As it is, pretty much every game already generates its own depth map: it's the z-buffer used in helping to render each frame. Reading out that data, or re-purposing it into a standard format for use by the Bravia screen, would have a negligible impact on CPU and GPU time. The end result on the display wouldn't be pure stereoscopic 3D with individually rendered frames for each eye, but it would certainly look pretty impressive, and not far off the real thing.
While Sony was probably wise to retract the statement about 3D gaming coming to all existing PS3 titles, the potential remains hugely exciting: all new releases could work with 3D out of the box if the developer/publisher chooses to support it. The biggest titles in the existing catalogue could be patched with 3D updates, and there would be no performance penalty whatsoever.
If this theory holds water, Sony would be able to take a commanding position in a new technology where Microsoft would be unable to compete for years. It would be leveraging its position as a leading light in both games and HDTV hardware to create an actual synergy that would position PS3 as the pre-eminent brand in 3D gaming, and if the effect is suitably impressive enough, early adopters would help drive a hell of a lot of sales of Bravia 3DTVs. As the Kotaku report linked earlier suggests, in the fullness of time, when it is ready for the mainstream, Sony could relocate the technology inside the Bravia set back into the PS3. An HDMI 1.4 controller would then send a conventional 3D signal out of this revised PS3 hardware that would work with all 3DTV sets - not just the Sony ones.
So, aside from the fact that it doesn't make TVs, what's to stop Microsoft following suit, perhaps by partnering with another major HDTV manufacturer? First up, there's a question of basic bandwidth. Xbox 360 runs with an HDMI 1.2 controller, with a 4.9Gbps link to the display. Good enough for 1080p60 with compressed surround sound, but that's the limit. On the other hand, all models of the PlayStation 3 feature HDMI 1.3, where bandwidth is effectively doubled to 10.2Gbps. While this is in place to support richer colour formats and lossless surround sound, there is nothing to stop Sony using the proprietary protocol Quaz51 suggests in order to beam over the depth map invaluable in creating a proper 3D effect. Sony already has "form" for proprietary HDMI links - there's the BraviaLink built into the PS3 Slim, for starters.
Of course, there have been dalliances with stereoscopic 3D already on both HD consoles, most notably Blitz Games' Invincible Tiger: The Legend of Han Tao, covered on Eurogamer a while back and released on both Xbox Live Arcade and PSN a few weeks ago. It supports true stereoscopic 3D, along with its infinitely crappier 1950s-based relative, the more familiar red/blue anaglyph 3D. Despite full HD claims from Blitz, it's clearly a native 720p game scaled up to 1080p, using 640x720 or 1280x360 resolutions depending on the mode you choose. It's an intriguing experiment, but simply can't be compared with the results Sony has shown on bleeding-edge PS3 games at both CES and CEDEC 2009.
It all suggests that the stars are moving into alignment for Sony and the PlayStation 3 as we move towards 2010. The company is set to launch its own wand-like motion controller in the spring, and it already has head-tracking technology built into the forthcoming Gran Turismo 5. Combine these elements with a stereoscopic 3D solution, and we have a battery of technological advancements for both the mainstream audience, as well as hardcore early adopters looking to upgrade their HDTVs to get the full experience. From a personal perspective, a fully working head-tracking solution integrated into proper stereoscopic 3D would be enough to make me upgrade my own display.
Whether the specifics of the theory presented here turn out to be the case or not, Sony's move into 3D gaming isn't vapourware. The event demos are clearly a fully working reality and should be greeted warmly as an example of how the platform holders are innovating in new areas over and above the amount of pixels pumped out from their hardware. Four or five years into a console's lifecycle, we'd typically be looking forward to the next generation of technology, but things are clearly moving in different directions this time. Both Microsoft and Sony are taking hugely innovative steps to ensure that we get maximum value and a range of new experiences from the kit we already own, and in many ways that's even more impressive (and certainly cheaper) than developing consoles.