Microcosmologist

Pentax is blowing up the interwebs this week with news of their newest sweet photo machine, the K-01. New sweet photo machines are a dime-a-dozen, but what makes this one interesting is that it’s the first mirrorless interchangable lens camera (we’re going to stick with the acronym MILC here) that actually shares a lens mount with their normal SLRs, meaning you can attach any of your existing K-mount lenses right on to this baby. Now that’s progress! No need to repurchase thousands of dollars in precious glass to be totally outfitted.

You’ll also notice it’s offered with a shockingly thin (read: skimpy?) 40 f/2.8 prime. That’s an odd novelty, but with glass that tiny, does it actually take decent images? Anyway, I don’t care much about that, back to the body–It’s cool to finally see someone make a mirrorless camera that doesn’t demand a whole new army of lenses. It’s about time!

And still no MILC from Canon… I’m waiting fellas!

Runner up for interesting new camera in the mirrorless world would be that Fujifilm X-Pro1. It’s predecessor, the X100, was featured on here a while back and remains on the short list of extreeeemely lust-worthy photo gear that I would totally buy if it were only a little more affordable. The X-Pro1 would probably get a whole post devoted solely to it as well, if it weren’t even further into the netherworld of unaffordability. At $1700 for the BODY ONLY, I wonder how many units of this thing Fuji is going to be pushing. For that price, you could get a rock solid Canon or Nikon DSLR AND a lot of great glass too. Unfortunate. But it does have enough unique features that I want to blab about it for a minute. Let’s start with some choice blurbs from the official press release:

That multiple exposure thing is pretty rad. Every camera should have that, using preview on the LCD. That’s science. The coolest thing about this camera is definitely the color filter array though. Okay kids, put on your nerd caps, cuz shiz is about to get hardcore up in here:

Every digital camera uses an image sensor to collect the photons (light) that make up the picture. Since the sensor itself cannot discriminate colors, filters are used to split the light into RGB (red, blue, & green) components. By combining RGB in different combinations, you can then spell out any color imaginable. Old school photographic color film accomplished the same idea by having (at least) 3 seperate layers of silver halide salts which were dyed differently. Since film was a chemical/analog process, those salt crystals weren’t arranged in any sort of perfectly aligned matrix, they were just scattered all over, however they happened to fall. The characteristics of any given film (contrast, sensitivity, resolution) were determined by the crystal sizes and the amounts of the silver halide for each different color layer. Fancier films had up to 12 different layers used to reproduce colors! So it wasn’t just RGB, it was a whole lot of different colors being combined, and each “pixel” in the film was of variable size.  Click the image of those Kodak T-grain silver halide crystals to see more Scanning Electron Microscope (SEM) images of different films.

To obtain digital color images, almost all sensors use the standard 2×2 “bayer” color filter, named after a clever Kodak engineer who came up with the scheme during the 70s. (If you’re wondering, Bryce Bayer chose to use an extra green pixel to emulate the human eye’s sensitivity to green light, which is kinda neat!) So nearly every digital camera uses this method of light collection, and even those rare cameras which deviate from the standard color filter configuration still have some kind of repeating, contant pattern. Although it’s not completely clear from the Fuji PR-speak, if their color filters really are randomized across the whole sensor area, woah, that is a major difference in the way light is gathered… and a pretty ingenius mimicry of analog. As a giant in the tradition of film and therefore owner of masterful knowledge upon the characteristics of it, seeing Fuji bring the lessons learned over decades of experience in color reproduction using silver halide emulsions to bear in the new digital era is… way cool. If anyone can do it, it’s them (or Kodak?).

The story of the transition between the analog epoch and the digital epoch is a tale chock full of “back to the future”-isms. In the sense that as digital technology gets better and better, the goal seems to be emulating the way analog things used to be. We see this is audio mastering (see TRacks-tube/tape emulation mastering software), audio reproduction (vacuum tube pre-amp stereos, and progressively higher sample rates to recreate a more ‘analog’ waveform), musical instruments (amps and effects pedals that strive to recreate the analog ‘warmth’, synthesizers that model their vintage predecessors), and even video games with roms and emulators built to resurrect the early days of gaming. To me, it makes perfect sense that photography, when the technology gets good enough, will undergo this same trend of digital modeling to recreate an analog era gone by.

Right along those lines, Fuji’s film simulation modes are an ultra-sweet concept that I have seriously been wishing someone would do for many years now. I wonder how those work, exactly… searching for information doesn’t turn up any explanation; I’d be willing to bet that the only difference between the three film simulation modes are simple tweaks to contrast and saturation. The example images shown on this excellent webpage would seem to support that theory. Although it’s complex to compare film ‘resolution’ to an equivalent digital resolution, analog film has somewhere between 15-25 megapixels of ‘resolution’ depending on who you ask. My 18MP Canon t3i starts to inch into this territory. As sensor design pushes even further into the higher megapixels, it’d be really cool to see the characteristics of old school film emulsions replicated not just at the post-processing (software) stage, but at the light-acquisition (sensor/color-filter) stage. Software enhancements are never going to beat photoshop at its own game… but hardware that captures color in a new way? That would be an innovation.

What would it take to truly create an honest digital carbon-copy of film emulsion? You’d need a variable color filter and as many megapixels as you could throw at it. As in, a color filter where you could actively switch which pixels recieve which color, and do it on the fly. We’d be aiming to mimic the salt crystals as you see in the SEM image, with variable sizes and variable locations for each individual exposure. You’d have some kind of randomization that would rearrange the configuration every time a new photo is taken, within a given set of parameters for each different film emulsion. To really nail it, you’d also want to not just use RGB in your color filter, but a larger variety of color shades to mimic the dyes from many-layered emulsions.

I lazily mocked up the idea here with a grid over an SEM image of Kodak Portra 160VC  Professional portrait negative film. Average grain size: 1μm.  I only colored in some film grains because it would take forever to do the whole thing and I got tired of clicking.  I did RGB and CYM(K) with a lot of K, I guess.  But you could use many combinations of colors in your color filter, both echoing classic film emulsions, or even getting creative with more funky configurations.

That kind of light capturing would offer something you could never do in photoshop, due to the fact that it would change the way the actual sensor is classifying colors of photons which make an image *as they come in*. I have no idea how one could create a variable or “active” color-filter, but man, if you could do that, it would open up incredible possibilities in the way light is captured. Think ultraviolet and infrared too. It’s all conjecture, but I find this stuff awesome to daydream about.

</nerdery>