A library of 'natural' movements must be made available, possibly by using the output of some commercial motion capture system.
Gesticulation and Talking Figure within a Scene (Case 3): The inclusion of a full 3D environment, with background scenes and additional 3D objects for the human figure to interact creates a rather complicated problem. A question over the amount of data that can be handled in a set-top box arises. Spline modelling and specifically the animation technology used here are built around the idea of low storage demands for figures and movements, but there remains the challenge of realizing a complicated scene.
Virtual World (Case 4): A full virtual world with multiple figures and user immersion, apart from the storage limitations posed by the set-top box already mentioned in case 3, raises the challenge of realizing a virtual world within the technology limitations of today. One of the challenges is to identify which parts of VRML ( VRMIL 1997) should be made compatible with the SAM graphics technology and how an integration of such structures could be performed.
Overlay display (Case 5): Compared to the situation in cases (1) or (2), an additional technical challenge will be how to realize overlaid windows on the set-top box. Different settings such as graphics-over-video, graphics-over- menus, graphics-over-graphics, rely on hardware technical features of the set- top box. For example when dealing with case 3, it would be useful to have graphics-over-graphics overlay when considering a moving figure and a mostly still background, as this would reduce the load on the graphical pipeline.
Other (Case 6): Among potential scenarios for future projects, the inclusion of speech input, and not just output, seems to be closer in practical realization and usefulness than others such as a talking head on the remote control. Given the amount of work other more realistic scenarios demand, a more detailed technical assessment, case 6 for the current projects is premature.