Originally Posted by JamesN
For example, in the 3DBD of MvA, many people point to the "bridge" scene as exhibiting a good deal of ghosting. Again: high contrast (red cables against a light colored sky) and wide parallax (separation between left/right bridge cables). A good experiment might be to have the converter convert the 2D version of that same scene. Is the parallax between the bridge cables still wide in the converted version? If no, there will be less visible ghosting in the converted version.
I've noted this in a couple of bad demos I've seen with the Samsung LEDs. In each case, the displays had been set up improperly. In one store, the Samsung rep had set up the display with a regular Blu-ray player, not the 6900. The TV was doing the conversion. In the other case, the TV was the 46" 3D LCD and the Blu-ray was the 3D Model 6900, but the disc may have been the 2D version. I never found out for sure, because the salesman at this American Electronics store kept insisting that it was set up properly. It wasn't.
At any rate, the TV was doing 2D to 3D conversion. I noticed as soon as I removed the glasses that the offset of the two images was not nearly as great (or as consistent) as it was when the native 3D MvsA disc was being played. If you watch without glasses while the TV is doing the conversion, it's also quite obvious that the offset sometimes shifts significantly even within a single shot - in ways it won't while it's in true 3D mode. Indeed, without the glasses, you can understand a little better how the set is going about creating an artificial parallax view. In the bridge scene, for instance, the support cables sometimes seem to vibrate like piano strings. Again, you can see this happening if you watch 2D to 3D conversion without the glasses.
Another way to know for sure if the MvsA disc you're watching is showing in true 3D mode is to watch lines of text in some of the menus. Instead of a flat paragraph of words, 2D to 3D conversion turns those words into undulating rows, some of which fall off into the background.
That's why the effect is so disconcerting for me. I noticed while watching the new Star Trek movie on the Samsung 55" LCD set that foreground and background sometimes seemed to shift position within a single shot. That's crazy-a$$ 3D and the main reason the conversion is so ineffective. Conversion by the Samsung TVs doesn't simply create a mild but consistent 3D effect. Especially if it's a complex, fast moving scene (like the battle between the USS Kelvin and the Romulan ship in Star Trek) it creates a 3D effect on speed or acid. It's 3D gone horribly wrong, and I just want to turn it off.
IMO, the algorithms for 2D to 3D conversion aren't nearly sophisticated or fast enough for real time. They may get there one day, but I suspect that day is several years into the future. Never say never, though. I didn't watch the whole movie, but I actually found the 2D to 3D conversion of the filmed stage play of "Rent" fairly convincing. Of course, there aren't too many movies that are filmed stage plays. I suspect it worked fairly well because it had very little movement. Virtually none of the background objects moved.