MX Simulator

I have a question
Is the 1st person camera based off 1 source of imaging ?
A human has 2.. (EYEBALLS)

Food for thought....

Binocular vision is more for depth perception than a wide field of view. The real weakness for conventional 3d with a wide field of view is that the central part of your view is shrunk down to a tiny area due to the way things are projected onto the view plane while the peripheral areas occupy almost the entire screen. Basically the opposite of your eye which has more resolution in the center and less at the edges. Some kind of hardware fisheye would be nice but I have no idea how hard it would be to do efficiently. Probably the easiest way would be to render the scene multiple times at different FOVs and then blend them together in the final shader.

What I've always wondered about is even if what is displayed shows everything your eyes could see in focus, you are still focusing on one thing at a time. For example, I normally use 110 FOV in first person, if I'm focusing on a corner, the sides of the track still aren't in focus. Maybe I'm understanding what you're saying wrong but to me it seems like it still works how it should. I suppose this argument depends on the size of the monitor and viewing distance from it also.

If your monitor is 20 inches across, the projection should be correct for a 110 FOV if you close one eye and put your other eye ~14 inches away from the monitor (20 inches / tan(110 degrees / 2)).

As far as making the resolution track your eye (if that's what you're suggesting), I don't think it's all that necessary since it's not a big deal to draw a reasonable scene at full resolution. The problem is if you want to do a one pass fisheye where the center of the screen looks like it's at 90 FOV while still showing a 170 FOV fisheye at the edges. The tangent of 170/2 is 11.43. The tangent of 90/2 is 1. So that means it I wanted to render the scene at 170 FOV and then warp it so that the center looks like it's at 90, I'd have to render at 11.43 times the resolution to have the center be at full resolution after the fisheye warp. That's why you'd have to do it in multiple passes for it to be feasible.

I just wanna take this time to say, JLV is one smart dude. It has always impressed me how 99.9% of the time when someone asks a technical question or gives a suggestion, you are on top of it. Its never "hmm I'll have to check into how to do that". You always have the science behind your answer.

Anyway, enough of my fan-girling lol. Carry on!

jlv wrote:If your monitor is 20 inches across, the projection should be correct for a 110 FOV if you close one eye and put your other eye ~14 inches away from the monitor (20 inches / tan(110 degrees / 2)).

As far as making the resolution track your eye (if that's what you're suggesting), I don't think it's all that necessary since it's not a big deal to draw a reasonable scene at full resolution. The problem is if you want to do a one pass fisheye where the center of the screen looks like it's at 90 FOV while still showing a 170 FOV fisheye at the edges. The tangent of 170/2 is 11.43. The tangent of 90/2 is 1. So that means it I wanted to render the scene at 170 FOV and then warp it so that the center looks like it's at 90, I'd have to render at 11.43 times the resolution to have the center be at full resolution after the fisheye warp. That's why you'd have to do it in multiple passes for it to be feasible.

I play on an 80" flat screen TV.... Any suggestions on what the correct FOV is? (I dont quite understand what you mean)

For a 110 FOV you want to be .7 times the width of your screen away from the screen. 80" is diagonal so I can't say without knowing the aspect ratio.