In this paper, a novel approach to the transcription of polyphonic signals is investigated. The method makes use of top-down analysis and bottom-up reconstruction of the time-frequency contents of the signals. Time-frequency contents of the signals are obtained using a bank of narrow band filters with very sharp transition band. The filters are designed using the frequency response masking technique. The high selectivity of the filters has enhanced the top-down analysis and increased the accuracy of identification over the simple frequency transform techniques. The method was tested with musical notes generated from an acoustic piano. Results show that for single notes and chords up to four notes, perfect identification can be achieved.