| Popis: |
The Room Impulse Response (RIR), which indicates the propagation characteristics of sound, is used in various acoustical applications. In particular, multi-point measurement of RIRs is important for applications that require spatial information, and measurement of these is difficult. Previously, spatial interpolation methods of RIRs based on physical models have been proposed. Especially, the sparse Equivalent Source Method (ESM) with Image Source Method (ISM) acheive the spatial interpolation of the early reflected sounds with higher accuracy at high frequency. However, this method has a problem that when the order of reflected sounds to be estimated increases, it becomes difficult to find the optimal solutions because the number of bases increases. In this study, to reduce the number of bases in each optimization problem, we proposed a sparse ESM with the image sources grouped by the arrival times of reflected sounds. The optimization problem for each group of image sources is individually solved. In the simulation experiment, it is revealed that the proposed method can estimate higher-order reflections with higher accuracy compared to the conventional method. In addition, we found that the estimated region with high accuracy became broader. |
