Go to advertiser website.
Go to advertiser website.
The DCS Pro 14n: worth the wait? - Continued

DCS Pro 14n colour is a good news, bad news story. The bad news is that 14n colour, as of firmware v4.1.2 and Photo Desk 3.0, is weak for a camera aimed at quality-minded pro shooters. The good news is that Kodak's DCS group can make changes to the colour processing in both 14n firmware and software, and all signs from Kodak indicate that they'll do just that in the months ahead.

A digital camera's colour is derived from both the response of the sensor and its colour filter array (CFA), and from the colour design that a manufacturer chooses for it. Of the two, the latter has the greatest impact on what a digital camera's photos look like. For that reason, there's reason to be optimistic that 14n colour can get a whole lot better.


14n photos lack colour variation. From the earliest sample photos from Kodak we'd been puzzled by the 14n's colour look. Even carefully-lit portraits lacked punch and life, in a way that wasn't explained simply by a lack of saturation or contrast. But we couldn't put our finger on what we were seeing. Or not seeing.

Once we began shooting a 14n side-by-side with an EOS-1Ds, the answer revealed itself. Photos from Kodak's high resolution SLR lack colour variation. This is the clunky description we've given to the 14n's colour sameness across scenics and portraits alike.

The morning skyline photo below, repeated from the Detail and Noise sections of this report, is a good example of what we mean. The EOS-1Ds version shows the reds, blues and browns of the downtown buildings, the green of the grass in the foreground, the orangy-yellow bridge lights and more. On a properly calibrated monitor, what you'll see in the 1Ds frame is a pleasing and reasonably accurate, if somewhat contrasty and saturated, look at Calgary's skyline that morning.

The 14n version is flatter and less saturated by comparison, but that's only a small part of what's going on. Look closer and you'll see that the some blues are mostly grey, the green grass isn't green, the sky and the bridge lights are both the wrong shade and lifeless. Then, take a step back and compare the entire 14n photo with the one shot with the 1Ds. The 14n photo shows, well, less colour variation.

DCS Pro 14n, ISO 80 (Zoom | Full-Res)

EOS-1Ds, ISO 100 (Zoom | Full-Res)

Boosting saturation and contrast in Photoshop helps a little, but it doesn't reintroduce color variation that seems to be lacking in the first place.  We've tried other ways of addressing the problem, including the building of custom camera profiles with two different high-end profiling packages, careful adjusting each photo with the extensive white balance controls in Photo Desk and processing the skyline and other photos with the Product HiColor and Product HiColor Hold looks from Kodak's extra-cost DCS Custom Looks package. Camera profiling was a bust. The DCS Custom Looks helped, though primarily to increase saturation. And while finding the optimum white balance on a photo-by-photo basis certainly helps get the best results from the camera, the lack of colour variation always remains.

Kodak's Steve Noble calls this a lack of midtone colour contrast, and that may more aptly describe the problem. It's a drag on colour quality that is apparent in most of the 14n photos we've taken. For scenes that don't have much colour variation, such as this photo, or the photo of the flowers below, it's primarily contrast that's needed to get the photo ready for printing. So, under the right conditions this colour problem isn't a deal-stopper.

DCS Pro 14n, ISO 80, WB: Daylight 6000K (Zoom | Full-Res)

EOS-1Ds, ISO 100, WB: Cloudy (Zoom | Full-Res)

But, almost without exception, 1Ds photos have colour that is more varied and more photographic. Not perfect: we're not trying to paint the EOS-1Ds as having flawless colour. But even without additional, colour improving steps such as building a custom camera profile and/or processing its files through Capture One DSLR, 1Ds colour is a really good starting point for editing in many cases.

Case in point: working up the two pictures below for printing, the 1Ds version showed good color variation, good contrast and reasonably accurate hair, skin, sweater and background colour but a slight overall green-yellow bias. Fixing that brought the rest of the picture into line.

The 14n photo, by comparison, renders the hair, face and even the sweater as somehow from the same colour family, while the blue in the background is wrong. No amount of effort in Photoshop restores the truer and more pleasing overall look of the 1Ds frame. Be sure to look at least at the zoomed versions of each photo to see what we mean.

DCS Pro 14n, ISO 100, WB: Custom WB (Zoom | Full-Res)

EOS-1Ds, ISO 100, WB: Custom WB (Zoom | Full-Res)

One final example of this is found in the team photo below. In the 14n version, the owners' suits in the middle row are almost all about the same shade of grey/black. In the EOS-1Ds version, the actual colours come through: some suits are blue, others are gray/black. As with most things we're describing on this page, if your monitor isn't well-calibrated and profiled you may not see what we're talking about.

DCS Pro 14n (ZoomFull-Res)

EOS-1Ds (ZoomFull-Res)

Again, the degree to which this is a problem seems to be dictated by what's being photographed, more than anything.

14n photos have a red bias. Like the lack of colour variation, most 14n photos we've taken share another characteristic: an overall red bias, even when the picture has been balanced off a gray card. Couple this with the lack of colour variation and you have a particular colour look to some 14n photos that is difficult or impossible to correct.

Note: There's a bug in Photo Desk 3.0 that causes a slight red-yellow shift in all 14n pictures that have been custom white balanced in the camera (unless that white balance is subsequently overridden in Photo Desk). We discovered this early on and switched to performing a Click balance in Photo Desk 3.0 instead, which works as expected. So, while we've labeled all photos whose white balance has been set off a gray card as Custom WB, in all cases those photos have actually been balanced using the Click function. A near-term update to Photo Desk should fix this, enabling both Click and Custom WB to once again produce the same result.

You can see this effect in the 14n portrait of the woman above, which has an overall redness that doesn't stem from an incorrect white balance. And you can see it in the hair of the soccer portrait below (especially the full-resolution version), and even in the blue jersey.

DCS Pro 14n (ZoomFull-Res)

EOS-1Ds (ZoomFull-Res)

It's also plainly visible in this mid-morning silhouette, and impossible to remove by adjusting the white balance in Photo Desk (though we've taken care to chose a white balance setting that minimizes it, as we have with all photos in this report). At least, not without causing a different and worse colour problem.

DCS Pro 14n, ISO 80, WB: Daylight 4500K (Zoom | Full-Res)

EOS-1Ds, ISO 100, WB: Daylight (Zoom | Full-Res)

In fact, if you look at just about any 14n photo in this report you'll see the red bias we're describing. You'll also see that reds themselves tend to be off. If you would like additional examples of the red bias and lack of colour variation that typifies 14n colour, download the latest batch of studio portraits posted on Kodak's web site.

Interestingly, though this characteristic has mostly diminished 14n colour quality, we've taken two photos that have benefited from it. The skyline photos below look like they were shot on different days, and not within seconds of each other as they were, so different is the overall colour rendering. The 1Ds version is more accurate, but in this case we really like how the red bias has given the entire 14n picture, including the sky, an especially warm glow that simply wasn't there.

DCS Pro 14n, ISO 80, WB: Daylight 5000K

EOS-1Ds, ISO 100, WB: Daylight

The same applies to this pre-dawn shot of the Golden Gate Bridge. To enhance the coolness of the morning light, both cameras were set toTungsten WB. The EOS-1Ds gave the result we expected: a cool blue cast. The 14n version looks like just the right mood-enhancing filter was placed over the lens. But it's not so: the camera's red bias did all the heavy lifting.

DCS Pro 14n, ISO 80, WB: Tungsten 3000K (Zoom | Full-Res)

EOS-1Ds, ISO 100, WB: Tungsten (Zoom | Full-Res)

As a default colour rendering, the red bias is a problem. But it does appear to have a place in the 14n toolbox.

One colour quirk is completely solvable. The colour shift visible in the photo below results from a soon-to-be-fixed bug in firmware v4.1.2. But, to the extent that you might see it, or shoot important pictures with the affected lens prior to the bug's squashing, here's how to work around it.

The 14n processes the RAW image data differently based on the lens in use, to compensate for a red-green colour shift that can occur across the image. The shift results from the angle at which light strikes the CMOS sensor. Kodak has tested a range of lenses and built internal tables that group lenses into one of three categories. It then applies the correct compensation automatically and the shift is gone.

DCS Pro 14n, Nikkor 85mm f/1.8D (Zoom | Full-Res)

As of firmware v4.1.2, however, the Nikkor 85mm f/1.8D is miscategorized. Until a revision of firmware appears that corrects the error, the Lens Optimization menu on the camera must be set to Type I manually when shooting with this lens, then set back to Auto afterwards. The correction is applied to the RAW data and can't be undone, so Lens Optimization must be set correctly in the camera before the picture is taken.

Out of the bag full of lenses we used with the DCS Pro 14n, the 85mm was the only one affected. In all other cases, leaving Lens Optimization on Auto meant that the camera applied the correct processing based on the category that lens was in. And we expect to see a 14n firmware update soon that fixes the problem with the Nikkor 85mm f/1.8.

Our experience here should serve as a general heads-up, though: if you use lenses from a variety of manufacturers, or shoot with fairly ancient or obscure Nikkor glass, Lens Optimization's Auto setting may falter, since Kodak can't possibly test all lenses out there that fit on a Nikon mount. New purchasers of the 14n would be well-advised to shoot open sky or some other flat field with all lenses and determine which ones, if any, require a Lens Optimization setting other than Auto.

The 14n delivers its best colour in the studio. Or, in our case, when using studio lighting on location. Though we haven't been satisfied with the colour from the 14n in most situations where we've put it up against the 1Ds, we have observed that the colour differences have been the smallest working in a studio-like environment. As we discussed in the Noise section, the 14n is a camera that likes to be surrounded by big strobe lights. Used that way, you may find the colour at least acceptable. Though, to be honest, we don't.

Kodak's Steve Noble indicates that the colour aim point for the 14n was the same as for the DCS Pro Back and DCS 760, though it's clear that aim point was not met. If Kodak can achieve that aim point with the 14n through colour processing changes, that should bring about a noticeable improvement in colour from this camera. Based on the response from Kodak to our colour criticisms, we're optimistic that 14n colour will get better, perhaps a lot better.

But, the EOS-1Ds has raised the colour bar significantly since its introduction, both because its native colour processing is good (using a smart mix of Color Matrix 1 and 4, depending on the photo), and because EOS-1Ds users have a powerful colour processing alternative in the form of Capture One DSLR. So, it may be necessary for Kodak to look beyond the 760 and Pro Back and develop a new colour reference.

Next Page: Conclusion
Go to Page: 1 2 3 4 5 6 7 8 9 
Go to advertiser website.
2000-2013 Little Guy Media. Not to be reproduced without written permission.