Bill Tuthill said:
I already cited this URL, but because I have lots of patience with people
who don't read all the thread (myself being one of them) I will cite it
again. This guy Thierry Burlot complains that his SA-21 crops 3mm total
down to 21mm (1.5mm at top and bottom) from his film feeds:
http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=008LJg
This is at odds with what Kennedy has been saying, and also at odds with
Nikon's manual, which says cropping is only 23.3 vertically. I don't know
the truth of this matter, not having any friends who both care about it
and own an SA-21. Nor has anyone followed up with precise measurements
in the above photo.net thread.
OK Here are some figures I just measured for a Nikon LS-4000ED using
4000ppi scans together with the Nikon specifications:
Using SA-21 motorised adapter with unmounted film strip of 6 frames:
Full area: 3946 x 5959pixels (25.1 x 37.8mm)
Nikon specification: 3946 x 5959pixels (25.1 x 38.0mm)
Active area: 3703 x 5643pixels (23.5 x 35.8mm)
Nikon Specification: 3654 x 5646pixels (23.3 x 36.0mm)
The active area is after cropping in Photoshop to remove all of the
border, whether that it due to the edge or the frame on the film itself
or a crop due to the SA-21 scan aperture.
When using the MA-20 slide adapter:
Full area: 3946 x 5782pixels (25.1 x 36.7mm)
Nikon Specification: 3946 x 5782pixels (25.1 x 36.8mm)
Crop using Gepe mounted slide (just normal type, not a full frame):
Active area: 3605 x5498pixels (22.9 x 34.9mm)
The active area was again cropped in Photoshop. Note however, that the
slide mount had slightly rounded corners and this crop was internal (no
black corners). Consequently, depending on where exactly on the rounded
corner the crop was made, there could be some trade-off between the
width and the height in an acceptable crop. Also, when scanning
oversized 38mm square slides in standard 2" mounts, the full area (25.1
x 37.8mm) of the MA-21 was available for the active image. In other
words, the crop was definitely the slide mount not the scanner.
Several points from these measurements are clear.
1. In specifying the active area of the SA-21, Nikon are conservative.
In my result the width of 35.8mm above is the actual edge of the frame
recorded by the camera on the film. With an oversized camera frame the
SA-21 would actually capture an active area of 23.5 x 37.8mm. In terms
of height, 50 pixels (0.3175mm at 4000ppi) more than the Nikon
specification are available within the cropped area of the active image.
2. With the MA-20 the number of pixels is *exactly* as the Nikon
specification (differences in physical dimensions are probably rounding
as I just read these numbers off the Photoshop Image size dialog box,
which is only accurate to +/-0.1mm).
3. The nominal height crop that occurs with the SA-21 automatic feeder
is actually LESS than the crop produced by a standard Gepe mount (both
glass and glassless mounts were checked, since I wanted to see if I had
any full frame mounts to hand - I didn't, and both types I had were the
same dimensions).
I don't know what Thierry Burlot was doing to lose a total of 3mm from
the height of his frames in the SA-21, either physically with the film
or in terms of calculating the lost dimensions. However his results are
not in agreement with my data above nor with the Nikon specification.
Perhaps his scanner is faulty or he is feeding the film into it in a
very strange manner to offset the frame relative to the adapter gate, or
his camera is out of specification and recording the image
asymmetrically on the film - although that much would overlap the
perforations - or his arithmetic is askew, or his measurements just
aren't up to par. Who knows?
Irrespective of Thierry's problem with his scanner, film, camera or
arithmetic, the data above are the precise measurements of a Nikon
scanner. Unfortunately I don't have any of my older Nikon scanners
around anymore, to check if the older SA-20 motorised feed cropped
tighter, but I don't recall there being a significant difference. Also,
I have some oversized scans taken from some of those 38x38mm slides with
an ancient LS-20 scanner, and an old FH-2 film strip holder I have from
that scanner has frame dimensions of exactly 24x36mm. So I have no
doubt that all Nikon scanners have scan areas which are at least as
large as the nominal 35mm frame. ie. the scanners do *NOT* crop a
nominal 35mm frame!