Since advanced 3D video systems employ multiview generation techniques using depth data to provide more realistic and comfortable depth impression, many research works are related to depth generation exploiting a time-offlight (TOF) camera. Most of them use time-consuming depth enhancement methods since the TOF camera generates noisy and unstable depth maps. In this paper, a 3D scene capturing method using stereoscopic cameras and one TOF camera is presented. The proposed method generates two viewpoints' depth videos corresponding to those of color cameras using depth up-converting and 3D warping techniques instead of using complex and time-consuming depth enhancement processes. In addition, the proposed method is implemented with the multi-thread technique to develop a fast system. Experimental results show that the proposed method generates two viewpoints' depth videos at 15 fps and those are used for generating multi-view images.