Originally Posted by tgm1024
Yes, but he makes a mistake; A mistake similar to many other mistakes that academics make by trusting calculations over observation and ending up with a bumblebee that cannot fly. That corner distortion he's talking about is correct only on paper. As humans, we've long since evolved to recognize straight vertical lines off to the right as straight vertical lines even if the top and bottom of them are further away.
I challenge anyone looking at a flat rectangle on the wall to see the corners as somehow curving outward similar to the dual (upright+inverted) keystone distortion he's talking about. You won't.
Yes, there is something "turned around" about the logic there and a bit self-refuting if you stare at it long enough. Sort of similar to some who I've seen argue that high frame rates, or frame interpolation looks fake because "it looks smoother than real life, in real life motion is blurred to our eyes." Uh...except since you aren't, "in real life," seeing the HFR as blurred, then these people should notice their proposal is self-refuting.
But back to the purported geometric compensation of the curved screen. Take a flat screen and a curved screen. I have a flat ruler the width of the TV screens. I hold the straight, flat ruler horizontally against the screen. Now I want to reproduce this, so we show an image of that ruler (same life size) on each screen.
Now which screen is going to reproduce the experience of seeing that flat horizontal line (i.e. an image of that same ruler, same size as the ruler)? Obviously: the flat screen. It will reproduce the ruler flat and unbent, just like the real ruler is flat. It's producing the same geometry to our eyes as the real ruler. To the degree our eyes, our visual system, introduces distortion to straight lines, the image of the flat ruler will be distorted just like the real flat ruler. Thus preserving accuracy of the real thing.
Whereas the curved screen will be ADDING a curved distortion that the real ruler does not have. Whatever "compensating" it is doing, it's not in the direction of accuracy to the image of the straight line/ruler.
As for curved screens in general, I've tried numerous times viewing them in the stores from various distances, usually trying to get an immersive sense of the image, then comparing to the flat screens, I found no advantage to the curved screens. I only end up noticing distortions, of the type mentioned above.