This web site is no longer maintained and the content may be outdated.
Please visit for up-to-date information.
No upcoming events...

Home / Graduate / M.S. Theses Completed
  Ali Taylan Cemgil, 1995  [download thesis]    

Thesis Title

Automated Music Transcription


We investigate efficient digital signal analysis techniques based on the wavelet theory when applied for pitch detection of digitized monophonic musical signals for transcription purposes. The aim is to reduce the computational burden by multirate signal processing techniques and still meeting the heavy analysis resolution requirements. In this respect the properties of the wavelet transform are investigated. This approach is equalent to a time scale representation, a more natural representation for musical signals than e.g. short-time Fourier transform.

Pitch tracking methods are classified into two domains: time domain methods and frequency domain methods. Under time domain techniques, FIR and IIR-lattice subband filter bank solutions with different regularity and phase conditions are compared against conventional methods. All frequency domain methods are presented as parametric tilings of the time- frequency plane. Efficient numerical algorithms for the calculation of the respective transforms and corresponding simulation results are given.
Boğaziçi University Department of Computer Engineering
Address: 34342 Bebek, Istanbul, TURKEY
Phone: +90 212 359 4523-24 Fax: +90 212 287 2461
general information:   webmaster: