For each of respective viewpoints, the capturing condition at the viewpoint, and each frame image captured from the viewpoint in accordance with the capturing condition are acquired. One of the viewpoints is selected as a reference viewpoint by using the acquired capturing conditions. Predicting coding and intra-coding are performed for the acquired images. For each of the respective viewpoints for each frame, the coding result of the frame image captured from the viewpoint by predicting coding or intra-coding is output. When performing predicting coding for an image captured from the reference viewpoint, predicting coding is performed by referring to the image captured from the reference viewpoint without referring to images captured from the viewpoints other than the reference viewpoint.