IntroductionStructure-borne sound is an important issue in the field of architectural acoustics. In order to predict such transmission characteristics of vibration, prediction methods using wavebased numerical analysis have been proposed [1][2][3]. Toyoda and Takahashi [1] proposed 3-dimensional elastic wave analysis for structure-borne sound by the finite-difference time-domain (FDTD) method. In this method, the target building structures are spatially discretized by 3-dimensional meshes in the x, y, and z directions, and accurate modeling of the vibration field is possible. However, the computational costs are greater than those of the simulation with 1-/ 2-dimensional models. In order to cope with such a problem, the development of an efficient calculation method, which has less memory requirement and is able to perform faster calculation, is necessary. For this reason, vibration analysis methods applying the 2-dimensional plate model have been proposed [2,3]. In such research, the computational costs of memory and calculation time are aimed to be reduced by decreasing the discrete meshes to 2 dimensions. Nicole and Kessissoglou proposed a method for predicting the power transmission through a simple L-shaped structure composed of two plates using a combination of the modal method and the traveling wave method, which results in an exact solution. With this scheme, the vibration characteristics can be solved. However, it is difficult to apply it to structures with a complex shape, owing to the feature of the exact solution. As a more flexible method for structuring the model, Takahashi et al. [3] proposed wave-based numerical analysis using the finiteelement method (FEM) for a building structure model composed of multiple plate elements. In contrast to such a frequency-domain method, methods in time-domain such as FDTD method and the CIP (constrained interpolation profile) method are widely used. By using such methods, the time development characteristics of sound/vibration propagation can be directly obtained from the calculation results, which greatly contributes to the understanding of the mechanism of the propagation process. For this reason, the authors propose a