Show description

Computational approaches have been proposed in the literature for the automatic emulation of listeners’ behavior [128]. An alternative approach to segmentation is based on the knowledge of music theory, in particular 46 Music Processing for classical music. According to music theorists, music is based on the combination of musical structures [72], which can be inferred by applying a number of rules. A MIR system exploiting this latter approach is presented in [89], where the segmentation algorithm is the one proposed in [10], where differences in melodic intervals and relative note durations are used as clues to automatically detect boundaries between segments.

In this case, the queryby-example (QBE) paradigm can be useful to help casual users in describing their information needs in a intuitive way. Normally, casual users are interested in the acoustic form, because their goal is listening to or collecting music. Furthermore, they may not be able to interpret the symbols in the symbolic score. Like for most of the media, casual users represent the majority of potential users of a MIR system. Some typical information needs are reported in the following, as a representative sampling of the needs that have been presented in research work in MIR addressing casual users.

1. , all the ensemble play simultaneously, performed by a symphonic orchestra. Compressed formats have been developed to overcome this limitation. Among the different compressed formats, the MPEG 1, Layer 3 (MP3) is the best known [94]. 4. Formats of Musical Documents 25 Fig. 2 Time domain representation of an audio excerpt corresponding to a performance of the score shown in Fig. 1. tional parameter of compressed formats is the bitrate, which is the number of bits needed to encode a second of music.