I've begun including links in our reviews to a Thumber-generated
index page for our test shots. The Thumber data includes a host of
information on the images, including shutter speed, ISO setting, compression
setting, etc. Rather than clutter the page below with *all* that detail,
I'm posting the Thumber index so only those interested in the information
need wade through it!
Default exposure (no exposure compensation),
Daylight white balance. Exposure is just right (it just barely
holds the strongest highlights), but midtones will look a little
dark to non-pros.
One notch of exposure compensation (+0.5EV)
and the highlights get totally lost. I really wish the S2
had 1/3 step exposure comp adjustments!
Excellent color (apart from a slight greenish tint from the white balance),
great detail, low noise.
The extreme tonal range of this image makes it a tough shot for many
digicams, which is precisely why I set it up this way. (And why I shoot
it without any fill flash or reflector to open the shadows.) The object
is to hold both highlight and shadow detail without producing a "flat"
picture with muddy colors.
The S2's default settings produced excellent results here. Surprisingly,
the camera got the exposure nearly perfect with no exposure compensation
adjustment. (This is quite unusual. - This high-key subject almost always
tricks exposure meters into underexposing. The default exposure was technically
just right, but I'd have preferred a slightly brighter image. The minimum
boost I could give it with the S2 was +0.5 EV though, which was really
too much. (I'd have liked to try +1/3 EV.)
I ended up shooting this with the Daylight white balance setting, which
resulted in a slightly greenish tinge. The shot with Auto
white balance was slightly more greenish, and the Manual
option produced a rather warm result. (I found the warm cast from the
Manual option more pleasing, but it was technically less accurate, so
I went with Daylight.)
Apart from the slight greenish cast, color is very accurate and pleasing.
I really like the S2's handling of Caucasian skin tones, an area in which
it's predecessor the S1 also excelled. Besides the skin tones, the S2
got the blue flowers almost exactly right, a color that seems to give
most digicams I test fits.
Resolution is excellent, with lots of crisp detail, and noise is quite
low, even in the shadows. Overall, an excellent job with a difficult (way
contrasty) subject.
Minus 0.5 EV of exposure compensation,
and the exposure is just about right.
The default exposure was a bit bright
in this shot.
Once again, excellent skin tones. This shot required one notch of
negative exposure compensation though.
Once again, the S2 did a beautiful job with Marti's skin tones. Detail is
exceptional, with the individual strands of her hair very delicately and
crisply rendered. (And this is in the S2's 6 megapixel, uninterpolated mode:
The SuperCCD interpolation does indeed bring out slightly more detail,
due to the diagonal orientation of the sensor pixels.)
Exposure details here are similar to the shot above, except this time
I dialed in -0.5EV of negative exposure compensation, as the default exposure
left the highlights in Marti's hair just a bit too "hot."
The extreme contrast of these outdoor/full sun shots are intended to
really test a camera's ability to handle high contrast, wide dynamic range
subjects. Judging by how poorly most cameras do with this subject, I guess
it's a pretty tough test. Accordingly, I was quite impressed with how
well the S2 Pro did with it. It managed to hold onto highlight details
without producing overly dark midtones, and while simultaneously holding
good detail in the shadows. Also, there's no question that the S2 Pro
is a great camera for portraits - Even in this horrible lighting, it manages
to do an excellent job with Marti's skin tones. Very impressive!
Good illumination from the internal flash, nice integration with Nikon
external speedlights.
The S2's flash illuminates the subject well, with good color and brightness.
I found that I needed to use a little negative exposure compensation with
the internal flash operating in its normal mode (top photo at right),
but a little positive compensation when I combined the flash with a longer
shutter speed, working in shutter priority mode (middle photo). Finally,
I attached a Nikon SB-80DX speedlight to the S2, and set it to bounce
from the ceiling, but with the little reflector card popped up to throw
a little direct light on the subject. In this mode, I found I need to
add a full f-stop of positive flash exposure compensation, but the result
was beautifully even lighting. (As always, I strongly recommend external
flash units for any digicams that can work with them - the results you
can achieve are so dramatically better than anything the onboard flash
can do.)
Good color with custom white balance, although still a slight greenish
tinge.
This shot is always a very tough test of a camera's white balance capability,
given the strong, yellowish color cast of the household incandescent bulbs
used for the lighting. The S2's white balance system did a pretty good
job with it, but I'd have liked to see the Auto setting produce a slightly
(not much, but a little bit) more neutral image, and the Custom white
balance option resulted in a slightly warm-toned image. (Overall though,
I prefer a slightly warmer look to this image, to preserve some of the
mood of the original lighting.
It should be no surprise that the Incandescent white balance preset produced
a very warm-toned image here: The S2's incandescent white balance is doubtless
balanced for professional tungsten studio lighting, which has a color
temperature of 3200K. The household incandescent lighting used in this
test has a color temperature of only 2440K.
As in several of the studio shots below, I found that the white balance
setting affected exposure significantly, with the Auto white balance setting
producing brighter exposures than the others. In this case, the Custom
white balance option need a full f-stop more positive exposure compensation
than did Auto. It looks like the S2 balances its color by reducing the
gain of the "hot" color channels, rather than by boosting the
gain of the ones with lower signal levels. Thus, the more the camera corrects
the color balance, the more the effective exposure is reduced. This can
lead to a good deal of uncertainty in exposure determination, a highly
undesirable camera behavior.
ISO Series: The S2's noise performance is rather interesting. Its images look
"cleaner" to the eye than their actual noise numbers would suggest.
I think this is because the camera manages to keep the luminance noise
quite low, with most of the noise appearing in the color or chrominance
channels. However you slice it though, the images I shot here of Marti
at a range of ISO settings look pretty good, even at ISO 1600. (Apologies,
these shots are all slightly underexposed, as I didn't compensate enough
on these for the Custom white balance setting's impact on exposure)
Manual white balance is dead on, excellent color, fantastic resolution.
The Auto and Daylight
white balance settings were just a bit off, but the Manual
white balance was spot-on, producing almost an almost perfectly neutral
tone on the white paint of the bay window. Resolution is really exceptional
here, clearly approaching the limits of the test target itself. (Which
is pretty amazing, considering that the original was shot on 4x5 film
with a very sharp lens and scanned to a file size of 500 MB.)
In evaluating these test images for the S2, I've been comparing them
to the matching ones shot with the Canon D60 and Nikon D100. As I work
my way through the library of images, it's becoming increasingly clear
to me that the S2 has a slight but definite edge in resolution and sharpness
over both competing cameras. In this image, the S2 seemed to do a better
job of resolving the fine details of the branches against the sky, above
the central gable of the house. I was particularly intrigued by the S2's
interpolated resolution mode. Fuji has made much of the unusual diagonally
structured array of photosites in their SuperCCD sensors. The process
of optimally converting this diagonal array of original pixels into the
rectilinear arrangement of pixels in the final file results in a 2x increase
in the number of pixels, a factor of about 1.4 in both horizontal and
vertical directions.
Interpolation has a lot of negative connotations in the digicam market,
because it has been used by a number of companies to artificially inflate
the pixel counts of their cameras. In the case of the SuperCCD though,
it appears that there is indeed some validity to the interpolation that
occurs as a natural consequence of the diagonal pixel structure. It's
difficult to directly compare sharpness and detail in images of different
sizes, so I took the interpolated image of the house here, and resampled
it down to the same size as the "uninterpolated" 6 megapixel
file. I then applied unsharp masking in Photoshop to both images, in the
amount that seemed to produce the best rendering of the fine detail in
the tree limbs. The result was very interesting, as shown below. - There's
no question that the downsampled and sharpened interpolated image
is crisper and shows more detail than the original 6 megapixel file from
the camera. (And the resulting image is likewise noticeably sharper than
anything I could extract from the 6 megapixel files from the Nikon D100
or Canon D60.) Here's a comparison of images from the three cameras. (The
S2 Pro and D100 used the same lens at the same focal length, the D100
at f/10, the S2 Pro at f/8. (A difference in aperture, but I think not
enough to be responsible for the sharpness difference seen here.) The
D60 image was shot with Canon's much-acclaimed 100mm f/2.8 macro lens
at an aperture of f/6.3. Unsharp masking for each image was set to produce
the maximum apparent sharpness, without introducing artifacts visible
in other parts of the image.):
Detail
Comparison, 3 Pro SLRs
Fuji S2 Pro
6 megapixel
USM 150%, 0.3 pixel radius
Fuji S2 Pro
12 megapixel file, downsampled to 6 megapixels
USM 170%, 0.3 pixel radius
Extraordinary
detail, with accurate color and a good dynamic range.
This image is shot at infinity to test far-field lens performance. NOTE
that this image cannot be directly compared to the other "house"
shot, which is a poster, shot in the studio. The rendering of detail in
the poster will be very different than in this shot, and color values
(and even the presence or absence of leaves on the trees!) will vary in
this subject as the seasons progress. In general though, you can evaluate
detail in the bricks, shingles and window detail, and in the tree branches
against the sky. Compression artifacts are most likely to show in the
trim along the edge of the roof, in the bricks, or in the relatively "flat"
areas in the windows.
This is my ultimate "resolution shot," given the infinite range
of detail in a natural scene like this. The S2 really excelled here, capturing
an extraordinary level of fine detail throughout the frame. The fine details
in the tree limbs and foliage above the roof are rendered with exceptional
delicacy. Shadow detail is excellent as well, and the strong highlights
on the bay window on the front of the house stop just short of blowing
out. (This shot was taken under slightly hazy conditions though, so the
test is less severe than as if it were shot on a crisp fall day, when
the lighting would be even more contrasty.)
On this photo, I again found that the "secret" of downsampling
and applying unsharp masking in Photoshop that I tumbled across on the
House poster test yielded even more exemplary results here. Below are
shots comparing the raw 6 megapixel image as it came from the camera,
the interpolated 12 megapixel image, and a sample of the 12 megapixel
that has been downsampled to 6 MP in Photoshop and then had careful unsharp
masking applied to it (200%, 0.2 pixel radius). (Click here
to see the full image that received the downsampling/unsharp masking treatment.)
Dave's
"Secret Sauce" applied to S2's 12 MP file
This is a series of clips from the
far-field shot, illustrating the effect of downsampling and sharpening
on the interpolated 12 megapixel file size.
The results are fairly dramatic, revealing more detail than can
be seen in the 6MP files produced in-camera.
(Not shown is the result of simply sharpening the 12MP image
- It just didn't seem to respond as well to unsharp masking as
did the downsampled version.)
(Note added a bit later in the review: Hmm - Actually, the usual
approach of unsharp masking an image shot with in-camera sharpening
turned off seems to work at least as well as the above. - I'll
try to experiment a bit more and report further at a later date...)
Original 6MP camera file
(6 megapixel file size, no unsharp masking, exactly as came from
the camera - Default in-camera sharpening.)
Original 12MP camera file
(12 megapixel file size, no unsharp masking, exactly as it came
from the camera. - Default in-camera sharpening.)
Original 6MP camera file, slight sharpening
(6 megapixel file size, Unsharp masking applied: 200%, 0.2 pixel
radius.)
Original 6MP camera file, more sharpening
(6 megapixel file size, Unsharp masking applied: 200%, 0.3 pixel
radius.)
Resolution Series: The table below holds pointers to the full range of resolutions and
JPEG image qualities produced by the S2 Pro. I didn't include the RAW
format image, thinking it would be of little use to anyone without Fuji's
special software. Likewise, the 12 megapixel interpolated TIFF file was
just too huge (37 megabytes) for me to want to incur the bandwidth costs
associated with thousands of downloads of it. - If you really think
you need to see the TIFF, email
me and I'll arrange to get you a copy.)
ISO Series: Image noise in the S2 is pretty well controlled, with low levels of
luminance noise, but more noise in the chrominance components than some
competing models.
Color Series
The S2 has four settings associated with the "Color" soft key
under the rear-panel LCD data readout. STD is the default color saturation
setting, HIGH produces more saturated images, while ORG produces less
saturated ones. (Don't ask me why they call it "ORG" instead
of "LOW".) Finally, B/W renders
the image as black and white.
Tone Series The S2 also provides options for tonal variation, with values of STD
(default), HARD (high contrast), and ORG (lower contrast). The table below
shows some examples of the far-field shot with each of these options.
The differences are fairly subtle (as I think tonal adjustments should
be). The "ORG" version is a bit lighter overall - to get reduced
contrast with no change in exposure, it looks like you'd want to dial
in about -0.5EV of negative exposure compensation.
Sharpness series: The S2 provides three options for in-camera sharpening. Standard (STD),
increased sharpening (HARD), and no sharpening (OFF). The results of each
setting are shown below. (Click on any image to see the full-sized image
with that sharpening processing applied.) Most interesting (to my mind
anyway), is how well the unsharpened image responds to careful unsharp
masking in Photoshop. - See the fourth image below, and note the delicacy
with which fine details are rendered. (These were all taken from images
shot at the S2's 6 megapixel file size.)
I routinely shoot this series of images to show the field of view for
each camera, with the lens at full wide angle, at maximum telephoto, and
at full telephoto with the digital zoom enabled. Since the S2 accommodates
a wide range of Nikkor lenses, this test is somewhat meaningless for it.
Still, since I had the lens on the camera from the far field shots above
anyway, I shot the following with the 24-85mm f/3.5-45. G AF-S zoom that
I used for a lot of my test photos. (This lens seems like a very nice
match for the S2, IMHO, corresponding to a range of 36 - 127.5mm on a
35mm film camera.) Following are the results with this lens at its maximum
wide angle and telephoto settings.
This shot is often difficult for digicams, as the abundance of blue in
the image can trick white balance systems into producing a overly-warm
color balance. The S2 Pro did a very good job here, producing only very
slight casts with the Auto and Daylight
white balances. I once again chose the Manual
setting as the most accurate, but none of them were too far off. The strangest
thing here though, is that the Auto white balance image is noticeably
much brighter than the other two, even though the exposure settings were
the same for all three. Skin tones and the blue robe are pretty much spot
on. Resolution is very high, but frankly, the resolving power of the S2
is greater than the level of detail in the poster used as the target for
this shot.
Macro Shot
100mm Macro lens
Macro w/Flash
Not purely a function of the camera, but the Nikkor 100mm f/2.8 macro
lens gives some incredible shots!
This is another test I don't normally bother with for Digital SLRs, since
you can put whatever lens you'd like on them. I couldn't resist slapping
my 100mm f/2.8 Micro Nikkor on the S2 though. - This is really excellent
lens, the full capability of which isn't really revealed by the lighting
on this subject. Give it a serious look if you plan on much macro work!
(If you've looked at many of my test shots snapped with consumer and prosumer-level
digicams, you'll immediately be impressed by how sharp and distortion-free
the image from the Micro Nikkor is, corner to corner.)
Besides the 6 megapixel shots linked at right, I also shot and uploaded
some interpolated images, both with and without
flash.
Accurate color and saturation, excellent shadow detail.
Once again, the Auto white balance setting resulted in an oddly brighter
exposure, even though the aperture and shutter speed were identical for
all three shots. I chose the Manual White balance shot as the most accurate,
although it was still just slightly warm.
Detail in the shadows is excellent, and image noise is quite low. There's
a lot of highlight detail as well, but that's no surprise, given the overall
slight underexposure of all but the Auto white balance shot.
Color is very good, but I was a little surprised to see that the subtractive
primary colors (cyan, magenta, and yellow) were a little dull, particularly
the large yellow block of the MacBeth chart. Still, a very nice result.
ISO Series
ISO options on the S2 extend from 100 to 1600. Image noise is fairly low
and well-controlled at all ISO settings, although as noted in the main
review, I found more chroma noise in the S2's images than in those from
either the Nikon D100 or the Canon D60. Here's a sampling of shots taken
across the range of available ISO settings:
Color Series
The S2 has four settings associated with the "Color" soft key
under the rear-panel LCD data readout. STD is the default color saturation
setting, HIGH produces more saturated images, while ORG produces less
saturated ones. (Don't ask me why they call it "ORG" instead
of "LOW".) Finally, B/W renders
the image as black and white.
Tone Series The S2 also provides options for tonal variation, with values of STD
(default), HARD (high contrast), and ORG (lower contrast). The table below
shows some examples of the Davebox shot with each of these options. (Actually,
it's a little hard to see the differences between the three with this
target.
Absolutely amazing low-light capability, with very low noise.
With shutter speeds out to 30 seconds (plus Bulb exposures) and ISOs
as high as 1600, the S2 has no trouble capturing bright images under very
dim lighting. More surprising though, is how incredibly clean the
images are. The shots in the table below were snapped at the lowest light
level I commonly test to - 1/16 foot-candle, or 0.67 lux. At ISO 100 and
200, the noise levels are as low as many cameras manage to achieve in
broad daylight. There's also no sign of "hot pixel" noise, despite
the fact that the S2 doesn't appear to use any form of dark-frame subtraction.
(I suspect this is yet another consequence of Fuji's SuperCCD technology.)
Noise at ISO 1600 was much more pronounced, but still nothing near what
I would normally expect from that ISO and this low a light level. (Comparing
the S2 in this respect to competing 6 megapixel SLRs, it handily beats
the Nikon D100 in noise performance, but is itself trounced by the Canon
D60.)
The table below shows the best exposure I was able to obtain at 1/16
foot-candle (0.67 lux), the darkest I normally test at (since that's as
low as my light meter goes). Images in this table (like all of our sample
photos) are untouched, exactly as they came from the camera.
All
Shots at
1/16 fc
0.67 lux
ISO
100
ISO0200
ISO
400
ISO
800
ISO
1600
(only
one series, since no separate noise reduction option)
30 secs
F2.8
15 secs
F2.8
8 secs
F2.8
4 secs
F2.8
2 secs
F2.8
Love high ISO photography?
Hate noise? Check out Fred Miranda's ISO-R noise-reducing
actions for Photoshop. Incredible noise reduction, with
*no* loss of subject detail. (Pretty amazing, IMHO.) Check
it out!
Flash Range Test
Good intensity to 11-12 feet with the lens used. (f/4.5 aperture)
The S2's flash showed good intensity out to about 11-12 feet with the
lens used for this test. (The Nikkor 24-85mm f/3.4-4.5 G) This is actually
a bit better than Fuji's assigned guide number of 39 feet at ISO 100 -
Divide by the f/4.5 aperture used here, and that should equate to only
about 8.7 feet. I tested the flash at the default brightness level, with
no flash exposure adjustments. Below is the flash range series, with distances
from eight to 14 feet from the target.
Astonishing
performance, strong detail to 1600+ lines per picture height (!)
The S2 performed spectacularly on the "laboratory" resolution
test chart, although I have to hasten to point out that SuperCCD sensors
are optimized to produce their maximum resolution along the vertical and
horizontal axes, while more conventional CCDs with rectangular pixels
actually produce their highest resolution along axes tilted 45 degrees
from vertical.
Even allowing for the fact that the detail in this test target is oriented
favorably for the SuperCCD sensor, the resolution results are quite remarkable.
It's also interesting to note that the SuperCCD interpolation really comes
into its own with this target, with the 12 megapixel files resolving much
more detail than the 6 megapixel ones.
It must be noted that there are some artifacts that creep in,
with some regular patterns that look a little like stitching appearing
on the diagonal lines at the 1000 l/ph level. There are also artifacts
in the vertical and horizontal hyperbolic resolution wedges that appear
as patches of ghostly-looking blotchiness. Still, while there are artifacts
present, there's clearly what I'd call "strong detail" out to
about 1650 lines/picture height in the horizontal direction, and 1600
l/ph in the vertical one. "Extinction" of the target elements
doesn't occur until 1850-1900 l/ph. This is by far the highest resolution
I've seen to date (late July, 2002) on any digicam I've tested.
On consumer cameras, I normally comment here on the optical distortion
and artifacts. Since these are functions of the lens, not the imager,
it doesn't make sense to report on them for removable-lens SLRs. It is
worth noting though, the difference a really good lens can make: I shot
these photos with the Micro Nikkor 100mm f/2.8 Micro Nikkor, one of the
sharpest lenses Nikon makes, and one that also has about the least distortion.
As a result, chromatic aberration is exceptionally low, and the image
is sharp corner to corner. (I also shot with the lens stopped down to
f/8 for this shot, so I'd come as close as possible to testing the sensor
and image processing algorithms, not the lens itself.)
The S2's SLR optical viewfinder tested out at approximately 96 percent
frame accuracy (vertically), slightly better than Fuji's claimed 93%
vertical/95% horizontal figures. I generally prefer viewfinders to be
as accurate as possible, so the S2 does pretty well here. Flash illumination
is even throughout the frame. (No surprise, this was shot at a little
distance, not wide angle at all.)
Waiting for an S2? RitzCamera.com is taking orders for the
S2 on a first-come, first-served basis, and have already shipped
out the first units to happy IR readers! Get
in line now for a Fuji S2 Pro!