A network-based device creates composite video streams containing video images of users for display on video conferencing screens. Video image streams are cascaded through a series of quad boxes to create combined video streams and ultimately the composite video stream. The composite video stream produces a display grid comprised of a plurality of image boxes. The device uses quad box input port mapping to direct a specific video image stream to a specific image box by controlling the cascade path of each video image stream through the quad boxes in creating the composite video stream. A user interface permits a user to selectively determine the layout of user images on the video conferencing screen. Screen layout extends to size of each user image being incrementally controlled and location of each user image being selectively determined depending upon the size and the number of the various user images. By controlling size and location of each user image, the screen display area available for other uses, such as the display of applications programs, is increased.