In development of mechanical products development time and cost are critical factors. In trying to minimize these simulation driven development is a hot topic where the ultimate goal is to be able to design a product solely based on modelling. To do so all aspects of the product will have to be examined before production starts.

The purpose of this master's thesis is to examine if the sound field created by a vibrating structure, a plate, can be predicted by means of modelling. The problem is split in subproblems named vibration, sound generation and listening each contributing their part to the final auralization. A combination of finite elements and acoustical modelling is used to estimate the sound radiation.

The thesis compares data from measurements and analytical work with simulation results to determine if finite elements analysis results are accurate enough to be used for sound generation.

The final outcome of the algorithm is sound which does share some resemblance with the one from the physical object. Still it is concluded that the generated and recorded sounds are not identical, neither objectively nor subjectively.