Minolta Dimage Scan Dual Film Scanner Test Images
 |
|
"Musicians II" image: (820k) The main image here was scanned at 1200 x 800 pixels, and several tonal adjustments were made. The original default scan (640k) was a little dark on our uncorrected Mac monitor, but very dark when we set the screen gamma to the graphic standard of 1.8. To compensate, we pulled the midtone slider of the Dimage Scan Dual's histogram tool down a fair bit, and brought in both the black and white-point slider controls. Color accuracy and purity on the adjusted shot (820k) is good, with bright colors and natural skin tones, although the blue of the Oriental model's robe is a bit under-saturated. (NOTE that this is NOT the identical "Musicians"
image as used in our digital cameras test! It's very similar, but the
models are different, and the digital-camera version is a couple of reproduction
generations removed from this particular version.) |
|
|
|
"Musicians II"
detail clip: (664k) This shot reveals the
good detail captured by the Dimage Scan Dual at its 2438 dpi maximum resolution.
Cropped from the Musicians II slide, with the resolution set to maximum,
the image shows very good detail, particularly in the fine, spiky foliage
of the flowers in the garland on the model's head. Interestingly, although
the Dimage Scan Dual seems to have a fairly high native resolution relative
to other scanners it competes with (as evidenced in the resolution test
below), like its big brother the Dimage Scan Speed, it seems less prone
to emphasizing the film grain. We attribute this to the diffuse light source
it uses (a special fluorescent tube), which doubtless works in the same
way that diffusion heads do on darkroom enlargers: Many pros swear by the
diffusion heads, rather than the more-common condenser ones, because of
exactly this effect. We found that high-resolution scans from the Dimage
Scan Dual took sharpening in Photoshop very well, as evidenced in this
version (1092k!) of the image, which had unsharp
masking applied at a radius of 1.2 pixels and an intensity of 120%. |
||
|
|
Kodak Royal Gold 25
"House" detail clip: (488k) This
is a detail clip from the same negative used to produce the "house"
poster for our digital camera tests. It was shot on Kodak Royal Gold 25
film, which is extremely fine-grained, but which has very different color
characteristics from more ordinary color negative films. Most scanners we've
worked with have had difficulty with RG 25's color balance, and the Dimage
Scan Dual largely followed suit, as shown in this
scan (484k), taken with the default settings.
On the Dimage Scan Speed, we achieved a color correction we were satisfied
with rather quickly, but got a bit pickier with the Dimage Scan Dual. The
result was a fair bit of fiddling, but a fairly good result, as seen in
the main shot. (488k)
This particular negative has either a layer of very fine dirt on the emulsion,
or voids in the emulsion, producing tiny white spots in the final scans.
We found these to be very noticeable on scanners with highly-collimated
light sources, but considerably less obvious with the Dimage Scan Dual,
thanks to its more diffuse film illumination. |
||
|
|
"Train" Shot
(Extreme shadow detail): (716k) This slide
is an exceptionally tough test of scanner dynamic range: It contains areas
of moderately bright highlight, but the shadows are very, very dense. As
would any 8-bit scanner, the Dimage Scan Dual had a harder time on this
shot than do scanners with greater bit depth, showing a fair bit of noise
in the darkest areas. As you'd expect, the scan with
default settings (316k) produced a very dark
image, but we were able to correct both the tonal balance and color shifts
in the deep shadows quite well with Minolta's software. The Dimage Scan
Dual has an option for "autoexposure for slides" that compensates
for underexposed slides, presumably by boosting the gain on the signal coming
from the CCD. We found that it had little impact on this test though, probably
because of the relatively strong highlights in the sandy ground in front
of the train. Note for Mac users with uncorrected monitors:
Because there's so much going on in this image in the lower end of the tonal
range, the noise will be particularly apparent on monitors with gamma settings
of less than 1.8. If you have a screen with uncorrected gamma, you'll probably
find this version of the picture (688k)
more appealing, since it's black point has been adjusted for better viewing
on low-gamma screens. |
||
|
|
Q60 Color Target:
(156k) The Kodak "Q60" color target
(also the basis for the international IT8 target) is a good test of color
accuracy and tonal rendition. The main image (156k) was
scanned with the scanner's white point, black point and gamma controls on
the histogram screen adjusted to boost overall midtone brightness a little,
as well as correct for a rather reddish color shift. (These points were
adjusted for by clicking on the corresponding steps of the grayscale wedge.)
For comparison, here is the default scan (140k),
which is slightly warm-toned, but nonetheless has great color "out
of the box." Tonal range is quite good, with even the transition between
steps 21 and 22 of the grayscale being well-resolved, but with some noise
appearing in step 19, and moderate to severe noise by step 20. (Overall,
good performance for an 8-bit scanner, but not up to that of a 12-bit model,
as you'd expect.) |
||
|
|
Q60 Detail: (184k) (New shot, 8/4/99: Earlier
reviews will not include this) We've just begun including this
image in our scanner tests (August 99), so you won't find a corresponding
one in earlier scanner reviews. We've included this cropped, maximum-resolution
scan of the Q60 test slide, since it is a more-accessible test target for
others to use in comparing output from different scanners. Detail is very
good, and the image takes sharpening well, as we noted in the "Musicians"
clip above, and as shown here (216k). |
||
|
|
"Davebox"
test target: (508k) Here is our official
"weirdness of color negative film" test target. Most scanner seem
to have a hard time with this shot, taken on Kodak Gold 100 color negative
film. The Dimage Scan Dual actually did a very credible job of scanning
it though, as shown in this unmodified default scan
(476k). The default scan did look a bit washed-out
though, as the midtone values were rather high, and the white point down
a little relative to the available tonal range. This proved quite easy to
correct, by pushing the "Gamma" midtone slider up, darkening the
midtone values, and pulling the white point down, brightening the whites.
ON this image, we found ourselves a bit frustrated with the Dimage Scan
interface: The preview image provided to shw the results of manipulating
the histogram tool didn't accurately represent the effect of the changes
we made, leading to multiple attempts at the color adjustment. This in turn
led us to strongly wish for the scanning software to retain the prior preview
scan, to avoid the lengthy pre-scan process each time we wanted to make
another change. We also wished that the eyedropper control on the histogram
interface would give a readout of the RGB values after the histogram
change. (Hope the Minolta software engineers are reading this!) The end
result (508k) though, was a very snappy image,
with bright, saturated colors and good tonal range. |
||
|
|
WG-18 Resolution Target Horizontal Clip: (40k) The full WG-18 (Actually now ISO 12233) resolution target is very large (see below), so we cropped-out these clips to show the scanner resolution on this familiar target. These scans were made from a target shot on Kodak Technical Pan black & white negative film. This film is extremely fine-grained, with perhaps the highest resolution of any commercially-available 35mm emulsion. The target was shot with a Nikon 50mm, f1.4 lens (a notably sharp lens, particularly when stopped-down), at an aperture of f8. Thus, while not a "laboratory" grade target, this represents about as much detail as you'll ever see in a conventional film image. The downside of this target is that the Tech Pan emulsion is a little "thin," lacking density. There's thus a fair bit of compensation required of the scanner controls to restore it to a full contrast range. The image here had the white point adjusted pretty far downward, to pull the anemic "white" of the raw film up to the full value of 255. (This image was scanned in black & white negative mode, to best show the maximum detail possible with the scanner.) This is an interesting shot, because the visual resolution appears to
be quite a bit higher than should be theoretically possible. Given the
roughly 2300 vertical pixels in the image, you should really only be able
to see about 1150 lines of resolution. Instead, detail is clearly evident
down to about 1400 lines per picture height. We frankly don't know how
to explain this, other than by observing that the contrast ratio at that
point is pretty far down from the 100% level set by the lower-frequency
elements of the target. Possibly the low amplitude at that frequency helps
to conceal the aliasing, along with the natural frequency rolloff provided
by the non-"laboratory" target quality. Whatever the cause,
the Dimage Scan Dual did the best of any scanner with similar resolution
(~2400dpi) we've tested to date! Color aliasing is also virtually nonexistent:
A RGB color version (64k) shows
no sign whatsoever of color aliasing. |
||
|
|
WG-18 Resolution Target
Vertical Clip: (40k) Here's the corresponding
vertically-oriented clip of the WG-18/Tech Pan test target. Again, visual
resolution appears to extend to at least 1400 lines per picture height.
Once again, the color version (40k)
shows no sign whatsoever of color aliasing. |
||
|
|
Full-Size WG-18 Resolution Target: (1,295k!) For the real masochists, here's the full-size resolution target, scanned at the maximum resolution of 2438 dpi. A side note: We don't have an explicit test for frame coverage by scanners, but this test shot goes right to the edges of the 35mm frame. With the slide holder, the Dimage Scan Dual covers all of that and then some, but we found that the strip-film holder crops the image somewhat, on the order of about 3% in each direction. This image was scanned in black & white negative mode, resulting in a single-channel, grayscale-only image. WARNING: This JPEG expands into a 7.3 megabyte
file, which may crash your broswer if viewed directly! To view it, you
must first download it directly to your hard drive (right-click in Windows,
click & hold in Mac Netscape), then open it in an image-editing application.
here is the link to the RAW
JPEG IMAGE (1,110k) (No surrounding HTML file.) |
||
|
|
USAF 1951 Resolution Target: (204k) (Elderly technoids only ;-) Old-line lens and film testers will be well-familiar with the "USAF 1951" resolution test target. (1951 is the year it was created, giving you an idea of what we mean when we say "old-line.") This was scanned at the maximum resolution of 2438 dpi from a laboratory-grade target (chrome on glass slide) before being cropped down. Because this target doesn't have the sort of extended structures that the WG-18 one does, it doesn't permit the visual interpolation our eyes can do on the WG-18 pattern. As a result, the USAF target consistently yields much more conservative resolution numbers. As expected, the USAF resolution target in fact gave more conservative
resolution figures, with the scanner resolving cleanly down to 28.5 line
pairs/mm, (724 line pairs/inch), and with some aliasing, down to 45.3
lp/mm (1151 line pairs/inch). |
