The Fuji X100S Lens Rendering

Warning: This post contains extremely esoteric crap and highly subjective non-scientific study based solely on visual observations. It’s meant solely for consumption by absolute visual geeks like me.

There we go, with the disclaimer out of the way we can move on to visual geekery. First ask yourself this question; What does the Fuji X100 series of cameras have in common with a Nikkor 28mm lens from the early 1970’s? The answer; More than you might think. Oh, beyond the visual lens rendering geekery I’ll probably drop a few tidbits having far more value of just playing around, projects, and how one thing leads to someplace completely different with very little ability to predict where. Something I continue to be amazed by, an extension on the thoughts I posted along with LaRoque over the last few winter seasons.

Photo Rendering Geekery

I mentioned somewhere on the blog in passing that a 28mm lens I got for the princely sum of free had quite a few characteristics I liked. Some of those characteristics were a shock at just how good one of the first retro-focus designs from so long ago performed compared to the huge expensive monstrosities so popular now. Some of those characteristics where physical such as how tiny it is and how good it feels to use and focus. The last thing I liked about it was how it rendered. In fact I also mentioned somewhere it reminded me a whole lot of the way the 23mm f/2 Fujinon built into the X100 series renders.

Not so, so clinical wide open, out of focus areas look really nice, super sharp a couple of stops down, and all that. Today I decided to compare the two lenses from different universes side by side. Not an exhaustive test, not really even a performance comparison at all. I just wanted to see how close the rendering of the two optics really were. Turns out my visual memory in terms of over all look proved to be closer than I imagined.

Exhibit A: The shot at the top. X100S f/5.6 one of my favorite StupidCrap™ test subjects for analysis, plastic blinds shot out of focus. The out of focus was to get an idea of how dark/light subjects are rendered by both of these diverse lenses. Exhibit B below is believe it or not the Nikkor 28mm f/3.5 shot at f/5.6 as well using the same highly scientific patented out of focus technique. Namely racking the focus on both cameras as close as they would go and shooting a subject a little farther away. I also did the reverse — as in focused at infinity and shot the same frame. These results are more definitive as the blinds were only two-ish feet away. The infinity shot was far more blurry which tells you less about rendering qualities as more blurry is actually harder to judge.

Pretty stinking close huh? Forget about minor exposure or WB difference here that’s not what we’re looking at. I was shocked at how close the rendering or drawing properties were. Why f/5.6? Funny you should ask — I’ll get to the important reason in a moment but a less important reason is I wanted to see how very different aperture configurations looked — again shocked at how close the rendering is considering the difference in number of blades, aperture shapes, etc…

For the more technically minded of you here are a few things having nothing to do with OOF rendering you might find interesting:

  • Note the very low distortion of the 28mm with no software corrections. The Fuji has some sort of built-in corrections.
  • The 28mm Nikkor from ancient history shows a major technical fault here as it’s the perfect storm of image to show those off. Note the strong vignette combined with the cyan shift in the corners and edges. This is mostly gone at f/8 and is much more pronounced on digital. A lot of really great and expensive optics do the same thing that were designed for film. People on the internet tell me it has to do with lenses that have very sharp angles of projection at the edges of the sensor or some such thing. I mostly shoot BW with this lens so I don’t care and as far as vignettes go. I can fix this if I want but I usually keep it in mind as a tool to be used at f/5.6 and larger.
  • Note the strong flare of the 23mm Fujinon in the upper right. Funny stuff, this lens has more flare even at 5.6 it’s hard to not have it do this. Especially funny because the Nikkor has a lot of flare but was no issue shooting the same scene. Theoretically it should be much worse, it’s single coated, has a far larger front element — huge, and has a wider field of view with no hood. Yes, I made a few with no flare on the X100S but it was actually difficult to not have it. Again a feature to be used and controlled as much as it is a fault.

You might be saying something along the lines of “you’re cheating, the subject is too simple to evaluate out of focus rendering”. You’re right, it was a starting point. Let’s look at some more complicated StupidCrap™ How about some highlights and things…

First the Fuji 23mm, again at f/5.6

I like the way it renders, always have. One of the reasons I love the Fuji X100’s and also the reason I shoot out of focus things on purpose occasionally when the mood strikes and the subject makes any sense beyond StupidCrap™ tests.

Now here’s the Nikkor 28mm at f/5.6

It’s way more out of focus. Hmmm, oh yeah, it’s 28mm slightly. Let’s not bother with nits. I would definitely say they rendering although displaying different degrees of focus are absolutely in the same family.

How about one from the Fuji at f/4. Wow both the Fujinon and the Nikkor are super similar in the way they render OOF stuff. At least from what characteristics I’m looking at.

Okay, I won’t bother with showing the other three dozen shots of various complexity. Take my word for it shot after shot proved I wasn’t imagining things when I had the notion that the 28mm Nikkor from 1973 and the Fujinon 23 rendered images with extreme similarity. Both are in the category of likable to me vs the other category — hateful.

More Philosophical Matters

I’ve mentioned in the last four dozen posts I’m attempting to actually finish — as in out the door done — a project I shot in 2014. The project was born of an accident of just doing something on a whim. I happened to shoot all of it on the 28mm lens referenced above. A lowly 28mm f/3.5 Nikkor Non-AI lens I got for free. I knew I like how it rendered, it was plenty sharp when I wanted it to be, small and unobtrusive, perfect. The project was shot on a mix of black and white film as well as digital (Nikon Df).

While fooling around one day with how to shoot a particular look I wanted on a particular day for that particular project a thought crossed my mind. What the heck, I’ll free-lens/lens-whack/poorman’s tilt-shift whatever you want to call it with the 28mm. Very cool I actually made a few shots that worked — at least I think they worked. Nice.

After working on a few more sessions for that project I had occasion to want the same look again. Free-lensing is dicey, it’s also got a very limited range of what you can do unless you take the lens apart to give you more leeway to move it around and achieve a focus strip where you want it at the distance you want it. Given I really didn’t want to bust up my little free gem of a 28mm another thought cross my mind. Hey, one of those thousand hunks of software I own has some sort of out of focus simulation feature that looked kinda nice when I played with it a year ago. I also remembered it had a few very specific lenses that were allegedly modeled in terms of the way they render. I tried it out to see if I could get a good match for the shots I did manually in terms of the way the simulation looked.

I’m not quite ready to show the whole project yet. A few contact sheets to cull yet, a few very hard decisions on which shots make the cut in terms of prints when evaluating the whole but here’s a StupidCrap™ version of my findings…

First up let’s take a look at an in-focus shot from the Nikkor 28mm f/3.5 shot at f/5.6

Next let’s take a look at the same shot but out of focus at f/5.6

And last let’s take a look at the sharp shot rendered out of focus using the simulation of a particular lens via software…

Just a quick and dirty job. I could probably get closer with another try but I’m lazy and who cares. Plus I already have experience matching them up from working on the project for the last few weeks. It’s so so close that I have no issue mixing shots between the lens-whacking I did in-camera on the Df and the strait-up shots I did on film and then simulated the focus-plane shift using software. Here’s the punchline… two punchlines to be specific.

Punchline #1: The lens simulation that turned out to be the closest just happened to be that of a Zeiss 28mm at f/5.6. Funny considering that particular lens I think they are referring to cost so so much.

Punchline #2: In most circumstances for this project I am using software designed to simulate film just for the lens blur simulation it offers on real film. I made no free-lens shots on film… to error prone, too expensive. Conversely I also confirmed a suspicion while messing with this that I have a new favorite in terms of grain emulation. I found it much easier to match my digital shots to actual film using this software than my previous go-to solution which was Nik Silver EFEX Pro 2.

What software is this? Well it’s Alien Skin Exposure 7. I had an inkling that I would prefer it for a few reasons in terms of grain simulation way back when I got the version 6 update. This is the first time I actually dove deep on a real project and tried to match up real film with digital shot for the same project.

Here’s a random sample from the project under consideration at the moment…

RB

blog comments powered by Disqus