Abstract:

This work presents an artificial intelligence solution for the harmonic classification problem. The proposed artificial intelligence model divides the harmonic classification problem in subproblems. Intelligent solutions are indicated for each subproblem. The subproblem's solutions interact into the model in the way to find the solution for the harmonic classification problem. The subproblems found are: chord identification, chord classification, chord inversion classification, music tonality classification and harmonic degree's classification. The model indicates connectionist solutions for the chord and tonality classification subproblems, and indicates symbolic solutions for the chord inversions and harmonic degree's classification subproblems. The chord identification problem is partially solved by an algorithm solution. The model was implemented in an appropriately software and hardware that allowed connectionist and symbolic solutions, and the utilization of MIDI interface as music source. The model validation was performed using musical parts from great erudite composers. The model performed an acceptable classification of these music parts showing that cognitive musical problems can be solved by Arfificial Intelligence solutions.