An implementation of automatic transcription of monophonic music with a Blackboard system
Juan Pablo Bello, Giuliano Monti, Mark B. Sandler · 2000
A system is proposed for the automatic transcription of monophonic music. It consists of two main parts: the front end, using a segmentation system that detects the note onsets and offsets present in a musical track and performs the time-frequency analysis of the signal; and a blackboard system, that processes this data and obtains the pitch of each note working through an hypotheses hierarchy. Emphasis is given on the explanation of the latter, and satisfactory results are shown within a limited database of real-world sound samples.