Investigation of some parameters of a neuro-fuzzy approach for dynamic sound fields visualization

3D Visualization of Sound Fields Perceived by an Acoustic Camera

“…The collected data are periodic signals with variable amplitude and constant frequency of the noise source (the beeper). The investigations for proper parameters of our algorithm [10,12] were carried out using a static beeper as a sound source, as well as two static beepers. Based on the results it was decided that 11 MT filters, 30 ENS neurons and 3 clusters suit well to demonstrate the performance of our algorithm for this particular sound source.…”

“…1. Since all the experiments were carried out with Brüel & Kjaer acoustic camera, the description of the raw data pre-processing is based on its specific characteristics (e.g., the microphones positions and their correspondence to the observed by the acoustic camera "screen") as described in our previous works [6,7,10,12]. The acoustic camera we have consists of 18 microphones array placed randomly in a wheel grid with mounted optical camera at its centre.…”

“…Thus we returned to the originally proposed in [8] idea to choose the projection(s) between neurons having the biggest number of maxima of the probability density distribution of their equilibrium states. Although even in this case there is a chance to obtain more than one projection, it appeared to be the best option among the other tested approaches [10,12].…”

“…The second separates the data into predetermined number of clusters and is faster. During preliminary experiments with test sound sources with constant frequency [6,7,10,12], it appeared that 3-4 clusters were enough to obtain a clear "sound picture". Hence, for creation of 3D visualization here we will present the results with only three clusters for simplicity.…”

“…Next in [7] the approach was extended to dynamic visualization of the perceived by the acoustic camera sound field, thus creating a three dimensional picture of the sound waves propagation through time. Next in [10,12] investigations of the proposed algorithm aimed at its refinement and tuning of its key parameters continued. Experiments with two sound sources, as well as with a moving sound source were carried out too.…”