Badger,
I am afraid that there is no real straightforward answer here. Part of what you are looking for might be here:
http://www.cambridgeincolour.com/tutorials/cameras-vs-human-eye.htmBut there are several other things to consider. You wouldn't necessarily want to try to match the limits of human vision in a picture because with such a wide angle of view, you'd have lots of distortion. That doesn't look natural in most cases when projected onto a flat surface. We don't see distortion like that in our visual field because, at least phenomenologically, we are "in the scene". It wraps around us.
And it isn't as simple as just considering the angle of view of our eyes when looking at a fixed point straight ahead. When looking at a real scene, we don't take it in all in one glance or fixing our eyes in one direction. The eyes move in saccades, darting around the scene, taking lots of samples and assembling an impression in the mind based on the information integrated from lots of such samples. And the head moves too. So when we explore what might be near the corners in our projected picture, where there would be lots of distortion, we don't see any distortion in the real world. That area, when seen with eyes, is always in the center of our visual field and the camera (our eyes) is pointed straight at it. Also, our retinas are round, not flat like the film or sensor plane in a camera or the image plane in a rendering engine. So that also plays a role in the lack of distortion.
Everything we pay attention to is always in or very near the center of the visual field with our eyes pointed straight at it. When we stand facing a large print, our eyes move to center the area of interest on the fovea, but when we look at the corner of the picture, there is a mismatch between the projection plane of the image and the angle of the retina.
Really, the only way to solve this problem is to render a portion of a spherical projection of the scene and print it on a big concave surface such that your head can be in the center of the sphere and you can look around at it from there. But that is problematic for a number of obvious reasons.
I think that if you want the viewing experience to feel as natural as possible, maybe a better way to look at it is to simply decide first how large your print will be and what the best viewing distance is, and then select an angle-of-view for your TG camera that would match that. This way, the picture has a projection similar to what you'd see if you were looking through a window of that size. You need to do a little trigonometry:
2*atan((picture dimension/2)/viewing distance) = angle of view for one picture dimension
So a 30x18inch window seen from 30 inches would involve an angle of view of about 53x33 degrees.
If you want to do this calculation, make sure your calculator or software is set to degrees rather than radians. And 'atan' might look more like 'tan
-1'. With the Windows calculator, when in scientific mode, to get 'tan
-1', you need to hit 'inv' first. That'll turn the trig functions into their inverse versions.
As for aspect ratio, if I remember right, the typical cinema aspect ratios were selected with an idea similar to yours. They were trying to mimic the human visual field as much as possible. But there is just no simple answer there, because our vision falls off gradually toward the periphery. Where you draw the line is rather arbitrary.
And in your peripheral vision, notice that if you look straight ahead, you can't see as far above as you can below. I can see almost straight down and almost straight to the sides, but probably only about 45-50 degrees up from straight ahead. This makes sense from an evolutionary standpoint. Seeing the ground is more important to us than seeing the sky. So how would you deal with that if you are trying to mimic the visual field?
