Original Version from the camera.
A greenish wall of construction material with barbed wire in the foreground.
I've spent a lot of time this week reading about and researching topics around A.I. and photography. There is not just one slippery slope to contend with, there are many. While I'm opposed to creating fantasy images out of a combination text descriptions coupled with the endurance of being able to hit the "generate" button some 2,000 times before selecting a machine proffered, computer-created picture that one likes ( his practice according to an A.I. poster of such images on Instagram), I'm also reconsidering where one draws the lines/boundaries in something as simple as image tweaking. Or its more complex sounding brother: post processing.
It's important to remember that even digital cameras set to all neutral or standard settings are still a design team's working idea of what a digital image should look like and are just two dimensional representations of the reality existing in front of the camera. Move from RAW file to Jpeg instead and the ability to introduce these kinds of subjective "tweaks" goes up dramatically. But where is the line between "sin" and "salvation" for a typical image that's meant to be shared with an audience?
I walked by this elevation of a building project when I was over near the UT campus yesterday afternoon. I liked the color of the wall but the added attraction which actually stopped me and convinced me to take a photograph was the barbed wire running across the bottom fifth of the frame. When I got back to the studio and looked at the .DNG file on my computer the overall image seemed flat and lacking the "pop" I experienced when I initially saw it. I'm sure the top photo is more accurate a representation of what's there but if you consider that I was wearing my polarized, anti-reflection coated, prescription sunglasses at the time of the reveal you might understand why my memory of the wall doesn't quite jive with the reality of the wall. At least from my point of view. Probably a good reason not to wear your sunglasses around when you are out photographing....if you are looking for faithful concordance with the "facts."
Here's what I did to make the image match what my brain first saw when I turned the corner and happened on the wall...
The first thing I did with the RAW image was to go to the lens correction menu in Lightroom Classic and look to see if the lens I was using was represented. I scrolled down to the Voigtlander submenu and a then scrolled through a couple dozen profiles to find the 40mm Ultron f2.0 SLii Aspherical profile and I selected that. On the screen you could see the building straighten out and the gentle barrel distortion get corrected. Look closer and you could see the vignetting in the corners brighten and the corner details tighten up. The image also seemed brighter overall; a result of subduing mid-frame vignetting perhaps.
My next step was to see what the noise reduction menu would serve up. I wasn't so interested in noise reduction as I was with another selection in that submenu; it's called: "Raw Details." This uses some degree of A.I. to bring more observable fine detail to the files. Yes, it's perhaps technically A.I. but not sentient by a long shot. I think the extent of the feature's intelligence is to look at the basic structure of the file and make assumptions about how sharpness gets robbed in the process of going through a camera's circuits and internal processing and then trying to reverse that sharpness and detail degradation by interpolating using existing structures and then augmenting them with presumed "lost" data.
It seems to work very well but the effect is subdued. At least in the files I worked with. Why not just click the noise reduction option? Well, the file was shot in bright light at ISO 50 and a close observation at 100% didn't show me any reason to fool around with noise reduction in what was basically a noiseless file.
My next step was to apply a "user preset" I've engineered for Leica SL 601 raw files. It boosts the white slider in the basic control panel, also increases shadow exposure, adds a +20 to the clarity slider awhile adding +20 to the vibrance control and +20 to the saturation slider.
Most of these settings are "canned." Which basically means I've processed enough similar files to establish a look that I like. Not a neutral look, or even a more accurate look, but a look that conflates with my own subjective appraisal of what a generic, sunlight scene needs to have added in post production to make it look the way my brain wants stuff to look. Call it my "Pollyanna Profile" since it is almost inevitably brighter and more colorful that the camera's initial point of view.
If I was incredibly busy, day-to-day, and a lot of my work involved endless post processing of similar files, I would probably spend a bit of time creating individual presets for not only every camera's raw files but all manner of subject types and lighting conditions. Examples: Portraits in open shade. Portraits in full sun. Portraits in artificial light interiors. Portraits at dusk. Architecture on cloudy days. Architecture on sunny days. Architecture at dusk, etc. etc.
While I bump up against a prejudice (mostly from engineers and people firmly and fully invested in some form of linear logic...) that states that we should all be invested in creating photographs that are as close to the objective reality of a scene as humanly possible that's really not for me. Sorry, that's not part of my "religion" as a photographer. I know that we're "interpreting." That's where the fun lies...
I think my work should match my subjective intention for a scene instead. But where are the edges of the slippery slope and at what particular angle does it become dangerously out of whack? I guess that's for each of us to figure out in our own work. The steeper the slope the more exciting the journey; but the same thing that makes the journey exciting increases the chances of peril.
Just thought these samples might give you some insight into my point of view. Thanks.
"Lightly" processed wall. To my taste.
Camera and lens?
The fabulous Leica SL (601) and the Voigtlander 40mm f2.0 Ultron ZF.2 SLii.
And here's one more, just for grins:
Advertising aimed at anyone who might want to shave their armpits.
When writers or bloggers or vloggers write about processing I think
it's only correct to supply samples of their own work when
explaining their choices. Just pontificating isn't as
instructive.
New Eyeglasses arriving Wednesday. Can't wait.
close shave....
ReplyDeleteI have to agree with your point of view on this, Kirk. As for myself, I have found that, in recent years, I have been deemphasizing " post production" - perhaps without consciously realizing it.
ReplyDeleteAt any rate, I currently have no real photo-processing suite - certainly not the way I used to. Lightroom is gone and so is Capture One. All I have is Apple's Photos app that allows me to crop, adjust exposure and maybe tweak colors and tone. I just try to get it right in the camera - with varying degrees of success.
I don't believe this is the result of some photographic "religion." But perhaps I'm fooling myself. There's no doubt that a lifetime in broadcast journalism has had an impact on my creativity and the way I see the world. I find that I experience a bit of cognitive dissonance if I don't try to keep things "real." I also find that, as a journalist, the further we descend into the current "post fact" and "post reality" era, the more frustrated I become. Perhaps that will change in retirement. Perhaps not. But AI and "augmented reality" aren't helping.