I took my time to watch a few videos. The realtime demo of the Move controller in the Sony Entertainment panel was absolutely hillarious - hopefully Natal can perform as good as the Move, or Microsoft will be in trouble.
I would like to watch some more, but I just don't have spare 40+ hours to watch all of them
![Smile :)](./images/smilies/icon_e_smile.gif)
Maybe later. I generally like to just read through PowerPoint slides or maybe read the transcripts rather than watch the full videos - it just takes much less time. The slides usually contains all the key points, and very few presenters actually elaborate anything beyond what is presented in the slides; the live discussion format is sure different enough to warrant watching of full live footage though.
I find it interesting that most panelists talked about fundamental issues with the game design, in particular user interface (3D cursor, icons/tags at the depth of the object, transparent GUI, HUD text etc.) and camera setup (pop-out, camera angles, separation and convergence, transitions, etc), as well as game level design and limitations of the older shader hardware. I think many people here at MTBS3D forums tend to underestimate the amount of artistic changes needed for a proper stereo 3D conversion; hear it from the involved parties.
In practice, it means that full S3D support is viable only for projects that are just starting and will be completed in 2-3 years, since developers can take their time to design everyting with stereo 3D in mind right from the start. These games should be designed around D3D10, and should look and play definitely better than current D3D9-era games where stereo support was mostly implemented in a quick and dirty way (some panelists even talked about "native Nvidia 3D vision support" when in reality all they did is correct the most nasty visual issues with the user interface and allow the 3D Vision driver to automatically convert the world for them).
As for all the confusion about separation (3D depth) and convergence (pop-out) levels in the live footage, well I think the comments were entirely legitimate. If you are going to showcase your technology, you need to do it in the best possible way. The panelists should have taken into account the fact that they were going to show their clips on a much larger screen and in a much larger auditoruim than they typically do in their development labs, but it seems like they didn't bother to test the live footage on a big screen. No wonder that it backfired.
The solution is quite easy, both stereo 3D games and "automatic" drivers should just offer some basic settings such as diagonal screen size and distance to the screen, and should allow not only basic computer monitor sizes but also take into account large projection screens and large auditoriums. Hopefully the panelists learned this lesson and will take their time to properly setup the demos in the future.
I don't think this will be a problem at E3 since it's a live floor show and attendees use their own display setups which can be tuned for an optimal depth and separation well before the show.