ANALYSIS-RESYNTHESIS OF MUSICAL TONES USING SPECTRAL MODELLING TECHNIQUE
|
|||
Note: This paper was presented at a conference, citation information below. | |||
Ong, B.S. & Ang, M. (1999). “Analysis-Resynthesis of Musical Tones using Spectral Modelling Techniques.” Malaysian Science and Technology Congress, MSTC’99, Symposium C. Confederation of Science and Technology Associations in Malaysia (COSTAM), Johor Baru, 6-8 December 1999. Proc.Vol.I. pp. 32-39. | |||
ABSTRACT Analysis-resynthesis is a modelling technique for the generation of musical tones. It looks at the spectrum of a real musical instrument and tries to recreate it through the application of additive and subtractive resynthesis algorithms. It encompasses different techniques that have a three-step process in common: firstly, a recorded sound is analysed, next, the musician modifies the analysis data, and finally, the modified data is used in resynthesising the altered sound. The sound of a musical instrument is contributed primarily by the addition of the harmonic partials that occur in the sound waveform. The analysis of a recorded sound is performed through the estimation of its power spectrum. The input waveform is windowed, with window widths based on the amplitude envelope parameters, obtained through analysis of the amplitude of the waveform with respect to time. The Fast Fourier Transform [FFT] is then applied to these discrete sections to obtain a spectral model of the sound signal. This data is then used for the resynthesis of the original sound. The modelling technique is tested using digitally generated waveforms with known spectral components and amplitude envelopes. Results indicate that the method developed is able to return the results expected, and may be used for the analysis and resynthesis of real musical instrument sounds. The complete paper is here. You will need the freely available Adobe Acrobat Reader to view it. |