Title :
Recognition of musical tonality from sound input
Author :
Izmirli, Özgur ; Bilgen, Semih
Author_Institution :
Dept. of Electr. & Electron. Eng., Middle East Tech. Univ., Ankara, Turkey
Abstract :
Work on the problem of computer based determination of the musical mode and tonality of a melody from sound input is presented. The study proposes a tool for determination of a time dependent tonal context vector for melodies from acoustical input. Each element of this vector, called a tonal component, indicates the extent of a specific scale usage. The system is made up of two stages: The first is involved in digital signal processing and is used to convert sound input of melodies into a sequence of musical note intervals and their associated occurrence times. The second stage uses finite state automats to match patterns of the interval sequences with pre-programmed scale patterns. The output consists of a time dependent tonal context vector whose elements are relative strengths of various tonal components regarding the melody under consideration
Keywords :
acoustic signal processing; electronic music; finite automata; music; acoustical input; computer based recognition; digital signal processing; finite state automats; interval sequences; melodies; musical note intervals; musical tonality recognition; pre-programmed scale patterns; time dependent tonal context vector; tonal component; tonal components; Acoustic signal processing; Automata; Context modeling; Digital filters; Fast Fourier transforms; Filtering; History; Pattern matching; Real time systems; Sampling methods;
Conference_Titel :
Electrotechnical Conference, 1994. Proceedings., 7th Mediterranean
Conference_Location :
Antalya
Print_ISBN :
0-7803-1772-6
DOI :
10.1109/MELCON.1994.381110