120 CONVERGENT VERSUS VERTICAL PHOTOGRAPHY, DISCUSSION
invite our American colleagues to provide the
audience with some hard facts and figures. That
is a real contribution to the evaluation of the
method. That is one question.
Furthermore, you will find mentioned two
different experiments. One is the famous Monti
di Revöira which has been evaluated from the
point of view of precision by the Swedish
Society. Mr Möller was kind enough to give me
an advance report about this. The mathematical
statistician who works on this said that it is very
difficult to come to any conclusion about it.
Whether it is so dangerous, I doubt.
I mentioned a few figures which I derived
on the way I am using for this and found there
that there is a great difference between the
results of the convergent camera and of the
vertical plate and film. We came there to a mean
square error reduced to microns in the plane of
the negative for signalised points to 17 microns
for the plate, 27 microns for the film camera
vertical which is also a high figure, and 34
microns for the convergent photography.
It may be that somebody who is viewing this
from the point of view of statistics would say
that you cannot come to any conclusions from
this. For the time being I am inclined to say
that all these figures are worth very little. I am
afraid that there are other sources of errors in
the picture in this case which in particular make
these figures, I would say in my own terminol-
ogy, useless, but that is rather a strong expres-
sion and therefore it is better to say of very
little use.
In recent months we have experienced very
bad things regarding films. I regret very much
that I must admit this. We measured in Delft the
fiducial marks of the Monti di Revóira pair be-
cause also the Dutch results were unsatisfactory.
A study of this finished in measuring the fiducial
marks, measuring the distance, transforming
on two points and then looking to see what is the
difference in the position of the other two. I
hardly have the courage to say what was the
difference, but it went up to 60 microns. We
cannot talk any more about 10 microns or some-
thing like this, this makes no sense whatsoever.
Another case of film still worse was where
we recently had three rolls of films taken in a
reseau camera. We made prints of two succes-
sive photographs of the same roll, put them in
the stereocomparator and measured the differ.
ence in size of grid squares. The observed grid
squares are with base O the same. The differ-
ences went up 40 microns at the edges of the
film. This are two pieces of film from the same
roll at a distance of 25 centimetres. I will not
say that the film is always bad, we also have had
very good examples. You find these mentioned
in my paper as experiments from Frankfurt,
given to me by Dr Fôrstner. He has had better
luck than we, and than the participants of the
Monti di Revôira test. He finds very nice figures,
7 microns for the convergent photography in the
plane of the negative and for the vertical also
7 microns. For the heights, derived from the
verticals he found 13, and from the convergent
as can be expected, a smaller value, 9 microns.
There is nothing wrong.
These combinations of figures make it very
difficult to determine, at least on the basis of the
solid material I had in hand at that time, which
was the best. ;
At the end a few additional remarks. I have
always expected that one roll of film for both
images, as it was in the pre-war Zeiss Twin
camera, that was a better method than having
two separate rolls as we have now. But after the
bad experience with these three separate reseau
films we are not sure about this. A point of
doubt, however, is that these three rolls are not
processed by a professional organisation, but by
RAF people more or less as a training subject.
But I am still even more than before in doubt
as to what a film really means. There are several
excellent publications about the high quality of
film, but as I have said many times, I would like
to look into the waste paper baskets of all those
people who have made experiments to see
whether they never discovered a film as bad as
ours was in Delft. But we have a kind of
privilege in always discovering bad phenomena.
For the time being that is all I have to say.
Now I would like to ask the audience whether
somebody likes to raise questions.
Discussion
Prof B. HALLERT: I would like to ask some-
body on the panel whether they have tested the
basic geometric qualities of convergent pho-
tographs. In other words, taking into account
the different influence upon the geometry of,
on the one hand, radial distortion from the lens
and on the other hand curvature, reflection and
perhaps other disturbances. This is a question, in
other words, as to whether the basic regular or
systematic errors have been investigated, and
further whether the residual errors have been
investigated.
have
no €
me i
pub.
of 1
friet
tion
our
racy
acci
gre:
coll
Cus:
lica
the
as (
by
hap
lish
bric
con
of :
hac
tior
me
COI
30.
ra
acc
gre
tio
avi
bu
mz
fou
ge
Cel
to
Sc
ea