#Human Eye Perspective?
10 messages in this thread
I wonder if anyone has used 3DS in asituation that needed the camera view to
show what the human eye would perceive from a given distance from objects in
question. For example, say you are designing a Billboard and want to show the
client what people will see 100 feet away versus 50 feet away. Is a 35mm or
50mm setting on the camera accurate to standard human vision? (I realize I just
touched on a huge window for variation with that word "standard vision") Any
ideas?
Paul
Paul, I believe the default cam. setting of 48.whatnot is as close to human
perception as poss. There is a std. ht. above the grnd that arch. folks use to
help w/ the illusion of human perception: about 5' 8" to 5' 10" but I'm not
sure exactly. Bob R.? Greg P.?
Wade and Paul:
For typical architectural human eye vantage points it is common to use 5'6".
Admittedly this is loosely based on a 5'10" average dude not a dudette. As a
rule of thumb I figure the eye to be just about 4" lower than one's height.
– Bob Ritger
Knew it was in there somewhere. Thanks.
Hi Paul,
The typical human field of view is 48 degrees which is provided by 3D
Studio's default lens length of 48.24mm. The problem that you must deal with
is that the field of view of 48 degrees represents a focused view. It does
not account for peripheral vision nor does it account for the fact our eyes
are usually "scanning" from side to side. Our field of awareness, as opposed
to field of view, is quite large.
The "natural" field of view will vary depending upon the environment and the
subject of your view. Our field of awareness shrinks as we focus more
intently on an object and as the environment closes in around us. Landscape
photographers frequently use very wide angle lenses (<30mm: large subject,
unrestricted environment), architecural photography often uses moderate wide
angle lenses (35mm to 40mm: limited subject, surrounded by walls or
buildings), while portrait photography uses normal to narrow angle lenses
(50mm to 135mm: tight subject, environment closed in by darkness or
backdrops).
If you have a 35mm camera with a wide angle zoom lens you might consider this
experiment. Find a billboard in an environment similar to what you are
modeling. Casually look at the scene. Look through your camera and adjust the
lens until the scene in the viewfinder is similar to what you saw without the
camera. Check the lens length and reduce it by a few mm's (that's because you
invariably concentrated a little too much on the scene and subconciously
narrowed your field of view). My guess is that you'll want to work with a
lens length around 38-40mm.
I know this isn't the answer you were looking for but vision is subjective.
Steven
Paul,
Another fact you might want to consider is the that people read with the center
5 degrees of their vision. Another words you can't read with your peripheral
vision. Just try looking at the side of the monitor while reading this message.
See what I mean.
Regards………Rick – Technical Animations
Good point Rick. Thanks also to others who responded. I realize the problem
lies in that vision is subjective as Steve mentioned. Here's what I have
decided to do. Take a 35mm picture from the vantage point I am rendering,
taking careful attention to making the camera viewpoint resemble what I see
with my naked eye. I'll scan in that view and transition from that to the view
rendered in the animation. At least that way I have established a reference
point based on actual human vision, not quesswork.
Thanks all who replied.
Paul…Fresh Look Design
Paul,
If you're going to put forth the effort of making a transition from a
photograph to a rendered image, why not do a "can't see it" special effect.
Make your new billboard a seamless part of the scanned picture by matching the
perspective and compositing it onto the photo.
In this way, you will have a real world setting for the billboard.
— James — Choreo Motion —
James,
Good idea. I have never done that before but will give it a try. Have you had
much experience doing that? Any tips?
Paul
> Have you had much experience doing that? Any tips? <
I've done a few. Lots of things can improve the quality of the finished image,
here are a few off the top of my head.
1) Use a still image for simplicity -or- make sure the video camera is fixed
tightly so it does not drift. Compositing over video looks better because it
adds a degree of realism. Leaves fluttering and passing cars look more
realistic that a static capture. (Unless the finished product will be a print,
in that case go with film.)
2) Record the camera's postion and the target's postition (and orientation)
to the best of your ability. This will be very helpfull in making adjustments
for matching perspective. A transit, a 50' tape measure, and an assistant are
handy addtitions.
3) Try to estimate the suns postion as a compound angle from the target. You
can do this by placing a pole next the the target and measuring the angle,
drection, and distance of the shadow from the pole. With a little trigonometry
you can back-figure the sun's angle. This will aid in matching lighting. (In
you're case, be sure to pick a billboard that faces the sun, so that the sun is
to your back when shooting.)
4) Take more than one picture to obtain the best scan. If practical, take
pictures from various vantages: 2D images always seem to look different than
they did when you were looking though the camera. Having choices during the
composition phaze is a plus.
5) If you're not a pro videographer or photographer, hire one. It will make a
difference in the end product.
6) Expend your effort in proportion to your budget. The project may not
warrent all the efforts I've described above.
Well, I'm sure there's something I missed, but this will give you a good start.
Experiment, take notes, and have fun!!
— James — Choreo Motion —