(Content-Based Retrieval and Analysis of Harmony and other Music Structures)
In every culture, music is a very important part of human communication. To understand better the musical language used by the great composers, musicologists have been analyzing written music for already centuries. During the last past decades some analyzing tools have been formalized in a generative and almost deterministic way. Thus, in some case the tedious hand work is not far from the point where it could be described as a computational problem and, therefore, to use suitable computer programs to do the work. Moreover, some work on musical psychology about what makes a musical work pleasing to listen to or "easy on the ear" have been described precisely enough to be applicable in computerized music analysis.
Content-Based Music (Information) Retrieval, CBMR (at times referred as MIR), is a research topic studied rather extensively during the last half a decade. One of its famous instances is the so-called "query by humming" or WYHIWYG (What You Hum Is What You Get) application. Given a large database of music, the task is to find those pieces of music that contain excerpts mostly resembling (in a musical way) the hummed query.
The C-BRAHMS project aims at designing and developing efficient methods for computational problems arising from music comparison, retrieval, and analysis. Particularly, the project concentrates on retrieving polyphonic music in large scale music databases containing symbolically encoded music. The project uses the findings in musicology and music psychology to achieve musically meaningful methods and results. Moreover, all the project output are planned to be exhibited in a freely available (under the GNU General Public License) query engine.