Re: Stereo Window - Poll time!

[P3D Larry Berlin <lberlin@xxxxxxxxx>]

>Date: Sat, 21 Jun 1997
>From: P3D john bercovitz  writes:
>
>John Bercovitz writes:
>> Some things to keep in mind during your deliberations.
>> Convergence is a weaker depth clue so don't think of the
>> eyes as triangulating on something to find its distance
>> unless it's very close.........
>
>Larry Berlin writes:
>> This comment is interesting. I'm inclined to see convergence 
>> as a rather strong clue in stereo images particularly. ........
>> The fact that the *eye/brain will establish their own infinity 
>> from scene context* is the mechanism by which a base for detailed 
>> relative triangulation is established. 
>.................................................................
>I'm not sure I understand precisely what you're saying here, 
>Larry.  I'm saying the map is built up by whatever means and 
>convergence is a secondary and weaker clue.  I think you're 
>saying convergence is the primary clue, the start of the map.  
>Have I got that right?  If not, there's no point in reading 
>further; let's discuss your thesis first.

******  While convergence may be somewhat secondary, I don't consider it to
be so much of a *lesser* clue, certainly not weak. I use convergence
triangulation all the time to a very strong degree, but not exclusive of
other sensations.  I'm saying that the original map is built up by various
means including significant and strong convergence factors.  It's a  factor
that helps sharpen the overall effect as well as build it. 

There are factors of scale here. Convergence to the window for example
defines the larger image space in certain key ways. Relative convergence
within the image is very important for providing details and specific depth
comparisons.

> 
>*************************************************************
> ........ You can find
>pairs with separation anywhere from 25 mm to 100 mm.  For ref-
>erence, I freeview from 400 mm because that's where my glasses 
>are set.  When I freeview a pair separated at 25 mm, I see 
>infinity lying at infinity, not at 600 mm.  When I freeview a
>100 mm pair, I see infinity at infinity, not 800 mm behind my
>head.

*****  At some precise geometry of image and viewing circumstances, infinity
points can be said to approximate true infinity. However, most persons can
arrange their eyes to move slightly wider than actual parallel. If the
perfect image arrangement is further separated by a small amount, you still
perceive infinity in the image as if at infinity, except that it has been
placed slightly farther away in real-space geometry, even to the point of
being beyond infinity's defined limit. The same holds true when the image is
reduced in size or brought closer together. There is still an infinity point
relative to the points within the image. It still looks like infinity,
though it has a specific place in space (convergence) that in actuality is
much closer than true infinity. Such are the artifacts of image perception
as they differ from reality. Stereo images, being fixed records of spatial
relationships, have a finite infinity. We can be successful at overlaying
that finite arrangement over real factors to simulate true infinity though
most viewing is seldom that accurate and doesn't need to be.

Larry Berlin

Email: lberlin@xxxxxxxxx
http://www.sonic.net/~lberlin/
http://3dzine.simplenet.com/


------------------------------