Re: Double Depth

[T3D Jeremy Hinton <Jeremy.Hinton@xxxxxxxxxxxxx>]

John B wrote:
>  So what is the real physiological(?) limit?  
(for allowable deviation) and then alludes to interesting 
experiments which demonstrate a thorough scientific approach!

>From a practical point of view, I think Bob Aldridges' point that
the limit comes from projection sounds good.

Back to the 'physiological limit'... Ancient Visual Perception 
texts (eg Davson, 'Physiology of the eye' section iv), discuss
PANUM'S FUSIONAL AREA. This is essentially the range of disparities 
with which objects can be perceived as fused. I'll relate a bit of 
his argument:

INTRO
Start with the corresponding points (to which the foveas of both 
eyes project). These have zero disparity. Points at other positions 
in space which also project to corresponding points on the retina are 
descrbed as lying on the HOROPTER. Points at greater or 
lesser distances have slightly disparate points on the retina 
(OK, I know you know this, but I have to start somewhere). Within 
Panum's fusional area, these points are perceived as fused, lying 
at different depths. Beyond Panum's fusional area, they are seen 
as double images. The disparity is usually exprssed in angular terms. 
Measured fusional limits seem to be between 5-20 minutes of arc.
(greater in peripheral vision, where acuity is poorer). 
This would extend both nearer and further than the point on 
which the eyes are converged.

But is this the answer?... NO!. 

When disparities become so great that fusion breaks down, two 
things can happen. 1) We change fixation (convergence), and 2) we 
perceive a double image. It turns out that there is no discontinuity 
in depth perception associated with this 'breakdown'. Disparities of as 
great as 7-9 degrees can be used as cues to relative depth. (Blakemore C 
(1970) 'The range and scope of binocular depth discrimination in man' 
J. Physiol _211_599-622) 

I'm sure Julesz (foundations of cyclopean perception) has some useful
conclusions, but I don't have the book to hand 8(.

Here's hoping I got the length and detail about right.

Cheers
Jeremy


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