Identifikation und Klassifikation von Musikinstrumentenklängen in monophoner und polyphoner Musik (English shop)

Preview

Table of Contents, Datei (100 KB)
Extract, Datei (89 KB)

The rapid development of network and storage technologies has made large multimedia databases with extensive music and sound archives available to a broad range of users. In order to manage such multimedia databases using classical, text-based functions, the individual contents must be annotated with metadata describing their content. Especially in the open databases of the Internet, however, this is only rarely the case, so that many multimedia contents cannot be captured by current text-based search engines.

Algorithms for semantic data retrieval solve this problem, since they allow content-based data retrieval in multimedia data together with automatic metadata extraction. The identification and classification of musical instrument sounds in monophonic and polyphonic music constitutes a specialised field in this context, in which metadata about the instrumentation of pieces of music or the source instruments of sounds can be obtained.

For multimedia databases and Internet search engines, metadata allow the formulation of similarity queries and content-based search requests for pieces of music with respect to the instrumentation used. Furthermore, new possibilities arise in the field of recording studio technology, such as the sound-based management of sound databases or the automatic doubling of audio tracks with similar-sounding instruments.

This book examines the identification and classification of musical instrument sounds in monophonic and polyphonic music. To this end, all relevant musical phenomena as well as the physical properties of musical instruments are considered extensively with regard to the resulting sound characteristics. Furthermore, various technical methods for feature extraction and classification are explained and analysed, and different implementations are presented and evaluated by means of comprehensive tests.