An automatic transcription system with octave detection

Octave detection has been an open issue in automatic transcription of polyphonic music over the years. In the literature, pitch detectors for polyphonic music usually fail to detect octaves for lack of information about the timbre of each instrument that appears in the music. To reveal the possibility that this difficulty in automatic music transcription can be overcome, here a music transcription system that is capable of octave detection is proposed. Preceded by a constant-Q time-frequency analysis front end implemented with a fast algorithm for nonorthonormal discrete wavelet transform [1], a classifier using support vector machine technique [2] serves as a novel octave detector in this system.