synthesis Method for Chinese Plucked-String Instruments-Pipa and Qin”, International Joint Conference on Neural Networks, 7/1999.
[27] S. F. Liang and Alvin W. Y. Su, “Recurrent Neural Network Based Physical Model for the Chin and Other Plucked-String Instruments”, Journal of Audio Engineering Society, vol. 48, no. 11, pp. 1045-1059, 11/2000.
[28] T. Tolonen, V. Valimaki and M. Karjalainen, “Modeling of Tension Modulation Nonlinearity in Plucked String”, IEEE Trans. On Speech and Audio, vol.8, no.3, pp.300, May 2000.