I’ve spent the past few weeks taking a break from my One Camera One Lens experiment with the Fujifilm X100V. While I intended to go back to it, I blundered into something extremely interesting: Sigma Foveon cameras.
I’ll try not to be too technical but it’s worth taking the time to explain why these cameras are so interesting to me and many other photographers. Nearly all camera sensors out there use what’s called a Bayer matrix.
This is a color filter that sits over the sensor. To see color (since all camera sensors are actually panchromatic) the Bayer matrix splits light into red, green, and blue (mega)pixels/channels.
This alternating grid-style color filter works very well for most types of photography, but it has a few disadvantages: each pixel is filtering light down to a single detected color, which reduces the sensitivity of the inherent monochrome sensor. Also, the camera uses what’s called Bayer Demosaicing to interpolate (calculate and infer - guess) which colors go where since patterns in nature aren’t as regular as the color filter array.
Even the Fuji X-Trans that I know and love does this. While it uses a more random color filter array for better (IMO) color reproduction, it is essentially a Bayer variation and relies on interpolation.
Sigma Foveon is one of the very few sensors out there that use an entirely different technlogy from either Bayer or X-Trans. The sensor architecture can be explained in a few different ways. But the way that makes the most sense is that it’s actually three panchromatic sensors stacked on top of one another. The top layer is skewed slightly towards blue, the middle towards green, and the bottom layer towards red.
Foveon sensors rely on calculating how far light penetrates each layer of silicon to create an image. What’s great about this approach is that it’s essentially seeing true color, or at least as close as you’re going to get to it. The tonal range is far greater and more accurate because there’s no Bayer Demosaicing to guesstimate which colors go where and in what intensity. The resolution of the stacked sensor design is also far greater than you’d expect, on par with modern full-frame sensors.
Oh, and the DP Merrills, like the one I have? They are under 400 grams each and have super sharp f/2.8 fixed prime lenses attached.
I think I failed at not being technical but at least most photographers should understand what I just said. So as great as it sounds, you’d think “why isn’t everyone already using one of these sensors in their cameras?” Because they are a pain in the ass, that’s why.
It takes far more processing power to render a Foveon image than an equivalently sized Bayer image. Also, the sensor architecture has significant weaknesses in terms of low light performance. Shooting beyond ISO 200-400 creates horrendous color noise in most situations. Sometimes you can shoot ISO 800 in some scenes. And in monochrome, you can push it to ISO 1600-3200. But it’s a very light-hungry design.
The battery life sucks, the autofocus sucks, the handling is…Pretty good, actually. Sigma’s photo processing software is…Slow but serviceable. Oh, and the DP2 Merrill is 9 years old and is essentially a dead technology as Sigma’s advanced to a new sensor architecture…
But the photos. The PHOTOS. You really do have to see them on your own computer to believe it. I’m sharing some here but you really can’t appreciate them unless you’re pixel peeping, using a high-resolution monitor, or looking at them in print. For $4-500, you can have images on par with a Nikon D850 or even a medium format camera in the right light environments, I kid you not.
There is just so much detail but it’s not just sharpness. There’s this micro contrast and subtle tonal gradation that make even the most boring subjects pop. It’s like looking with your eyes somehow, it’s really hard to describe what it is I like about the Foveon look. It’s very realistic, I think. The colors don’t really leap at you in the way most other cameras do. You can definitely create that more saturated style with less tonal gradation. But I went for a very natural edit with these shots to try and show the Sigma look.
There’s this snap that even my X100V can’t generate. I did a direct comparison between the two earlier and it was painful for me, I must admit. The Fuji photos look amazing until you pull up a Sigma Foveon image. And then they look more like watercolor renditions of the Sigma image. With enough processing, I can get them looking close, but there’s still a distinct Sigma look that I can’t recreate.
Honestly, I’m wondering if I’m wasting my time even showing images here, given the size of the previews and the lack of anything colorful during a Milwaukee winter. Make sure to zoom in on the photos if you can.
Color photography is where the Sigma DPs really shine. Landscapes, portraits, still life…Lots of light and no movement but things are so bleak right now…It’s pretty easy to find samples on Flickr and reviews from years back when Foveon seemed like the next greatest thing to rock the photography world, though.
So where will the DP2 Merril fit into my camera bag? I’m honestly not too sure. In terms of resolution, it’s actually TOO good because my screen can’t actually display it without me pixel-peeping. The focal length and resolution really demand a tripod as well to get the sharpest possible image. My X100V on the other hand is far faster, more flexible, more fun, and much easier to use.
They also seem to call for different subjects. When I’m rocking the Sigma, I’m looking for fine details in things. The grains in rocks, wood, and bricks, tangles of tree branches, delicate tonal transitions, bright punchy colors…It’s a very different way of seeing the world.
Fortunately, I have a Macbook Air Retina on the way anyway as it was time to upgrade my 2012 machine. I’m curious to see how much better the DP2 images look on a high-resolution Retina screen. Since I don’t print often I might have to send the DP2 back if I’m not able to appreciate the difference between it and my X100V without pixel peeping.
I do enjoy pixel peeping on occasion but it’s not something I care to do for every image I take. Still, the differences in color rendering are very obvious even without pixel peeping. With its 45mm equivalent lens the DP2 might make for an excellent portrait/landscape/monochrome camera for moments where I really want the absolute best possible image quality and I have the light to make it work. But it’s a fussier way of working than I’m used to.
I look at these images though and I do fancy the idea of adding the DP1 (28mm equiv) and Dp3 (75mm equiv) to my bag one day! If they weren’t such pain in the ass cameras.
That image quality though…