T3D Re: Mounting quality

[abram klooswyk <abram.klooswyk@xxxxxx>]

Tony Alderson wrote (21 Sep 1998 TECH-3D Digest 362):
>Has anyone ever done surveys of the quality of mounting in 
>stereo competitions? (...) what is the actual practice?

I mentioned (T3D digest 364, 23 Sep 1998) the survey carried out 
in 1975, by Jac. G. Ferwerda, Koo for his friends.
At the request of the Nederlandse Vereniging voor Stereofotografie, 
in 1975 the Stockton-on-Tees International Stereo Exhibition was 
combined with the Dutch International Stereo Exhibition, the 
only PSA approved exhibition ever held in the Netherlands. 

All entries were brought from Britain to Holland and judged
once more by a Dutch jury. They were supported by Pat and Jim Milnes 
from Stockton, with their long experience in running competions.
There were in total 613 entries from 27 countries (country with most 
entries: U.S.A., only 27 entries from the Netherlands). Ferwerda was 
one of the judges, Hugo de Wijs projectionist. 240 entries accepted 
for the show, held 5 times, in different parts of the Netherlands, 
for a total of nearly 800 persons.

Because there were reports of eyestrain in viewing the show of accepted
slides, Koo after some hesitation decided to measure a sample of the
show. From the 240 projected slides, he measured only 5P (Realist type)
slides. Because of limited time, from them only the first 79 consecutive 
ones and the last 50 consecutive ones. Not all measument could be made 
on all slides, some had hazy contours.
Results were published in the Netherlands Bulletin and in the one of the
British 'Third Dimension Society'. The latter publication is in the
literature references of Koo's book "The World of 3-D", but I forgot 
whether the book itself uses figures from the articles (couldn't find 
it in the index).

The English article is a treatise of 15 pages whith measurement results
intermingled with comments, opinions and advice on several stereo and
mounting problems, later on expanded in Ferwerda's book (1th ed 1982).
In the following I will summarize mainly the results. Deletions 
indicated by ... or (...), my comments in [].
-> For those who want to repeat a survey like this (which I would 
appreciate very much) there is a note at the end of this posting.
======================

Excerpts from: "Remarks on mounting and projection of Realist format
slides", by Koo Ferwerda. 
The Third Dimension, No 69, spring 1976, pages 24-38.

Our first showing was on 8 november 1975 ... many applauses for fine
slides ... our International Stereo Show seemed a succes.
... second showing ... after the show ... many said it was very
beautiful, yet rather tiring. Some ... good friends had gone home
because of headache. ... I expected ... good advertisement for 
stereoscopy ... disappointment for me.
... Suspected way of mounting ... decided ... to make measurements
... Realist size ... left out 24 x 30 mm slides ... spent several days 
in measuring ... found most astonishing errors. ... 
Measurements themselves also ... inaccuracies ... suppose ... error up 
to some 1/10 mm in each ... will not influence ... general trend.
[I remember discussing the method with Koo, suggested microscope with 
low magnification, crosshair ocular and x-y movable nonius device to 
hold the slides, this was not at short notice available, so Koo used 
another method, I forgot which one exactly]
(...)
Heigth Errors ... on each slide two pairs of clearly visible homologues
... near right [edge] and near left [edge] of the images. Distance of
these four points to upper edge of the [mounted] slide (on which it 
rests during projection) were measured ... difference derived ... [if 
left unequal to right pair of homoloques] mean has been taken.

Table 1  - (...) Heigth errors [in mm, on 118 slides]
------------------------------------------------------------------------
H. diff   =<-0.67    -0.45    -0.22    0      +0.22    +0.45     =>0.67
------------------------------------------------------------------------
class limit     -0.56    -0.33    -0.11  +0.11     +0.33    +0.56
------------------------------------------------------------------------
V.A. diff   -66'      -44'     -22'    0'      +22'     +44'      +66'
------------------------------------------------------------------------
Numbers      1         7        21     62       21       5         1
------------------------------------------------------------------------

[H.diff = heigth difference, V.A.diff = vertical angle difference]
[explanation]... figure 5 in last row: of 118 slides 5 in the class 
+0.45 mm (between 0.33 and 0.56 mm) ... right eye must be directed
44 min of arch upwards compared with the left eye [for a] spectator
[in the] orthostereo seat (...).

>From my own experience I assume that an angle difference of 20'-30' 
should be tolerable, but this may vary from one person to another. (...) 
only 14 out of 118 ... show this kind of misalignment ... not very
serious.
However, if two slides with respectively positive and negative heigth
error follow each other (...). 

(...) Twist. from the ... former measurements I derived also the twist.
... angle of rotation [left to right] image ...

Table 2 - (...) Twist (Rotational Error) [in min of arch, on 118 slides]
------------------------------------------------------------------------
Angle of twist <-85'   -68'    -34'     0'     +34'    +68'    >+85'
------------------------------------------------------------------------
class limits       -85'    -51'     17'   +17'     +51'    +85' 
------------------------------------------------------------------------
Numbers          1      4       33      45      23      10       2
------------------------------------------------------------------------
In addition to measuring the twist (...) checked by studying the bottom
edge of the picture (...). From my own experience ... twist of only 
about 30' is uncomfortable. [then] at least 50% of the ... slides showed 
this error. I think this is one of the main causes of tiredness.

Infinity Separation. ... Realist system ... fixed ... should be 63.4 mm
... [when] on screen larger than 6,5 mm ... eyes forced to toe-out ...
unpleasant for nearly everyone ... will cause eyestrain.
On slides with distinct infinity points (most landscapes) I measured ...

Table 3 - (...) Infinity Separation [in mm, on 65 slides]
------------------------------------------------------------------------
Infinity sep.   62.6   62.9   63.2   63.5   63.8   64.1   64.4   64.7
------------------------------------------------------------------------
Number           1      4      15     16     15     8      4      2
------------------------------------------------------------------------
[In the projection of the show] ... the too large infinity separations 
... many projector adjustments, especially of course when following a
small infinity separation. (...)

Parallax and Deviation. (...) accordance ... definition of parallax 
given in astronomy [Ferwerda studied astronomy and mathematics, wrote a
thesis on an astronomy subject] parallax of a point of the subject is 
the angle between the lines connecting the subject point with two 
different viewing points ... for instance two eyes ...[or] two lenses...

(...) Deviation ... from ... Earl E. Krause ... difference ... on slide
of [homologues of certain point] with ... infinity separation. (...)

Table 4 - Deviations (and Parallaxes) of masks in stereo landscapes
          [in mm, on 65 slides]  
------------------------------------------------------------------------
Dev. of mask   0.2    0.5    0.8    1.1    1.4    1.7    2.0    2.3
------------------------------------------------------------------------
Par.(converg.) 20'    50'    80'    110'   140'   170'   200'   230'
------------------------------------------------------------------------
Number          2     11     12      16     12     9      2      1
------------------------------------------------------------------------
On the slides used in table 3, I also measured the mask seaparation,
which appeared to vary between 61.9 and 62.6 mm and from these I derived
table 4. ... the parallax of the mask, that is at the same time the 
angle of convergence when the eyes change from looking at infinity to 
looking at the [images of the] mask (as seen from the orthostereo seat).
(...) [paragraph on double depth deleted]

Close-up and Macro Stereo. (...) infinity ... not included ... far- 
separation ...  near-separation ... range in deviation ... I measured 
it on all close-ups (subjects obviously nearer than 2 m) and macro's...
for near separation I took the mask separation. (...) negative range of 
deviation indicates that the window is *behind* the farthest point.

Table 5 - Deviation Range in close-up and macro stereo [mm, 55 slides]
------------------------------------------------------------------------
Dev Range   -0.4  -0.1  +0.2   0.5   0.8   1.1   1.4   1.7   2.0   2.3
------------------------------------------------------------------------
Number        1     1     2     9    10     6    10     7     5     4
------------------------------------------------------------------------

(...) range exceeding 1.2 mm ... farthest points ... separation beyond
6.5 cm on the screen ... eyestrain. ... mounting with ... double depth
... can improve this ... but ... stretch.
[paragraphs on orthostereo seat, on advice for longer camera focal 
length, on projector focal length deleted]

Conclusions. (...) correct mounting is very difficult ... 
I possess the patience to give as much as a quarter of an hour for 
putting only one stereo pair in the mask, I admit that ... some 
inaccuracy can be observed during projection of my slides. 
[but very few people actually succeeded in this!]
... three major errors ... 
a) twist, b) too large deviations, especially in close-uppery, 
c) ... infinity separation differing too much from 63.4.

[ideas about perforation mounting of 5P, double depth, using a longer
camera focal length to improve projectability of 5P, new type of macro
camera deleted]

[Koo Ferwerda 1976]
======================

Note for those who would undertake to repeat these measurements on 
slides of a present day exhibition:
When glass mounted slides are measured by placing them on a grid, care
must be taken to avoid viewing parallax. Already small movements of the
observing eye can change readings considerably.
Abram Klooswyk


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