Fuji FinePix S3 ProFuji updates their digital SLR with a 12.34-megapixel Super CCD SR II, for amazing tonal range.<<Included Software :(Previous) | (Next): Reference: Datasheet>> Page 13:Test Results & ConclusionReview First Posted: 9/7/2005 |
Test Results
In keeping with my standard test policy, the comments given here summarize only my key findings. For full details on each of the test images, see the S3 Pro's "pictures" page.
For a look at some more pictorial photos from this camera, check out our Fujifilm S3 Pro Photo Gallery.
For the real nitty-gritty image-analysis details, read my S3 Pro Imatest Results page.
Not sure which camera to buy? Let your eyes be the ultimate judge! Visit our Comparometer(tm) to compare images from the Fujifilm S3 Pro with those from other cameras you may be considering. The proof is in the pictures, so let your own eyes decide which you like best!
- Color: Excellent, accurate color. Overall, the S3 Pro produced excellent color and saturation throughout my testing, with pleasing, natural results in most cases. Analytically speaking, it rendered colors more accurately than the majority of cameras I test. (This more accurate rendering may be perceived as flatter-looking color relative to the majority of cameras with routinely oversaturate color slightly, but if you want a more saturated "look" to your photos, it's easy enough to select the more saturated film-simulation mode.) Skin tones were about right (if just slightly reddish in the "Sunlit" Portrait), though the always-difficult blue flowers of the bouquet turned out very close to their real-life color. I typically chose the Manual white balance setting as the most accurate, though the Auto setting produced the best skin tones on the Musicians shot (despite slightly warm color elsewhere). The S3 Pro had a little trouble with household incandescent lighting in our "Indoor Portrait" shot, but its Manual setting performed very well there. All in all, very accurate, pleasing color.
- Exposure: Accurate exposure, about average among the d-SLRs I've tested. The S3 Pro handled my test lighting quite well, accurately exposing most shots. It underexposed the very high-key outdoor portrait shot slightly at the default setting, but a small amount of positive exposure compensation did the trick and captured reasonably bright midtones. Other shots that typically call for exposure compensation seemed to need about the typical amounts, and shots that don't usually need exposure adjustment also did not with the S3 Pro.
- Dynamic Range: Absolutely amazing dynamic range, thanks to Fujifilm's "SR" sensor technology. (And helped along by Adobe Camera Raw 2.) I don't usually include a separate paragraph in these Test Results summaries for dynamic range, but the S3 Pro's performance in this area is so extraordinary that it really calls for separate discussion. For a full exploration of the camera's capabilities in this area, read my analysis of its performance on my "Sunlit Portrait" test on the S3 Pro image analysis page. Here's a brief summary of my findings though:
In a word: Incredible. (If you're a wedding photographer tired of blown-out highlights in bridal dresses, don't even think twice: This is the camera you need.) Taking JPEGs straight from the camera in its "Wide 2" dynamic range mode, results are numerically about 2/3 of an f-stop better than any other camera I've tested. Visually though, the results look to be more like a full stop or more. The really dramatic results come though, when you process its RAF-format RAW files through the Camera Raw 2 plug-in in Adobe Photoshop(tm). Under that condition, numeric results improve to 1 2/3 to 2 full f-stops over the rest of the field, and visual results verge on mind-boggling. Highlights that "should" be blown out by three f-stops or more show full detail, and colors are rendered entirely naturally. (Fuji's own Hyper-Utility 2 works can do slightly better than the in-camera processing, but doesn't begin to approach what Camera Raw 2 can do.)
Here's how the S3 Pro stacks up numerically against other cameras I've tested. (Numbers from Imatest.)
Dynamic Range (in f-stops) vs Image Quality
(At camera's minimum ISO)Camera Model Allowed noise, f-stops
(Image quality level)1.0
(Low)0.5
(Medium)0.25
(Med-High)0.1
(High)Fujifilm S3 Pro
(Via Adobe Camera Raw 2)12.1 11.7 10.7 9.0 Fujifilm S3 Pro -- 9.9 9.4 7.94 Nikon D50 10.7 9.93 8.70 7.36 Canon EOS 20D 10.3 9.66 8.85 7.29 Canon Digital Rebel XT 10.3 9.51 8.61 7.11 Olympus EVOLT 10.8 9.26 8.48 7.07 Canon Digital Rebel 10.1 9.11 8.47 6.97 Pentax *istDs 10.2 10 8.87 6.9 Nikon D2x -- 8.93 7.75 6.43 Nikon D70S 9.84 8.69 7.46 5.85 Nikon D70 9.81 8.76 7.58 5.84
Numbers are all well and good, but to really understand the magic of the S3 Pro, you need to look at some really pathological examples of "problem" photos. Here's one such. All three images were extracted from the same RAF file, a shot captured at +2.0 EV of exposure compensation, a good stop to stop and a half overexposed for the midtones, and at least three stops for the brightest highlights. The image on the left was extracted without dynamic range expansion (the "S" pixels were ignored). This is perhaps a bit worse than the typical d-SLR would handle this degree of overexposure. The middle image represents the best I was able to do with Hyper-Utility 2, using the full 400% dynamic range expansion setting. The final image on the right shows the results from Adobe Camera Raw 2, importing into Photoshop CS2. (Yes, it really is from the same RAW file.)
This is of course an extreme, pathological case: In practice, you hopefully would never have to deal with anything this extreme. - But it's nice to know that the capability is there, in case you ever need it.
There is one caveat though: When dealing with an extreme tonal correction like this, Camera Raw 2 does seem to generate some image noise over a fairly narrow tonal band. - The shot above shows a 1:1 crop from a Stouffer 4110 density target, and clearly shows the elevated noise in steps 11 and 12. I never saw evidence of this noise in "natural" subjects, but it consistently appeared in the dynamic range target.
Bottom line though, the S3 Pro's dynamic range under the most restrictive image-quality conditions extends to a full 9 f-stops, and to nearly 11 f-stops if you're willing to accept about a quarter f-stop of image noise in the shadows. Remarkable!
- Resolution/Sharpness: Better than average resolution for a camera with ~6 million effective pixels. The S3 Pro performed quite well on the "laboratory" resolution test chart. While it technically has 12.34 million "effective" photo sensors in its array, these are grouped together into half that many actual pixel locations, so the camera's resolution could be expected to be about that of a 6.17 megapixel model. In practice, Fujifilm's SuperCCD technology does deliver somewhat higher resolution along the horizontal and vertical axes than that obtained with conventional CCD arrays, and the S3 Pro showed this in its performance with the ISO-12233 resolution target images. It started showing artifacts in the test patterns at resolutions as low as 1,200 line widths per picture height in both directions (and there was a fair bit of color artifacting starting at about that point), but I found "strong detail" out to at least 1,800 lines, and I could still make out the lines at the 2,000 lw/ph point.
- Closeups: No data, macro performance will depend entirely on the lens in use.
- Night Shots: Excellent low-light performance. Good color and exposure, with low image noise, even at the darkest light levels of this test. Excellent low-light focusing ability too. The S3 Pro produced clear, bright, usable images down to the 1/16 foot-candle (0.67 lux) limit of my test, with good color at all of the ISO settings I tested. A warm color cast appeared at the lower exposures, but overall color was still pretty good. Noise remained low in most shots, and was lower than I'd normally expect, even at the 800 and 1,600 ISO settings. The camera's autofocus system was also more sensitive than most, able to achieve focus at light levels well below the minimum shown here, even with the AF-assist light disabled. (Note though, that it's critical to have the camera mounted on a stable tripod when trying to autofocus at very low light levels.)
- Viewfinder Accuracy: Average viewfinder accuracy for a d-SLR. The S3 Pro's digital SLR viewfinder was fairly accurate, showing about 94 percent of the final image area. Given that I like digital SLRs to be as close to 100 percent accuracy as possible, the S3 Pro has a some room for improvement here, but many competing models have only 95% accurate viewfinders also.
- Optical Distortion: No data, optical distortion will depend on the lens in use.
- Battery Life: Rather limited battery life for a d-SLR. A little finicky about specific NiMH cells? The S3 Pro uses a set of four AA-style batteries to power all camera operations. This has the advantage of low cost and ready availability of spares, but also results in rather limited battery life when compared to most other d-SLRs on the market. With modern high-capacity NiMH cells having true capacities of 2100 mAh, worst-case run time should stretch to about 2.75 hours. This is shorter than that of most other d-SLRs, but this is of course assuming continuous operation (without letting the auto-shutdown timer put the camera to sleep). Also, the use of conventional AA cells means it'd be easy and quite inexpensive to keep several sets of NiMH cells topped off and on hand at all times. One side note though: While I didn't have any way to test the S3 Pro's cutoff voltage directly (the battery voltage below which it refused to work anymore), my sense was that it was a bit finicky about the NiMH batteries I fed it. - With some brands and models seemed to shut down much earlier than with others, even though the cells involved had roughly similar actual capacities. Be sure to buy name-brand, high-capacity NiMH cells for use with the S3 Pro, and be sure to get a good charger as well.
- Print Quality: Excellent print quality. Slightly soft at 13x19, but very usable for wall display. High ISO images also usable for unusually large prints. Testing hundreds of digital cameras, we've found that you can only tell just so much about a camera's image quality by viewing its images on-screen. Ultimately, there's no substitute for printing a lot of images and examining them closely. For this reason, we now routinely print sample images from the cameras we test on our Canon i9900 studio printer, and on the Canon iP5000 here in the office. (See the Canon i9900 review for details on that model.) In the case of the S3 Pro, the resolution capabilities of its SuperCCD technology translated into 13x19 inch prints that looked a fair bit better than I'd normally expect to see from what boils down to a 6.17 megapixel camera. While close inspection revealed some softness in 13x19s printed from the S3 Pro, I'd have no qualms whatsoever about selling them to clients were I a commercial, wedding, or portrait photographer. Print quality at high ISO was also very good. Once again looking at 13x19 inch prints, photos shot at ISO 1600 looked somewhat softer, but colored noise artifacts were really minimal. Whether or not prints this big at ISO 1600 would be suitable for client presentation would likely depend a fair bit on the subject involved, but for available-light work at a wedding or other social/entertainment event, I suspect they'd be entirely acceptable. All in all, really excellent print quality, particularly in light of the relatively modest sensor resolution.
Conclusion
Pro: | Con: |
---|---|
|
|
Fuji's original S1 Pro did well partly due to its beautiful color rendering, and partly due to a price that was thousands of dollars less than other D-SLRs at the time of its introduction. The following S2 model genuinely amazed me with its resolution and how delicately it rendered fine detail. I really liked its user interface design, with the small rear-panel data readout and "soft buttons" below it, and was very impressed overall with the camera's capabilities. The S2 offered very nice color rendering as well.
Free Photo Lessons
Learn how to take stunning photos with simple pro lighting tips, in our free Photo School area!
Now comes the S3 Pro, a camera that takes a somewhat different path, trading off image resolution to achieve dramatically greater dynamic range. Its body size and design will prompt many to compare it with lower-end (and much cheaper) "prosumer" d-SLRs, but that's not really a fair comparison, given its amazing imaging capabilities. When it comes to preserving difficult highlights without disturbing the tonality or color rendition of the rest of the image, the Fujifilm S3 Pro is unquestionably without equal anywhere in the digital camera market. If you're a wedding photographer, perpetually challenged by the need to capture white-on-white detail under difficult lighting conditions, you shouldn't even think twice. -- This is the camera you need, it'll save you thousands of dollars in lost shots and missed opportunities, most likely in the first few jobs alone. Portrait and commercial photographers should find a lot of benefit in the S3 Pro's exceptional dynamic range as well, as it could help relax lighting requirements (and therefore setup times) quite a bit. The one minor negative note about the S3 Pro's amazing dynamic range performance is that you really need to work in RAW mode and process its files through Adobe's Camera Raw 2 Photoshop(tm) plugin to get the absolute most out of it. (That said though, JPEGs straight from the camera do noticeably better with difficult highlights than even images processed from RAW files from competing d-SLRs.)
Other camera parameters are fairly middle-of the road, with less than spectacular shot to shot speeds, and mediocre battery life. It's easy to forgive these minor shortcomings though, in exchange for the incredible dynamic range the S3 offers.
Bottom line, Fujifilm appears to have achieved exactly what they set out to do with the S3 Pro, delivering a camera with dynamic range that equals or exceeds that of film, with excellent color and tonal rendition in the bargain. Highly recommended, especially if your work involves a lot of tricky highlight detail. (An easy Dave's Pick.)
Reader Comments! --> Visit our discussion forum for the Fuji FinePix S3 Pro!
Follow Imaging Resource: