Here’s a question I might know the answer to. Someone commented “I’ve never seen a cardinal that red before!”
If I use darktable modules like exposure, levels, contrast, highpass filter, levels and tone curve to make what my eye likes best how do I know what real is?
Perhaps guy (or gal) needs a color bar they shoot once or twice each day, in front of backgrounds similar to what you are working with. This is a 700 pixel version the image mentioned above.
I suppose you could look at embedded jpegs too. But jpegs are just raw images edited by algorithm. Whose to say that’s what’s real? Someday I’ll have to figure out how to profile my 32" Dell monitor I suppose. But even then what you get is what you edit.
There’s no way to perfectly convey a scene accurately without transporting someone in time to go see for themselves, so it’s up to the photographer/post-processor to communicate the scene, and the viewer to choose to believe it… or not.
Having pretty much no short-term memory (drives my wife nuts), I really rely on the colorimetric response of the camera in my hands, and the color primaries/white point in my raw processor to deliver an image for me to subsequently mung into something I like. In that endeavor, I haven’t intuited much variance from my oh-so-poor recall. So, what I’m chasing in my raw processing is a minimization of the operations required to present the starting image: incorporating white balance correction in the working profile transform, minimizing stacked tone transforms, that sort of thing.
If I were to take up an imaging endeavor that required a consistent representation of color to what most would perceive from the real scene, I’d be incorporating a color target and session-specific profiles in my workflow. My use-case consideration would be something like catalog illustration, where the following customer exchanges is to be avoided: “That’s not the color I ordered…”…
I think there is a real difference in approach with photography versus photoillustration. If I do something to highlight what I think is important in the picture, then I’m traversing the divide. That isn’t to say a photojournalist wouldn’t be beyond a few seconds of dodge and burn in the darkroom for technical reasons, but even there if it’s overdone then the line has been crossed.
Let’s just say colors does not actually exist. What you see is different than the closest person next to you. A slight difference in color perception exists between you, the camera, and the person next to you. There’s no such thing as orange, or red. Not to mention optical illusions can show even the two shades are perceptually different.
Like for this example -
There’s also tools to try to callibrate colors, but even that isn’t perfect.
I should have qualified my original question. If we’re making art anything goes. In the special case of wildlife (and bird) photography most viewers want an image most closely matching what the eyes actually saw. But what did the eyes actually see? Perhaps there is no easy answer. Perhaps color cards the best tool for those special situations.
To my understanding, an attempt to address your problem is colour management and ICC profiles. In principle, it should make a scene illuminated by sun light and viewed by the (average) human eye close to what you see on a calibrated monitor – provided all components in the sequence are correctly calibrated.
All this will not work for different illumination, different detectors …
Editing a (calibrated) image is a completely different story from my point of view.
I bet that your eyes saw something quite different from what I saw.
Wouldn’t it be better to ask for a result that mirrors the way the colours actually were?
As an example, take PMS 300; once upon a time known in the trade as Volvo Blue.
If you shoot such an object, of course you want it to look the same on screen as well as in print (on glossy stock as well as on newsprint [shudder!]).
I don’t think that’s the case. If you take a look at the best World Press photos, you will immediately notice that most of them have been very heavily edited. Especially in photojournalism, I think it’s important to have the skill to highlight the most important thing. Besides the “right” selection of the motive, the processing plays an important role if it serves to highlight that information. For me, photo manipulation is the line that should not be crossed.
In my opinion, the same thing I mentioned above also applies to wild photography. Already the choice of angle, depth of field and background determines how and what the viewer sees. For example very strong color contrast between green background and bright red feathers of the bird in the photo above, already distorts the color perception as the red stands out.
The observer would have a completely different impression if the bird were standing in front of a background with a warm color. So here is also true, in my opinion, it is not important what you actually see, but what you want the observer to see.
Here’s a good way to reach true colors. Comes in handy when you shoot nature.
First, you know beforehand what colors stomps, grass, logs, asphalt have, right? You know, for example, asphalt is gray. You learned the color of asphalt early in your childhood. It’s common and everywhere the same, thus it can be used as a reliable reference - a signpost, lighthouse to guide you to true color.
Get that artistic BS out your head, there is no artistic perception if WB is incorrect.
For example, the stomp on the image has strong
red/magenta cast. Is dry wood supposed to be magenta? I don’t think so. . . That’s why the cardinal is so unusually red. In fact, it is not.
Hence the tempreture of light falling on the stomp. Midday almost white, dusk/dawn - reddish and so on. Consider it changing a tint of grey. It was obviously white, not magenta cast on the stomp, so no cast.
This is where you start. Once the colors are done as they are supposed to be, not to look. You can add as much artistism as you want.
This specific image won’t look correct even if the colours were confirmed to be spot on because the red is blown out losing the texture. Even a B&W with all the textures there could be perceived as ‘how you saw it’.
If you put the raw up for PLAY_RAW, you’d see a host of different takes on it but most of them would probably look right.