Here you can see a generated GretagMacbeth ColorChecker chart. Place your mouse over any of the labels below it to see the color reproduction in that mode. Select a camera/setting combination from the 'Compared to' drop-down to comparative boxes inside each patch.
Picture Controls (Nikon's name for its image parameter presets), can be modified, saved and transferred between the D3S and any other DSLR in Nikon's lineup to allow consistency of output between different cameras. Obviously the noise and noise reduction characteristics each camera will have an impact on the final results so the output might not be strictly identical at all settings. However, the results will be consistent enough that someone swapping between two cameras at a wedding or other event is unlikely to notice the difference in output.
|Nikon D3S||Compare to:|
Artificial light White Balance
The D3S doesn't do very well in our white balance tests, although the huge variation in temperatures that exists under the umbrella terms 'incandescent' and 'fluorescent' makes these types of lighting a difficult task for any camera. The Nikon D3S has an embarrassment of 'fluorescent' white balance presets - seven in all - and as we saw in our recent tests of the D300s and D3000 the 'white fluorescent' preset has done the best job of dealing with the standard domestic-type tubes that we use in this test.
It is good that Nikon has catered for the variation of types of fluorescent bulbs on the market, but to be honest, by the time you've taken enough experimental images to establish the correct one for your specific lighting, you may as well have created a custom white balance reading. Should you somehow be able to identify different types of fluorescent bulb on sight, we suspect that accurate white balance is the least of your problems.
It is worth noting that if you wish to shoot under the same or similar artificial lighting regularly, the color balance of all white balance settings (including auto) can be shifted independently on a amber-blue/green-magenta axis if necessary. Also, because the D3S uses focus information to bias its metering and white balance response, it is generally a little better in 'real world' shooting than these tests, taken in studio conditions, might suggest. This doesn't mean that the D3S's AWB system is 100% reliable - like many DSLRs it can deliver slightly different colors from shot to shot when working under some times of fluorescent and LED light, for example - but we're confident that it is at least as good as anything else on the market.
|Incandescent - Auto WB
Red: 7.9%, Blue: -16.8%, Poor
|Incandescent - Incandescent preset WB
Red: 7.7%, Blue: -13.8%, Poor
|Fluorescent - Auto WB
Red: 7.2%, Blue: -13.7%, Poor
|Fluorescent - White Fluorescent preset WB
Red: -0.1%, Blue: 2.2%, Good
Excellent performance here from the D3S, which has delivered extremely accurate color rendition in auto white balance mode when using flash. The flash in this instance was a Nikon Speedlight SB-900, facing directly towards the subject in both images. The D3S's matrix metering has done a good job here as well, and exposure is spot on in the portrait. There is plenty detail visible in both my subject's skin, and also his dark colored sweater.
Active D-Lighting has been included in Nikon DSLRs since the D300. It mixes a metering correction and tone curve shift (as per Canon's Highlight Tone Priority mode), with sophisticated dynamic range compression techniques, based on technologies from Apical (similar to those in the Sony DRO and Olympus SAT modes). ADL aims to preserve highlights and pull a little more detail out of shadows than would otherwise be possible in a single exposure.
The theory is that this results in images that are closer to the way the human brain perceives the scene. The cost is that you might see some additional noise in shadow regions of the images, and in DSLRs lower down Nikon's product line (the D5000 and D3000 in particular) the extra processing demands of the adjustment can slow the cameras down.
In the D3S, Active D-Lighting is off by default, and can be turned on in one of five modes - auto, then 'low', 'normal', 'high' and 'extra high'. Depending on the scene, the difference between any of the manual 'on' modes isn't always that noticeable, and I've shown an extreme example here, of the same scene taken with ADL turned off, and again, turned to 'extra high'. Notice how the dark stonework of this Norman church has been brightened, but highlight areas (the snow and the sky) have been pulled back slightly, to a pale gray, rather than bright almost-white.
This is an extreme example of the effect, and arguably not entirely natural, but it demonstrates the power of Active D-Lighting to push and pull the tones in a single exposure. It is also possible to apply Active D-Lighting adjustments to raw files in Nikon's Capture NX2, although unlike the in-camera adjustment, exposure is not affected (obviously) and ADL in Capture NX2 is essentially just an incremental tone curve adjustment.
|Active D-Lighting Off||ISO 400, 1/100 sec, F9|
|Active D-Lighting 'Extra High'||ISO 400, 1/200 sec, F9|
Overall Image Quality / Specifics
The Nikon D3S is designed primarily for maximum versatility. The D3X, and whatever eventually replaces it, is Nikon's benchmark for detail resolution at low to medium ISO settings, but the D3S sacrifices resolution for speed and exceptional sensitivity in unfavorable light.
We've seen in the comparison pages of this test that despite its (relatively) low pixel count, the D3S offers impressive detail resolution, but more important is that it can still deliver good levels of detail and sharpness at ISO settings that would have been unthinkably high until very recently. In fact, in the hands of a skilled practitioner, acceptable images can be captured right up to the D3S's ISO ceiling, of ISO 102,400 (equivalent).
Shooting in raw mode is advisable for best results, obviously, but even JPEGs still show a surprising amount of detail at these extension settings, and in the many hundreds of images that I shot in extremely low light conditions for this test, I see no evidence of any banding. This is impressive stuff, and there can be no doubt that the D3S sets new standards for image quality in extremely low light.
These dizzyingly high ISO sensitivity settings are not without their pitfalls, however, and depending on the situation, you may find that saturation drops off considerably when shooting JPEGs in exceptionally low light. Nikon's approach to noise reduction is to reduce chroma noise more aggressively than luminance. This stops things from looking too smudgy, but it comes at the expense of some grittiness. Also, because of the way in which chroma noise reduction works, color saturation tends to drop. You can see from some of my images of floodlit soccer that even with noise reduction set to 'normal' (default) skin tones can be rendered unnaturally (and unpleasantly) gray at high ISO settings in very poor light.
Naturally, however, most photographers will judge the D3S on its performance in the more commonly used ISO range of 200-1600 or 3200. Within this ISO sensitivity range barely a hair separates the D3S from its predecessor the D3. We've seen, too, that at the lowest ISO settings, the D300S also offers very similar performance to the D3S, only dropping behind at ISO 800 and above. In terms of image quality alone (putting aside other additional functionality) the advantage of the D3S compared to the lower-level D300S and D700 is simply that at high ISO noise starts to degrade images much later, which allows you to shoot at ISO settings like 6400 and 12800 without worrying too much about image quality.
Although raw capture is necessary to get the most out of the D3S' 12 million pixel CMOS sensor, sports photographers and photojournalists are far more likely to care about image quality in JPEG mode. Shooting JPEGs increases the camera's buffer to the maximum 130 shots per burst, and JPEG files are also a far more practical proposition when emailing or wirelessly transmitting images.
In JPEG mode, the D3S gives very similar results to the D3 and D3S at low and medium ISO settings, which is to say that detail capture is high, but not as high as a carefully processed raw file, and pixel-level detail is masked slightly by a general softness which characterizes JPEG output from all current Nikon DSLRs (and many of those from other brands as well). Turning the default sharpening up a notch crisps things up a little, but does not result in a noticeable increase in actual detail captured. Far better, when detail is an absolute priority, is to shoot raw and carefully sharpen images using an Unsharp Mask tool such as that found in Adobe Photoshop or Nikon's Capture NX2 software. The supplied View NX offers only a limited sharpness slider, which, although handy in a pinch, is something of a blunt instrument for more serious adjustment.
It is a shame (but not much of a surprise) to see that in-camera raw adjustment of the D3S's NEF files does not provide noticeably superior image quality to an equivalent JPEG. Although good results are possible from converting raw files in-camera, the advantage of the function is chiefly one of convenience, allowing you to correct for white balance mistakes, for example.
|JPEG (Contrast 0, Sharpening -3)|
|JPEG Standard (Contrast 0, Sharpening 0)|
|JPEG (Contrast 0, Sharpening +3)|
|Raw (ACR, no sharpening)|
|Raw (ACR, sharpening for maximum detail)|
It is obvious here that the Nikon D3S's JPEG output definitely needs some sharpening to look decent at 100% on screen. The default level of in-camera sharpening gives good results, but we wouldn't recommend turning sharpening up much beyond this point. As you can clearly see here, dialing in-camera sharpening up to its maximum setting is definitely a step too far, and the resulting image is over sharpened to the point where halos are evident around edges. Fine for a small print, but if that's all you need to make, you needn't worry too much about sharpening anyway.
Raw output on the other hand responds much better to sharpening post-capture (in Adobe Camera RAW in this instance). Although the final image in this selection, which I have sharpened for maximum possible detail rendition, looks a little 'overcooked' at 100%, it is clear that a lot more fine detail has been recovered than is possible from an equivalent JPEG file.