Fuji FinePix S1 ProA 3.5 megapixel "SuperCCD" gives superb color and amazing low-light capability in an under-$4,000 SLR digicam!<<Reference: Datasheet :(Previous) | (Next): Print-Friendly Review Version>> FinePix S1 Pro Sample ImagesReview First Posted: 8/1/2000 |
| We'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, we're posting the Thumber index so only those interested in the information need wade through it! ;) |
|
NOTE: The largest sizes of the files below apparently don't render properly in some versions of Netscape and Internet Explorer. If you see images that appear stretched horizontally, you'll need to download them to your hard drive and open them with an image-editing software package. NOTE TOO:
We also saw some odd "jaggie" artifacts in browsers
(IE only) in the DaveBox images, at all sizes, that were definitely
*not* in the images themselves. If you see odd artifacts, you
should again try downloading the images to your local hard drive
and opening them in an image-editing program. |
 |
Outdoor portrait: (1475k) (Wow, great color!) The extreme tonal range of this image makes it a tough shot for many digicams, which is precisely why we set it up this way. The object is to hold highlight and shadow detail without producing a "flat" picture with muddy colors and the S1 did a reasonably good job. We shot this image in the automatic ( 1466k), daylight (1447k) and manual (1443k) white balance modes, selecting daylight (or Sunny, as Fujifilm refers to it) for our main series. The manual setting resulted in a rather warm image, while the daylight and automatic settings appeared very similar. In the end, we chose the daylight setting because it had a slightly better white value. Overall color here is excellent. The skin tones in this shot are just the slightest bit ruddy (reddish) to our eye, but the colors in the flowers are excellent, and there are none of the odd color breaks we've seen in at least one other high-end SLR digicam(!) The S1 had little problem with the usually difficult blues of the flowers and the model's pants: Some cameras render these with a rather purplish cast, but the hues in the S1's image are virtually spot-on. The image appears very crisp throughout, with great detail in the shadow areas, and barely any visible noise. (These images were shot with the S1's medium resolution setting, corresponding most closely to the direct CCD resolution, rather than the interpolated settings. See our shots of the House, Davebox, and Far-Field tests below for some examples of the 6.1 megapixel file size.) Our main image was taken with a +0.7 EV exposure adjustment to get the best exposure on the face without losing too much detail in the highlight areas. We had to compromise a little here, as the highlight areas are already somewhat blown out, but lowering the exposure compensation made the skin tones and shadow areas a little dark. (This actually would have been a good shot to try the S1's reduced-contrast tonal setting, but we didn't have time this day before the sun hid behind a cloud. - Afternoon clouds are an environmental hazard in Atlanta in the summertime!) We could perhaps ask for the skin tones to be tweaked a little, but overall this is some of the most appealing color we've yet seen on this shot!) The table below shows the results of a range of exposure settings from zero to +1.7 EV in the daylight white balance mode. Exposure Compensation Settings:
|
||||||||||||||||||||||||||||||||||||||||||||
 |
Closer portrait: (1466k) The S1 also did a beautiful job on this closer portrait shot. On an objective basis, the skin tones are again a trifle on the "hot" side, but the overall effect is exceptionally appealing. (We suspect the "original" color setting would bring the skin tones into a better range, although at some cost in color saturation everywhere.) We again shot with the daylight white balance setting for our exposure series, with our main shot requiring a +0.7 EV adjustment. (Interesting, as this shot generally requires less exposure adjustment than our longer portrait above, but the S1 seemed to want the same compensation setting for both.) We also snapped images with the automatic (1432k) and manual (1464k) white balance settings, achieving similar results as in the larger Outdoor Portrait. Resolution and detail look tack-sharp in this shot, although we did notice very faint color artifacts visible in the model's hair, as first noted by Phil Askey of www.dpreview.com. (His model's black hair made them much more apparent than they are here however.) Noise remains quite low in the shadow areas. The table below shows the results of a range of exposure settings from zero to +1.7 EV. Exposure Compensation Settings:
|
||||||||||||||||||||||||||||||||||||||||||||
|
Indoor Portrait, Flash:
(1355k) (Late-breaking word on the indoor flash test:
It turns out the onboard flash is more powerful than we'd realized,
and in fact has a fairly long minimum working distance,
due to the high ISO rating of the S1! - It appears we were simply
trying to work with too large an aperture on this shot. - We'll
reshoot this shortly, using a more reasonable aperture setting!) |
||||||||||||||||||||||||||||||||||||||||||||
|
|
Indoor portrait, no flash: (1344k) This shot is always a very tough test of a camera's white balance capability, given the strong yellowish cast of the household incandescent bulbs used for the lighting, and the S1's automatic white balance system had a fair bit of trouble here. We tested the automatic (1352k), incandescent (1370k) and manual (1344k) white balance settings, choosing manual for our main series. The automatic setting produced extremely warm results, with orange tones, as did the incandescent setting. Manual white balance produced the most accurate results, although the skin tones still look a little pink. The flowers are also muted, with the blue flowers appearing purplish and the red flower looking also a bit pink. Interestingly, the S1 reported "Out of Range" when we attempted the manual white balance adjustment for this light source, telling us that the extreme warm tones were more than it felt it could properly compensate for. This was the first time we saw a message like this from a digicam with a manual white balance setting, and it was very welcome: We'd like to see more digicams tell us when they thought we were asking for something beyond their capability to produce. For our main shot (1344k), we chose a +1.3 EV adjustment. We also tested the camera's other ISO settings, shooting at 400 (1349k), 800 (1378k) and 1600 (1411k) (all of which have a +1.0 EV exposure adjustment). As you might expect, the noise level increases with the ISO setting. However, at the 1600 setting, which is the noisiest, the noise pattern is so tight that we didn't find it terribly distracting. We also noticed that the 1600 ISO setting produced a rather "milky" look in the shadow areas, apparently as a result of the image noise setting a higher "floor" on image brightness that was most apparent in the shaddows. Overall, we found the ISO 400 & 800 images very usable, but the ISO 1600 shot less so. In the ISO 1600 image, the noise level in the blue channel took a substantial jump upward relative to the lower ISO settings, some odd blue highlights appeared in the deep shadows, and the noise took on a strongly patterned streakiness along the horizontal axis. Undoubtedly better than flat-out missing an important shot, but we don't think you'll want to rely upon the S1 Pro for routine shooting at ISO 1600. The table below shows a range of exposure adjustments from zero to +1.7 EV using the manual white balance setting. Exposure Compensation Settings:
|
||||||||||||||||||||||||||||||||||||||||||||
 |
House shot: (2581k) NOTE that this is the "new" house shot, a much higher-resolution poster than we first used in our tests. To compare the image of the S1 with previously tested cameras, here's a shot of the original house poster in the manual (1513k) white balance setting. We shot this image with the automatic (1574k), daylight (1590k), cloudy (1580k), fluorescent 1 (1570k), fluorescent 2 (1581k), fluorescent 3 (1508k), incandescent (1515k) and manual (1477k) white balance settings, ultimately deciding on the manual setting for our main series. The cloudy setting produced a very warm image, while the daylight setting appeared only slightly warm. Automatic resulted in a nearly accurate image, although the color cast was still just a hair warm. The fluorescent settings all produced interesting color casts, with the fluorescent 1 setting showing a very warm image with an orange tint. The fluorescent 2 setting appears a touch magenta, and the fluorescent 3 setting appears very magenta. The incandescent setting resulted in a predominantly blue image. The fine details in the image look very nice, and sharpness is very good across the board. The shingles show just a moderate amount of noise, although that could be just the texture of the shingles in the poster. Shadow detail is exceptionally good, with very little noise. Color is excellent, very closely matching the original target. The table below shows the full range of resolution and quality settings for the S1, using the Aperture Priority exposure mode with an aperture setting of f/5.6 on our 105mm f/2.8 Micro Nikkor lens. With this target, we shot a full series of image sizes, including the 6.1 megapixel interpolated size. Comparing the fine details of the 6.1 megapixel images with those from an "2304" size image, we felt we did in fact see an ever so slight increase in detail in the 6.1 megapixel samples. The difference was really small though, leading us to recommend shooting at the 2304 x 1536 pixel size for routine work.
Sharpness Series We also shot a series of images with the camera's variable sharpness mode, which offers options of Off, Standard and Hard. The Off setting produced a very soft image, with no in-camera sharpening at all. Standard produced moderately sharp results, with Hard resulting in a very sharp image (with a slightly higher contrast). Even the Hard setting was fairly restrained in its effect though, as there was little evidence of obvious "halos" around objects. The Soft setting is welcome because it allows post-capture processing in Photoshop(tm) or other image manipulation software, where the sharpening effect can be more closely matched to the particular subject matter involved.
Color Series The S1's color adjustments also work very well. The Black and White setting produces a very even, monochromatic image. Original color appears slightly dull, while Standard color seems a little more saturated. High color increases the saturation even more, but manages not to overdo it. Other manufacturers should take note of this color-saturation implementation: We're starting to see this option appear in more and more digital cameras at the consumer level, but Fuji's implementation in the S1 is excellent. The variation between the "Original", "Standard", and "High" color settings is about what we'd expect to see in a range of film emulsions. The subtlety of the variation is just right to provide a range of viable color options that offer useful creative control. Very well done!
Tone Series The S1's tonal adjustments work very subtly with this test target. The Original tone setting produces a just slightly dim image, without too much saturation. Standard increases the tonal intensities a little further, and Hard tone deepens the tone and seemingly, the image contrast just slightly as well. We encourage the inclusion of tonal controls on digicams, in the same way that we do the color saturation options as discussed above. The tonal adjustments on the S1 are less successful than it's color adjustments are though. "Original" tone seems to primarily affect the shadows, which isn't where digicams run into trouble anyway. (Highlights are where detail is irrevocably lost: Shadows can usually be brought up in an image-editing program if the noise isn't too severe.) We'd like to see a tonal adjustment that affected both ends of the tonal range, bringing the highlights down and the shadows up simultaneously. The S1's Original tone setting could be helpful on high-contrast subjects, but you'll need to remember to bring down the overall exposure level a little to be able to see its effect in the highlights...
|
||||||||||||||||||||||||||||||||||||||||||||
 |
Far-Field Test (2551k) 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. For this test, we shot with the automatic (1467k), daylight (1482k), cloudy (1481k), fluorescent 1 (1432k), fluorescent 2 (1481k), fluorescent 3 (1434k), incandescent (1487k) and manual (1496k) white balance modes, choosing automatic as the most accurate. The automatic and daylight setting produced similar results as the manual setting, but both were slightly warmer, and cloudy showed a very warm cast. The three fluorescent settings performed as with the House poster, producing distinct color casts, while the incandescent setting resulted in a bluish image. This is the strongest test of detail of any that we do, and the S1 performs well. Sharpness looks great throughout (maybe just a hint soft, but see our note below about the lens used), and the fine details in the leaves and branches above the house are reasonably well preserved. The strong highlight on the central bay window of the house is very hard for cameras to hold detail in, and the S1 does an excellent job. Color balance looks good throughout and the shingles and shadow areas show a very minimal amount of noise. This shot called for a somewhat wide-angle focal length, which meant we couldn't use our ultra-sharp 105mm f/2.8 Micro Nikkor. We ended up using a Tamron 28-300mm "vacation lens", which was quite a bit softer than the 105mm. The effect impact of the lesser lens is noticeable in the slightly softer images here, which are more representative of the lens than the camera. (This underscores the need to use the highest-quality lenses to extract the best performance from high-end digicams.) The table below shows the full resolution and quality series in Aperture Priority mode with a lens aperture setting of f/5.6.
Sharpness Series We again shot with the S1's sharpness settings, which did a very nice job of increasing and decreasing the overall sharpness without getting too extreme.
Tone Series The tonal adjustments again performed well, applying subtle changes to each image, although we would have preferred it if the Original setting had pulled in both ends of the tonal range (shadows and highlights equally).
Color Series We also noticed nice color adjustments here, with the monochromatic image producing very even results and the High Color setting producing a nice level of saturation.
|
||||||||||||||||||||||||||||||||||||||||||||
|
|
Lens Zoom Range We've received a number of requests from readers to take shots showing the lens focal length range of those cameras with zoom lenses. Thus, we're happy to present you here with the following series of shots. Since the S1 doesn't come with any lenses, this test is a little meaningless (you can put anything you like on it from a ~15mm ultra wide angle to an ~1000 mm ultra telephoto. Nonetheless, since we had the 28-300mm Tamron on the camera for the Far Field shot above, we went ahead and shot examples at both 28 and 300 mm. Resolution looks reasonably sharp with the 28 mm, wide angle shot, shows the softness this lens is noted for at the 300 mm setting. (The conventional wisdom seems to be that this lens is reasonably (if not tack-) sharp at focal lengths from 28 to 200 mm, then softens considerably as you approach the 300mm limit.
|
||||||||||||||||||||||||||||||||||||||||||||
 |
Musicians Poster (2551k) We shot this test with the automatic (156k), daylight (163k), cloudy (164k), fluorescent 1 (159k), fluorescent 2 (162k), fluorescent 3 (157k), incandescent (155k) and manual (159k) white balance settings, choosing manual as the most accurate setting overall. The automatic and cloudy settings produced very similar results, both just slightly warm. The daylight setting appeared almost identical to the manual setting, although it was just a touch warmer. The three fluorescent settings again produced a range of color casts, with the first setting resulting in a very orange, slightly sepia image. The second setting produced magenta results, with the third fluorescent setting resulting in a very purplish image. Finally, the incandescent setting resulted in a very blue color cast. (It seems that the S1's automatic white balance system falls prey to the typical problem with this sometimes difficult target: Many digicams are tricked by the significant amount of blue in the image and overcompensate for it, making the image too warm overall, which is exactly what the S1's automatic setting tried to do.) In our manual white balance version, the skin tones look quite accurate and the telltale blue robe also looks about right, although slightly dark (this is a hard blue for many digicams to interpret properly, but the S1 does better than most). Resolution is also good, with nice detail in the bird wings and silver threads on the blue robe, as well as in the flower garland. (Really, this poster is only useful for color comparison these days, since the poster itself is beginning to limit resolution performance.) A moderate level of noise exists throughout the image, most noticeably in the background (some of which may be from the actual poster). Below is our standard resolution and quality series in the manual white balance setting, all shot in the Aperture Priority mode with a lens aperture setting of f/5.6. We also included a sharpness, tone and color series, which performed nicely, as with our other test shots. Resolution/Quality series
Sharpness Series
Tone Series
Color Series
|
||||||||||||||||||||||||||||||||||||||||||||
 |
Macro Shot (913k)
(THIS SHOT COMING SOON) Well, as in the wide/tele shot above, you're free to attach *any* lens onto the S1 that you'd like (including a microscope via an adapter!), so this test is a little meaningless here. We did though, use it as an excuse to show off the incredible performance of our 105 F/2.8 Micro Nikkor again. (Having paid as much as we did for that lens, and using it as seldom as we do, it's good to have an excuse to justify it as an investment. ;-) |
||||||||||||||||||||||||||||||||||||||||||||
 |
"Davebox" Test Target (2378k) We shot this test target with the automatic (159k), daylight (159k), cloudy (159k), fluorescent 1 (162k), fluorescent 2 (160k), fluorescent 3 (159k), incandescent (161k) and manual (161k) white balance settings, again choosing manual for its accuracy. The automatic setting produced similar results to the manual setting, although with just a slightly warm color cast. Daylight also produced slightly warm results, while cloudy resulted in a very warm image. The three fluorescent settings produced very heavy color casts, with the first setting appearing almost sepia. The second fluorescent setting resulted in a slightly magenta image, and the third fluorescent setting produced almost purplish results. Color balance with the manual white balance setting looks good overall, the large cyan, magenta and yellow color blocks on the left side of the target look pretty accurate, although maybe just the slightest bit under-saturated. The S1 does a nice job of separating the red and magenta color blocks on the middle, horizontal color chart, although the two hues appear a little flat (many digicams try to blend the two colors into one). Overall, we feel this is some of the most accurate color we've seen from a digicam on this target. Surprisingly, S1 slightly overexposed this shot, with the subtle tonal variations of the Q60 chart are only visible up to the "D" range. Shadow detail is quite good, but highlights are lost in the white gauze as well as in the Q60 target. (We apparently should have used a slight negative exposure compensation here.) Noise levels are quite good throughout the image. Below is our standard resolution and quality series, shot with the 105mm Micro Nikkor in Aperture Priority mode with a lens aperture of f/5.6.
Sharpness Series
Tone Series The S1 again performed well in this series, significantly brightening the color blocks with the Hard Tone setting.
Color Series As with the Tone Series above, the color blocks on the test target appeared much improved and brighter with the High Color setting.
|
||||||||||||||||||||||||||||||||||||||||||||
 |
Low-Light Tests The S1 Pro proved to be an incredible low-light performer, able to easily capture images all the way down to the 1/16 footcandle (about 0.13 lux) limit of our low-light test. We again shot with our 105mm f/2.8 Micro Nikkor lens, with the aperture set wide open. This should be a pretty good representation of what the S1 can do in normal applications, as f/2.8 is fairly standard for a "fast" professional lens, available even in zoom and moderate telephoto formats. Looking at our test images, several things immediately stand out: First and foremost, they are remarkably "clean", with very low noise levels, even at the lowest light levels we test at (1/16 of a foot candle, or about 0.13 lux). Not only is the noise level unusually low, but there are very few "stuck" pixels in evidence. The excellent results in terms of stuck pixels are doubtless helped by the fact that the S1's minimum ISO of 320 is obviously a very solid, conservative rating: Even at the lowest light level and ISO 320, our exposure times were only on the order of 1 second. (And the camera goes all the way out to 30 second exposure times!) Longer exposure times would doubtless reveal more "stuck" pixels, but based on our tests, that should only be an issue for astrophotographers, or people specializing in available-light coal mine photograph. ;-) The other thing that we found absolutely remarkable about the S1's low light performance was the incredibly rich, accurate color it could produce even the lowest light levels. To be sure, we perhaps "cheated" a bit, running the camera with the "custom" white balance setting, having set the white point with the barn doors on the photoflood we use for these shots set wide open. Still, this could be a very legitimate trick for location shooting as well: Walk over directly under the street light and shoot a white reference, then fire away in the darker areas with abandon. (Here's another thought: How about just slipping a piece of light diffusion gel over the lens and pointing it directly at the light source to get the white-point reading? That should let you get good white values even with very dim light sources.) One important note in low light tests is the camera temperature at the time of testing: CCDs are quite sensitive to temperature, as their noise levels double about every 6-8 degrees C. Prior to shooting these tests, we had the S1 in a fairly chilly air-conditioned room for several hours, and so would estimate the camera body was at a temperature of about 65 degrees F. (18 degrees C) This is the condition we try to set all the cameras in prior to testing, so our own results should be consistent across cameras. Keep this in mind though, if you see singificantly better or worse noise performance on other websites for the same cameras. Overall, the S1's low light performance is easily the best we've seen in any digicam we've tested to date. (We haven't tested the extended-sensitivity Kodak 620x model, and fully expect that would do better, but then again, that's a camera that costs nearly three times as much as the S1, and has only two-thirds the resolution.) At higher ISO values, the S1 still did very well in the noise department, only showing noise that we'd deem objectionable at its highest ISO rating of 1600. Again, the S1's noise levels in these tests was better than any other camera we've tested to date. (August, 2000)
|
||||||||||||||||||||||||||||||||||||||||||||
|
|
Flash Range Test (New) (This test was added in August 1999, so cameras tested before that time won't have comparison pictures available. As we go forward though, all the new models will have similar tests available). Fuji gives the S1 Pro's built in flash a guide number of 15 meters at ISO 100. In practice, since the lowest ISO rating available is 320, the useful guide number is a fair bit higher. In the manual, Fuji shows a range of 4.3 to 20.7 feet (1.3 - 6.3 meters) at ISO 320 and f/5.6. This is quite a bit beyond the test range of our studio, and in our testing, we found the S1's built-in flash to be quite effective all the way out to 14 feet without getting too dark. (The fairly long minimum usable distance for the flash might also explain the overexposure we initially experienced in our indoor portrait test.) The table below shows results obtained at a range of distances from seven to 14 feet.
|
||||||||||||||||||||||||||||||||||||||||||||
 |
ISO-12233 (WG-18) Resolution Test (2612k) This is likely to be one of the most contentious areas of our review of the S1, as we feel the results really support our assertion that the in-camera interpolation adds a slight but not dramatic amount of resolution over that seen in the uninterpolated 2304 x 1536 image size. In the test images below, you'll see that there is a noticeable decrease in aliasing in the images shot in the 3040 x 2016 resolution mode at certain ranges of spatial frequencies. In what we've been calling the "uninterpolated" resolution mode (2304 x 1536 pixels), the image is very sharp and clear, with visible detail out to nearly 900 lines per picture height horizontally and 800 lines vertically, but color aliasing is quite apparent beginning at 800 lines horizontally and perhaps 750 lines vertically. By contrast, the "interpolated" 3040 x 2016 resolution mode shows greatly reduced aliasing in the horizontal direction, with reasonably clean detail all the way out to 100 lines per picture height, although there is a patch of color aliasing visible around the 900 line mark. Vertically, the effect of the larger file size is less dramatic, with only a slight improvement in aliasing between 700 and 800 lines, but virtually none at all beyond that. So, overall we'd have to say that there is a noticeable improvement in resolvable detail between the 2304 and 3040 image sizes, but it isn't nearly as dramatic as the raw increase in file size would suggest. Bottom line, there does appear to be some improved interpolation achievable with the SuperCCD geometry, but not as large as Fuji would claim. Our recommendation? Shoot in the 2304 pixel mode for all but the most critical shots, and you'll save memory space and probably never notice the difference.
Sharpness Series
Tone Series
|
||||||||||||||||||||||||||||||||||||||||||||
|
|
Viewfinder Accuracy/Flash Uniformity Because we shot this test with a 105 mm lens, instead of a zoom lens, there are no wide angle or telephoto examples here. This really reflects the fact that the S1's viewfinder accuracy is much more a function of the camera than any lens that's attached to it. We did find that the framing in the viewfinder was somewhat sensitive to eye position, as we could "peer around" the edges of the framing mask in the viewfinder if we moved our eye back and forth or up and down. We first took a shot of what we called the "viewable area," in that we lined up the shot based on what we could initially see in the viewfinder without any eye movement. Our viewable area shots resulted in about 94 percent frame accuracy (448k) at all three image sizes. Next, we lined up what we called the "extended viewable area," which means that we utilized all angles of view as a result of moving our eye around to line up the shot. Our extended viewable shots resulted in 97 percent frame accuracy (439k) for all three image sizes. For this test, we shot in the Aperture Priority mode, with an aperture setting of f/5.6 We didn't measure optical distortion on the S1, since its body design accepts a range of lenses and distortion will vary with the lens. (Looking at the results from our standard 105mm f/2.8 Micro Nikkor though, illustrates at least one difference between a low-end zoom lens and high-end professional single focal length design: There's really no visible distortion, either in the form of barrel/pincushion distortion, or as chromatic aberration. |
||||||||||||||||||||||||||||||||||||||||||||
Follow Imaging Resource: