Abstract: Gaze positions of an operator wearing a virtual-reality headset displaying a video stream define preferred view regions of the display. Starting with a reference gaze position, and for each subsequent distinctly different gaze position, the virtual-reality headset sends control data, including three spatial coordinates and a time coordinate expressed as a cyclical video-frame index, to a view adaptor receiving the video stream. The view adaptor stores contents of a number of most recent video frames of the video stream in a circular content-buffer and control data of a number of most recent gaze positions in a circular control-buffer. A content filter within the view adaptor determines a preferred view region surrounding a gaze position according to control data held in the circular control-buffer and extracts a partial content of a respective frame held in the circular content-buffer according to the preferred view region.