When a group of still images are reproduced interactively, reproduction wait time to confirm the content of the images can be shortened. Image data representing the group of still images including the still images photographed from various viewpoints are formed into a hierarchy and recorded in an image file. A first layer has image data representing an index image comprising images viewed from 4 points whose resolution is ¼ of that of the original image. A second layer has hierarchical image data representing images viewed from 16 points and having a resolution which is ½ of the original image resolution. A third layer has hierarchical image data having the same resolution as the original image and viewed from 36 points. Upon reproduction, the hierarchical image data are reproduced from the first layer.