Re: 'useless' resolution

[T3D john bercovitz <bercov@xxxxxxxxxxx>]

John Ohrt writes:
> Could you please give a more detailed reference?


Here we go from T3D digest #190

> Gerald Westheimer's reply to Jim:
> "Westheimer G, McKee SP  (1980) Stereoscopic acuity with defocused and
> spatially filtered retinal images.  Journal of the Optical Society of
> America 70:772-778. Fig. 2 shows impairment of stereoacuity with spectacle
> blur, Table 1 compares uniocular and binocular blurs.  A 2.5 diopter
> monocular blur gives about twice the stereoacuity deficit of an equivalent
> binocular blur."

=======================

And from P3D:

> Date: Sun, 22 Aug 93 19:19:58 PDT
> From: bercov@xxxxxxxxxxx (John Bercovitz)
> Subject: stereo article #2

J. Opt. Soc. Am. Vol. 70, No. 7, July 1980, Ppg. 772 - 778
 
"Stereoscopic acuity with defocused and spatially filtered 
retinal images."
 
Gerald Westheimer and Suzanne P. McKee
Dept. of Physiology-Anatomy, Univ. of Calif., Berkeley
 
Abstract:  Spectacle blur decreases stereoacuity more than 
ordinary visual acuity, and this cannot be accounted for 
by contrast reduction or by accommodation and/or 
convergence instability.  Stereoscopic acuity was measured 
over a range of defocus exceeding 2 D in normal subjects 
with a 3-mm artificial pupil.  Blur of one eye leads to a 
decrement in stereoacuity that is at least as high and 
usually higher than blur of both eyes; blurring only the 
disparate element of the pattern while keeping the rest 
clear is nearly as detrimental as blurring the whole 
pattern.  In a separate experiment, the individual line 
elements of the pattern were subjected to spatial 
frequency filtering in the horizontal direction.  For best 
stereoacuity values, representation of the whole spatial 
frequency spectrum is required.  Patterns with diminished 
high-frequency content (>10 cycles/deg) show lowered 
stereoacuity and those utilizing predominantly a spatial 
frequency band centered on 22 cycles/deg even more so.
 
My synopsis:  The authors' apparatus presented the 
subjects with three bright, high-contrast lines against a 
dark background.  The outer lines were at the distance of 
the images and the inner line could be moved in or out 
from the outer lines in three steps; also there was one 
position of the middle line which was at the same distance 
as the outer lines.  The stereo pair was flashed on the 
screen for 1/2 second out of 3 and the subject threw a 
switch to indicate whether the middle line was closer or 
further than the outer lines.  In between test images, a 
square was put up to maintain the subject's convergence 
and accommodation.  If the subject was correct 75% of the 
time, he/she was considered to be at the stereo threshold 
for those conditions.  The authors compared the 
degradation of stereo acuity to the degradation of 
ordinary acuity effected by different strengths of 
positive lenses from one half to three diopters.  The 
authors also did a series of tests where they defocused 
the images with ground glass held at different distance 
from the images.  This series did not use an artificial 
pupil.  
 
The authors presented their results in graphs and made a 
lot of observations based on their data, much too much for 
me to spell out here.  They presented it compactly in the 
article.  They made comments to the effect that the center 
of a blurred image ought theoretically to be possible to 
find quite exactly but they were not able to help the 
situation by increasing image contrast.  One graph shows 
data from their locally-famous subject who has a stereo 
acuity of three seconds of arc under ideal conditions.  
Anyhow, this one's a great read.
 
John B

=======================

Kumar writes:
> Others of interest might be: "Disparity range for local stereopsis as a 
> function of luminance spatial frequency." Schor, C and Wood, I. Vision 
> Research 23 pages 1649-1654.

> "Spatial tuning of static and dynamic stereopsis" Schor, C. et al. Vision
> Research 24, pages 573-578.



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