Color (8.39)
Since color preference is largely subjective, our testing measures the accuracy of the reproduced tones by contrasting the camera’s rendered tones with the corresponding ideal. We do this by recording a series of exposures of an industry standard GretagMacbeth color chart and importing the results into Imatest Imaging Software. For cameras that offer multiple color modes as well as tonal and saturation adjustments, a series of tests are conducted at the various available settings, with the best results reported. Again, these are the most accurate color results achieved by the camera, not necessary the most striking or appealing to the user. It is far easier to embellish a “realistic” representation of a scene in postproduction than it is to try to revert it back to its actual appearance. Cameras that score high on these color tests will give photographers an opportunity to capture a more natural scene and then work with the file; it is not a gauge of how well the camera might be tailored to any given individual’s taste.
Below is a modified GretagMacbeth color chart, displaying the Fujifilm S3’s reproduced tones. The chart has been modified by Imatest Imaging Software to display the variance between the camera’s produced colors and the corresponding ideal from the original chart. For each color tile, the outer square contains the camera’s rendered colors, while the vertical rectangle is the ideal. The small inner square is a color-corrected version of the camera’s produced tones, approximating editing with a software application.
The chart below displays the S3’s color results in a more quantitative, linear manner. The circles represent the tonal ideal, while the corresponding squares are the camera’s produced tones. The line bridging the two shapes indicates the degree of error; the longer the line, the less accurate the camera’s reproduction of that particular tone.
The S3’s color performance was acceptable, earning an 8.39 overall color score; however, overall performance was not quite as strong as might be expected. While users can manually adjust color saturation in 3 levels, the S3’s rendering of most red and green tones (# 15 Red, #9 Moderate Red, #16 Yellow, #14 Green, #11 Yellow Green) were far from accurate. Blue channels were shifted, while red values were over-saturated and green values under-saturated. The inconsistency in the S3’s rendering makes it difficult for users to properly calibrate tones in-camera, making it necessary to adjust individual tones post-capture. For some users, this will not be a problem; however, those looking for a usable JPEG straight off the card might see this as a sizable roadblock.
Most of the inaccurate tones produced by the S3 appear to be deliberate adjustments by Fujifilm to add vibrancy and brilliance to the image. While many might prefer the aesthetics of the S3’s reproduction, those seeking flawless color representation will have to invest time some editing the file.
Still Life Scene
Below is a shot of our vibrant still life scene captured with the Fuji FinePix S3 Pro at 31mm, using a Tamron SP AF Aspherical XR Di LD (IF) 28-75mm 1:2.8 Macro [phi]67 lens at f/22.
Click on the image above to view a full resolution version (CAUTION: the linked file is very large).
Resolution / Sharpness (5.55)
The S3’s 23.0 x 15.5mm Super CCD SR II imaging sensor records 6.17 million pixels, but it has two photoreceptors for each pixel - an “S” photoreceptor for the basic image, and a smaller, less-sensitive “R” receptor for the highlights. When the camera is set for maximum dynamic range, the S and R data are combined to produce a single 6.17 megapixel image. The “S” site is very much like the ones on traditional CCDs, but the “R” site is much smaller and less sensitive to light. So, when the “S” is overexposed, the “R” site is not, which is key to the S3’s extended dynamic range. The camera can interpolate the CCD data into 12.1 million pixels, but not by using the “R” data as separate pixels.
To test the resolution of the S3 Pro, we captured several exposures of an ISO 12233 resolution chart and uploaded the images into Imatest Imaging Software. The software reads the imported file and detects the number of pixels active in forming the image. Our resolution tests are conducted under controlled studio conditions, using three softbox lamps totaling 350 lux of illumination. The exposures are recorded at various aperture settings and span the entire focal range of the applied lens. We report the best results attained, indicating the camera’s optimal performance capabilities. The results are given as both an exact pixel count and as a percentage score of the camera’s advertised resolution. When this is done, cameras that achieve a score exceeding 70 percent of their advertised megapixel count are viewed as "good" performers, while cameras that score beyond 80 percent are "very good" and anything exceeding 90 percent is designated as "excellent" and is quite rare.
Click on the chart to view full resolution file
Our tests concluded that the Fuji S3 records images with 5.55 megapixels of resolution. This is 91% of its advertised pixel count and earns the rare and coveted “excellent” distinction. By contrast, the Canon Rebel XT achieved a resolution score of 89% of its advertised count, though with an 8.0 megapixel sensor, the XT utilized 7.102 megapixels during imaging. But the 5.55 active pixels used by the S3 will certainly enable photographers the flexibility to tweak images to their client’s specifications and create 11 x 14-inch prints with stunning sharpness and tonal quality. The image displayed above was captured at a focal length of 41mm with an aperture setting of f/9.5. This was taken using the S3’s expanded dynamic range as most users willing to pay $2500 for a camera that features such a thing will likely keep the camera on that setting.
Noise – Auto ISO (0.0)
The S3 does not offer an automatic ISO setting, but contains a manually selectable 100-1600 ISO range, accessible in all shooting modes.
Noise – Manual ISO (14.17)
When shooting with the Fuji S3, users can choose a sensitivity rating between ISO 100 – 1600, in full stop increments, with the exception of an additional ISO 160 option. This is a flexible range, although it does not include an ISO 3200 rating which would significantly increase the camera’s low light capabilities.
We tested the amount of noise produced by the S3 Pro at each available ISO setting and imported the results into a regression analysis to determine an overall noise score. The incremental results are displayed in the chart below, with the available sensitivity ratings placed along the horizontal X-axis and the resulting noise plotted along the vertical Y-axis.
In terms of noise suppression, the Fuji S3 performed far beyond expectation, earning a 14.17 overall noise score. Even at ISO 1600, images produced by the Fuji S3 had impressive image clarity and minimal perceivable noise. To shed some perspective on the S3’s performance, the images captured by the S3 at its highest ISO setting (ISO 1600) contained less noise than images that would be shot from the Canon EOS 20D and Rebel XT at an equivalent ISO setting of 300 (as interpolated from our noise graph). When the cameras were set to ISO 800, the amount of visible noise produced by the 20D and EOS Digital Rebel XT was nearly double that of images shot with the S3 at 1600. The S3’s performance at 1600 even rivaled Olympus’ introductory/prosumer-level DSLR, the EVOLT E-300, at its lowest available ISO setting (ISO 100).
The tonal gradations and expanded dynamic range of the Fuji S3 are clearly strides beyond other available digital SLRs, and the S3’s ability to control noise is equally impressive. Although the S3 does not contain the resolution or speed offered by most competing models, the minimal noise and expanded tonal range of the camera results in images of superior quality. This test clearly shows why so many wedding and portrait photographers are willing to spend $2,500 on a 6.1 MP DSLR that can only capture high quality JPEGs every 1.02 seconds.
Low Light Performance (5.5)
To evaluate the Fuji S3’s low light performance, we captured a sequence of images at decreasing light levels. The images are shot at the camera’s highest available ISO setting without the assistance of the flash. This isolates the camera’s sensitivity to light and provides insight into the camera’s ability to capture shots at night or in dim indoor situations. Cameras are tested at 60, 30, 15, and 5 lux to simulate common low light conditions; 60 lux is comparable to an interior scene at dusk, while 30 lux appears similar to a room lit by a 40 watt light bulb. Images recorded at 15 and 5 lux indicate the camera’s ability to perform in near darkness.
|
Low Light Tests
|
 |
 |
|
60 Lux
|
30 Lux
|
 |
 |
|
15 Lux
|
5 Lux
|
Click on any of the above charts for additional image analysis
As you can see in the sequence above, the S3’s low light capabilities are what might be expected from the camera’s more generalized strengths and weaknesses. Images captured with the S3 under minimal lighting remain clean with only a slight degree of noise present. However, the four low light images display the S3’s auto focusing deficiencies, which become more pronounced as the available light is decreased. The camera also had difficulty achieving accurate calibration using the manual white balance settings in the expanded dynamic range mode in compromised lighting. The resulting images captured in less than 60 lux of illumination displayed strong reddish overtones in warm color values, while cooler tones were shifted and a bit muted. This is surprising and a bit unexpected considering Fuji’s professional reputation in color capability.
Speed / Timing
Startup to First Shot (9.41)
The Fuji S3 takes a relatively leisurely 0.59 seconds to start up. That’s the slowest start-up among the digital SLRs we’ve tested, but not by much.
Shot to Shot (8.98)
The Fuji S3 needed an average of 1.02 seconds between shots for JPEGs in extended dynamic range mode. Shooting RAW, shots went off every 0.67 seconds, but the buffer was saturated after only three frames, and it takes many seconds for the buffer to clear. The camera performs much better when extended range is turned off, indicating that the delay is the mainly the result of the significant data processing the camera performs to integrate the information from the two sets of photoreceptors. Since the camera’s main strength is its extended dynamic range, the lag time in that mode is clearly the one to consider when evaluating the camera.
Shutter to Shot (9.99)
The Fuji S3’s only good Speed score is an important one: we measured only 0.0001 seconds of shutter lag, so when you press the shutter, you’re going to get the shot.
