4.02.2018

Observations about cameras as a result of a nostalgic dive into three and four generation old models. Warning: This one is about traditional DSLRs, not Mirror-Free.


Everyone needs a hobby. Even professional photographers. My hobby is photography and in the pursuit of this I sometimes follow Alice through the mirror and have adventures that are....less than rational.

About a month ago I was bumming around Precision Camera, handling all the lights, asking to see weird lenses in the used case, and generally making myself an annoyance. Didn't seem to phase the staff who are either used to my shopping habits or just had nothing better to do at the time. They dutifully pulled out old Hasselblad lenses and ancient Broncolor flashes so I could play with the focusing rings or the knobs and controls before shaking my head and moving on to the next shiny object that caught my eye.

And that next shiny object was a very nice, relatively unscathed Nikon D2XS which was sitting, unloved, in the glass case with other cast off Nikon DSLR bodies. I had someone extract it from the case and I played with it for a spell, all the while remembering when I had owned and extensively used one, many years ago. The price was minimal so I bought it, rationalizing that I'd use it with some of my older Nikon manual focus lenses.  I came home and charged the battery and re-familiarized myself with the old and simple menu and then shot with it for a while.

There are a few things I remember from shooting murals with this camera back in 2007, one is that this camera is exuberantly happy at ISO 100, relatively content at ISO 200 and starting to get a little edgy at ISO 400. By ISO 800 we're veering into full blown noise anxiety. Shooting raw and post processing with finesse and experience might get you a relatively decent ISO 800 (if you nail exposure!) and a fairly usable ISO 1,600. I also remembered that when I shot my original D2XS at ISO 100 and used my best techniques the files that I could get were pretty much perfect on many levels and could be easily enlarged to just about any end application. I find that is even more true today with all the Adobe PhotoShop's constantly improving re-sizing tools.

The D2XS is huge and heavy and the shutter is loud like banging trash can lids together. But the whole package certainly has its charm for an old school photographer. I give muscle memory a nod for a certain amount of my current nostalgia --- decades of form combined with function make re-accessing old cameras just like getting back on a bicycle....

A week or two later I ran across another old Nikon I remembered from my past. It was a nicely preserved D700 and after I played with it for a while I remembered the beauty (especially for files used on the web or used smaller than 11x17 in print) of the large pixel files I routinely got out of that model. I decided to add it to my growing collection of "hobby" cameras. This purchase engendered a secondary purchase of a smattering of older lenses, hand-picked for their cheap pricing and their under appreciated sharpness and general performance.

The one lens I had that I wasn't entirely happy with was a used 50mm. It was too new and I wanted to find a nice, older 50mm f1.4 ais model to augment the plastic AF model. So I pointed the car north and went back to the store one more time----- just to look. As far as lenses go I came home empty handed but continued my collecting lunacy by buying a nice copy of the more recent Nikon D800e.
And that's what I wanted to write about today, the D800e.

The D800 and D800e were interesting cameras. At a time when 24 megapixels seemed like the resolution end game for 35mm framed cameras these two cameras took the whole industry up a notch to 36 megapixels of resolution. They were also the leading edge of a generation of cameras that, along with the Sonys, were becoming ISO invariant (sensor noise floors low enough that they could be raised dramatically in post production without provoking the shadow noise that has plagued digital cameras from the beginning).

One thing I did not remember from earlier research on the D800 series was the availability of uncompressed, 4:2:2 video from the clean HDMI set up. I'll be testing that when I have some down time....

I set up some studio flashes and used the Nikon 105mm f2.5 on the D800e to test the camera. I was fairly conversant with the menus having used a D810 extensively and relatively recently. For all intents and purposes the files I created in the studio were on par with those I had routinely gotten from my D810; noiseless at ISO 100 and with detail that just goes on and on. But the thing that I had forgotten, after my long immersion with Sony, Olympus and Panasonic, is just how good and mature the color science of Nikon cameras is. They've been doing digital for a long, long time and even though I wish they'd make the leap to mirrorless in at least some of their APS-C and full frame models I have to admit that they (and Canon) know the formulas for pleasing color.

Many articles recently have been tossing around the topic of "Color Science." The general understanding (at least how it pertains to Jpeg files from cameras) is that Olympus is a master at making colors that please most users, as is Fuji, and, that while Canon colors are warmer they too have a huge fan base of photographers who find the Jpegs from their 5Dx cameras to be subjectively, visually wonderful. Nikon has the reputation for generating files that are a bit more "analytic" and less pleasing OOC but which can be edited into submission without too much of a struggle. Panasonic was, for a long time, dinged for crappy skin tones but have made huge strides in fixing their Jpeg renditions in the newest series of cameras (GX8, GH5, GH5S and G9). Sony got low marks for their Jpegs until this latest generation and they finally have circled around and started delivering much nicer skin tones and generally pleasing color.

Many years ago Kodak and Fuji both dove deeply, and with huge budgets, into the "science" of creating two kinds of color for their film stocks. There were two different objectives in the making of color films and the objectives were often at cross purposes. It turns out that there is accurate color and then there is pleasing color. Accurate color is based on delivering a recording medium in which the colors match known references as closely as possible while delivering a saturation and contrast that also matches measurable targets. Kodak and Fuji both delivered several transparency and negative film stocks that were as accurate as their science could make them. But there was an issue with acceptance by the general public.

Seems that their general consumers (the people who made up the overwhelming bulk of the film buy-in market) didn't care nearly as much for accuracy as they did for what Kodak called, "Pleasing Color." And in North America that meant much more saturated colors, warmer skin tones, less accurate but richer yellow and blue hues and, in general, a much less "correct" approach to accurately capturing a photograph. I don't know exactly how this cultural vision evolved (and, yes, it is somewhat cultural according to studies by Fuji and Kodak...) but I conjecture it had to do with what people were seeing in regional movies and on television at the time. I think domestic advertising was also pushing more saturation and color in their work at the time of the "pleasing color tipping point" as well.

This led to a decline in popularity of accurate film stocks and something of an arms race to create film stocks with ever higher levels of saturation and candy color. Not at all accurate but happily embraced by millions and millions of hobbyists, moms and dads and even some pros. But Fuji and Kodak were kind enough(?) to continue to make and provide color accurate films to working professionals who might be working with critical color requirements (the Cheerio box, fashion make up or car interior color samples needed to be a close match to reality to prevent client revolt!!!) and we made good use of the neutral stocks, especially when making color ads for book covers and when shooting floor material catalogs.

So now we're in the age of digital imaging and the software of our cameras can be tweaked to deliver a range of colors, tones, saturation, contrasts and hues from the available sensors. Having precise metrics to aim for makes it easier, on one hand, for camera makers to proceed when creating a color/tonal menu assortment for their cameras if the end goal is accuracy but if the goal is pleasing color or most acceptable color ("color" including: saturation, hue, tone, and contrast tweaks) then the design of a camera's color space becomes more like gourmet cooking and less like slavishly following the recipes in "The Joy of Cooking." 

There are qualities such as the angle of the curve of the highlight rendition, mid-range contrast tweaks, color responses in the twitchy red and blue spectra, and a lot more. Fuji had a head start in the "pleasing" color race since they could access so much data from the film days. In almost every camera with pleasing color reponse there is a gentler roll-off in the highlight areas, a bit more contrast in mid-tones and a pleasing red/yellow combination in the area close to skin tone. The one area where there is more differentiation is in the depiction of blues.

If we move from Jpegs to Raw files some of the differences between cameras become less obvious as much of the color "flavor" is provided via the interpretation of the Raw processor being used. Files from Nikon Capture, DXO and Adobe are obviously different if one uses each Raw processor's defaults.

While it should be possible to create profiles or even LUTs (look-up tables) to make one company's line of cameras resemble another company's line I think it would take a deep foray into the camera's software to match them more precisely. A deeper foray than most photographers have the time of inclination for...

Camera companies decide where on the spectrum their spectrum will exist. There is a range from "very pretty color" to "absolute color" and it will be affected by things as disparate as the coloration of lens coatings to the regional markets in which the cameras sell most profitably. The bottom line is that companies are taking pretty much the same raw data off the same kinds of CMOS sensors and overlaying a look and feel that they feel will sell best.

It seems to make sense that cameras aimed at the lower end of the buyer demographics will have punchier, more saturated and more culturally nuanced color aim points than cameras aimed at much more exacting and demanding users such as advertising professionals. The files straight out of a Canon Rebel will look, to most consumers, better than the files from cameras with lower saturation levels and flatter profiles. Since the expectation is that most consumers will perform less post production the color science of a Rebel or Olympus EM10-iii is a "win" for sales. A more accurate color response would probably reduce sales, within specific markets.

If you are curious about the color accuracy or color delta of a camera you can use controlled and known lighting to shoot known color targets and judge the results on a vector scope which can show you how far the camera's response is from the accuracy of the original color as well as the degree of saturation for each color of the target. This, of course, presumes that your camera is able to output HDMI.

What it mostly boils down to is that Jpeg shooters should have a keener interest in just how the different camera companies choose to craft their Jpeg color science because, in Jpegs the color is "baked in" and harder to change without consequences than a raw file.

If you are a raw shooter who sometimes needs real color accuracy to produce accurate results for commercial client you may need to use a camera aimed at more absolute color renditions even though you might not like the "straight out of camera" Jpegs as much as cameras from other makers. But when shooting raw it should be possible to create settings that will get your camera closer to pleasing color and further from absolute color without too much effort.

So, what do I think about the files I'm getting from my collection of Nikon's older cameras? Interestingly the Jpeg files when used with Nikon's neutral profile setting (in camera) are pretty darn accurate. They got a lot correct back in the day. If you want to get closer to a Fuji, Canon or Olympus "look" you'll need to make changes to blue hues, overall contrast, mid-range contrast and parts of the red and yellow color saturation levels and a few other things. And none of what we've discussed here includes the various sharpening settings to which the cameras default....

Interestingly enough, the more controls camera makers include for videographers (see a Sony RX10iii menu to understand just how changeable files can be, in camera, before you poo-poo the idea) the more controls you have at your disposal to transmute the Jpegs to your taste (assuming you can access these profiles in regular photography!!!). Some folks on the web have even created mini-idsutries in fine-tuning camera colors.

I cut my digital teeth on Kodak's ancient DCS 660 and DCS 760 cameras which worked in raw only in Kodak's software for a long time. You had more control but you had more options with which to fuck up. The Nikon professional DSLRs seem to be set up to be conservative in their overall color responses --- a neutral color science. While it requires more tweaking before we can put it in the same "pleasing color" ballpark as some competitors the neutrality is welcome for demanding applications where built in casts are less welcome.

What will I buy next? It's a toss up. The 45mm f1.2 for the m4:3 system or a 20mm lens of the Nikons. All depends on what kind of job hits the inbox next...