Stereoscopic 3D contents editing workflow using depth measurement and compensation

Abstract

This paper presents an efficient method for creating depth perception for editing stereoscopic three-dimensional (S3D) contents. Unlike two-dimensional (2D) contents, S3D contents require a production of depth perception from the pre-production step. The storyboard and depth chart for each cut should first be established, based on which the shooting and editing work will be done. Most existing methods for editing S3D contents are completed at the expensive DI step of the final editing process, in which the stereoscopic imagery has been viewable. For this reason the depth perception of S3D contents is currently achievable only at the DI step and not at the non-linear editing (NLE) steps. The editing based on storytelling however creates primarily at the NLE (Non Linear Editing) process. This paper proposes a more effective method for confirming the depth perception of each cuts using a depth map, as an editor creates shots of sequences. The experimental results demonstrate that an editor can more efficiently identify the not-good (NG) sequences using the depth map using generated by the proposed method.