Acoustic time histories of vibrating partial sources

Sammanfattning: The research in this thesis evaluates a method designed for separation of individual or partial sounds from multiple sounds generated by a sound source (e.g. a diesel engine). The sound of sound sources can be recorded to reproduce a sound event in an environment, just like a photo or moving picture can be used to illustrate a visual impression. By modifying the recordings, desired sounds can be described as targets of an improved sound environment. Furthermore, the reproduced sounds of the recordings can form the basis for improvements on sources, sound paths and sound environments which affect the cause and the perception of the sound. This is the founding idea of sound design and work that strives to bring about improved product sound quality. In cases when the sound comes from a multipart product, questions may arise about which parts produce which sound. With knowledge about which sound is caused by which part, modification of the right part and hence the sound is achievable. After the different sound signals are separated from each other, the components of a multipart sound source can be identified and then appropriately tested. For example, the sound of the separate parts of a heavy vehicle engine may be assessed in listening positions around the vehicle or in its cabin. Time signals of the parts of the engine and the sound paths from the engine to the listening positions are calculated and measured separately. The sound of similar engines can therefore be determined in different cabins regardless as to whether an engine actually had been positioned there. The foundation for this research originates from signal processing and inverse problems, in the aspect of calculated time histories from sources, and methods for empirical measurements of transfer functions between sources and microphones. The method used consists of six steps: Recording of sound pressure signals, measurement of transfer functions, measurement of source strength, calculation of source strength matrix, calculation of filters and filtering of recorded sound pressure signals. The first four steps of the method can be used for automated sound source quantification of the sources' source strength or sound power; the subsequent steps are needed to determine the acoustic time histories of the sources as source strength at the source or as sound pressure in the listening positions. The results of the research demonstrated that this approach is a viable method for the detection and isolation of partial sound sources. As the research used two types sound source it is also apparent that additional research is needed before the model can be widely applied to other types of sources. In particular, the effect of the positioning of fictitious monopoles needs to be better understood and also the effect of distance on the positioning of microphones from a partial source.