Searching for more clues, I came across a number of forum entries on this topic dating back to the E3. Here is one example. The most intriguing Nintendo patent mentioned in those posts is the patent that brought ´cube mapping´ into view, here. It is entitled ´Video game play using panoramically-composited depth-mapped cube mapping´ and the abstract reads as follows:
Video game play rendered using a panoramic view of a cube map style rendering uses an associated depth map to supply three-dimensionality to the pre-rendered scene. The resulting panoramic rendering may be indistinguishable from rendering the original scene in real-time except that the background is of pre-rendered quality.I have to admit that I have never read this patent before, at least not in great detail. Often they can be very hard to get your head round. But this one appears amazingly straightforward: pre-rendered images in real-time 3D. That is a contradiction in terms to me, but that is precisely the claim. Reading on, there are some even more dazzling comments in the background description:
We have discovered a unique way to solve this problem in the context of real-time interactive video game play. Just as Alice was able to travel into a 3D world behind her mirror in the story "Alice Through the Looking Glass", we have developed a video game play technique that allows rich pre-rendered images to create 3D worlds with depth.There is no mistaking about that claim. And need I remind readers of the patent dealing with hidden objects and hiding objects I discussed in great length below? These two seem to fit together quite well, not to mention the apparent displacement mapping patent.
But how may this work? The more technical description is in the claims and is more of a mouthful:
A video game playing method comprising: loading a pre-determined environment-mapped image having multiple planar projected images and associated multiple depth map images; at least in part in response to real-time interactive user input, compositing at least one additional object into said mapped images, said compositing using said depth map to selectively render at least portions of said object into said mapped image to provide a composited mapped image; and panoramically rendering said composited mapped image using a desired viewing angle and frustum to provide interactive video game play.As far as I can make out, the technique would be to composite a number of pre-rendered sequences in such a way that would make them appear to respond to a player´s real-time input.
In the Gamedev.net glossary, Cube Mapping is described as follows:
An alternative to sphere mapping used in environment mapping, cube mapping gets a 'screenshot' looking in 6 different directions and arranges them in a rolled out cube. When applied, the object appears to reflect the environment around it. A 'cheap' alternative to raytracing reflections, cube mapping is fast enough for realtime.This seems to tie in with the quotes from the patent above and hints at how the apparent contradiction of pre-rendered graphics in real-time 3D may be overcome.
But why else might this be true? Firstly, naming the CPU and graphics chip ´Broadway´ and ´Hollywood´ respectively could be a clue. As has been pointed many times, Broadway is synonimous with live entertainment and Hollywood with pre-produced content. Naming their chips likewise suggests a fusion of the two.
Secondly, a lack in hardware power compared to the competitors has been reported from various reliable sources, yet Nintendo claims the graphics will be on par.
Thirdly - and this has not yet been mentioned in relation to this theory I believe - Nintendo has reportedly formed a division that is concerned with pre-rendered graphics. This was mentioned in conjunction with possibly forthcoming Zelda and Metroid movies. A year ago, Variety wrote:
Nintendo consultant and former prexy Hiroshi Yamauchi has put together a plan for the company to get into film production. He will soon present the proposal to the executive committee for consideration.What if this new division was, in fact, related to games? I would think it unlike Nintendo to venture into a completely new field altogether, moreover such a competitive one. Of course, Nintendo has enticing franchises. But - presuming that we are talking about CGI, since the division was reported as an in-house studio - Pixar, Dreamworks Animation and Blue Sky (Fox) really have got the market divided up among them. There is hardly room for another competitor, I would assume.
According to reports in the Japanese press, plan calls for Nintendo to create a pic based on one of its own franchises for theatrical release in 2006.
So, do I think this is really it? Let me put it like this: The cube mapping theory is old. So is this bit of news about a Nintendo movie division. But combine them and you get one of the best theories as to what the Revolution´s secret might be.
Source: PCVS Forum, Gamedev.net
Thanks to: GCFan2k5