Hello dear friends!
I try to explain, why sometimes the_out_of_rules
anaglyphs can be seen well.
We suppose that 2 degrees of deviation is the
ideal!
( Now I dont want open nothing about the otimal
deviation like, 2 or 1,7 or 1,5 degrees??)
The rule says: from the nearest point to the
infinity point, the deviation on the image must be 2 degrees or less. The
nearest point = any point on the stereo window. The farest point = any point
what is on the farest distance.
This is the rule, I think it is clear for
everybody.
For my opinion this interpretation is bad. We must
add that to this rule: " ... if the nearest and farest points are in the same
visual direction..."
Forasmuch an anagylph is viwed with our central
view, what is about 10-15 degrees of FOV, ( not the foveal view, what is
more less, about 1-2 degrees) we are moving our eyes in zigzag on the image
left-right and up-down.
On the Dariush's image we cannot find in the stereo
window ANY object what is visible together with the farest object of the image
when we consider our central vision. The farest point is visible when we seeing
the head of the lady. So we can consider that the nearest point on this image is
the head of the lady and the farest points are the gifts behind her. So
determining in this mode the stereo base what we must use, I think the base used
by Dariush is good. I divided the image in 3 horizontal part. On all parts the
deviation is good, the image is well visible.
The under part: dont has any point form the
infinity, so any object is viwed with the infinity. Middle part: thare are some
parts from the backgorund, but any close object is viwed with the same visual
line what is going to the background. The upper part: here is the background and
it is viwed in the same visual line with the head. Doesn't count that I aligned
this part of the image. If the deviation is good, we can move the red-cyan part,
like this part is visible on the original image.
So I think, that objects in the forefront must be
considered the NEAREST object when its height arrives at the horizont or hangs
slack from up above the horizont.
So if on the Dariush's images we put something
on the stereo window, what is in the same visual line with the farest point, so
we must consider it the nearest point. ( see attacment of the next message)
In this case the used base by Dariush is wrong.
Of course the near objects will soffer some
distorsions, like is visible on the Dariush's image, but if we shot on the
street a standing person, so he will be the nearest object, without
reference to the asphalt before him.
I hope my explanation is clear.
Cheers: Imre
(390K)