Concerning the topic at hand, it has now been suggested by a number of people commenting on this story that Nintendo may in fact be planning to use yet another type of mapping called ´parallax mapping´. As ever, I have direct evidence in the form of new patents, but let´s look at the technique itself first.
Wikipedia is a good starting point again:
Parallax Mapping (also, Photonic Mapping, Offset Mapping or Virtual Displacement Mapping) is an alternative to the bump mapping and normal mapping techniques applied to textures in 3D rendering applications such as video games. To the end user, this means that textures (such as wooden floorboards) will have more apparent depth and realism with less of an influence on the speed of the game.To see exactly what parallax mapping does, let us look at some samples, courtesy of InfiScape.
Parallax mapping is done by displacing the texture coordinates such that the texture occludes itself in accordance with a height map. Next-generation 3D applications may employ parallax mapping as new graphics algorithms are developed.
An easy way to understand this concept is to close one eye, take a pencil, point it at your eye, and move your head left and right. Parallax mapping takes that pixel on the far left of the pencil when it was facing you directly and stretches it accordingly to simulate your angle in comparison to the pencil.
Parallax mapping is also a way of faking displacement mapping.
The effects are obvious, particularly in the second example where it is combined with bump mapping. Basically, it works like this: Imagine looking at a raised point on a surface and then moving sideways. Due to perspective, the raised point will appear to move in the opposite direction. Parallax mapping can modify bump mapping to take account of such a move. This is referred to as offsetting. No extra polygons are needed. A height map is enough to calculate the offset value. It´s very simple to implement, is easy on the CPU and can thus be combined with other mapping techniques to achieve results, easily comparable to displacement mapping (though it does not actually alter the height of the pixels, I should point out).
Now, there are plenty of quotes in the patents mentioned in my earlier posts below that seem to relate to this technique. But I also came across this new one entitled ´ Method and apparatus for interleaved processing of direct and indirect texture coordinates in a graphics system´, which seemingly referrs (in some part, at least) to parallax mapping, as it often mentions an offset matrix that is applied to the textures:
In a graphics system having a memory containing texture data, a method of indirect texture referencing comprising the steps of: (a) using a set of indirect texture-coordinates to retrieve offset data from the memory; (b) multiplying the offset data by predetermined values forming elements of a texture offset matrix to produce a set of texture offset coordinates; and (c) using said set of offset texture coordinates for referencing texture data stored in the memory when mapping a predetermined texture to a rendered polygon.Reading further, we learn about what effects this technique might have on a hardware arrangement (formatting is mine):
To provide an enhanced repertoire of texturing capabilities for a 3-D graphics system, the present invention provides a versatile texturing pipeline arrangement achieving a relatively low chip-footprint by utilizing a single texture address coordinate/data processing unit that interleaves the processing of logical direct and indirect texture coordinate data and provides a texture lookup data feedback path for "recirculating" retrieved indirect texture lookup data from a single texture retrieval unit back to the texture address coordinate/data processing unit. The interleaved coordinate processing and recirculated/feedback data arrangement of the present invention allow efficient processing of any number of logical direct and/or indirect texture mapping stages from a smaller number of hardware texture processing units while preserving a fine granularity in the overall data processing flow.Concluding, it seems to me that this technique is the most likely to be used in the Revolution. It almost has the visual quality of displacement mapping while requiring next to no more calculations.
Image source: InfiScape