Bayou

Every RAW editor applies a different color profile to a RAW file when opened for editing...a starting point, if you will. Guess what starting point you get with a proprietary editor? Often it's substantially the same image you'd get with the default jpeg setting of the camera.

So of course the Olympus of Nikon software gives you a "truer" image right off the bat...true to the camera's jpeg engine. True to reality? What's that?

selmslie Loc: Fernandina Beach, FL, USA

The problem with this is that the concept of "original colors" is impossible to define.

For starters, what is recorded in the raw file has no white balance (WB) applied to it. That does not happen until the JPEG gets created. To make matters worse, the WB the camera used to create its JPEG is very likely to be "wrong" in the sense that something where you expect a neutral gray (JPEG red=green=blue) may not actually be neutral in the camera's JPEG.

So what does the proprietary computer software do with that questionable WB setting? It attempts to honor it by using the same setting for the JPEG it is displaying. And many non proprietary software programs do the same and call it "As Shot". The computer has no way of knowing whether the WB as shot was deliberate or accidetntal.

To make matters worse, the camera may use different default settings for saturation which also alter the appearance.

The photographer may try to mitigate this by adjusting the WB in the computer but trial and error may not achieve perfection.

To make matters worse, our vision adjusts to "correct" WB to accommodate small changes in lighting. So there is no way to actually judge a scene for ourselves and match what we remember to what the computer is displaying.

So while you might arrive at colors that are satisfactory for you, to call them correct or original is just not right.

selmslie Loc: Fernandina Beach, FL, USA

More accurate in what sense?

In all three examples, the neutral gray row seems to have been rendered very close to neutral (the RGB values seem to be within +/- one or two units). That did not happen by accident. It's unlikely that the camera came that close even if you set a custom WB to try and make this happen.

But the saturation of the "custom input profile" has been cranked up significantly. Several of the primary color squares actually have a 0 value for one of the primary or secondary colors (cyan, yellow, blue). That does not make them more accurate.

dbrugger25 Loc: Raleigh, NC

For many years, I ran a company that manufactured "Cast Stone Concrete" products. Those are custom products made using various casting methods that closely replicate the color and texture of real stone. In building restoration or renovation projects, we had to closely match existing stone. Some projects were hundreds of miles away and we couldn't send someone from our staff to derive an exact color match. Often the original stone quarry was closed many years ago. Sometimes the item required internal reinforcing which is impossible with stone.

We asked the contractor to tape a fairly new $20.00 bill to the existing stone using blue Scotch brand masking tape and to take a closeup photo in natural daylight, but not in direct sunlight. We could then process the photo in our design studio to the state where the colors of the currency, and tape were exactly correct on a calibrated monitor and printer. It worked and we almost never got it wrong.

Ysarex Loc: St. Louis

The colors in the photo of the scarf and of the color checker are closer to the colors of the objects photographed compared with the output from the generic profiles.

photonaut Loc: Albuquerque, New Mexico USA

I have amblyopia in my left eye and have had since birth. There is definitely a color shift between my right and left eyes. I've always wondered what life would look like with two perfectly functioning eyes. But my RAW files don't care what my eyes see so I can have as much fun as I want with the files in post-processing. :>)

selmslie Loc: Fernandina Beach, FL, USA

How would you know if you can’t measure the targets objectively?

Their color is a matter of reflectance that depends on a light source.

Everything would be different in broad daylight or open shade.

The way that light reflects from the fabric will be different from how it reflects from the matte squares or the plastic frame or the textured wall.

Ysarex Loc: St. Louis

Not interested in responding to stupid bait question today.

selmslie Loc: Fernandina Beach, FL, USA

Every wavelength in the visible spectrum gets past the red, green or blue CFA filters to some extent.

Under normal conditions, no colors can be saturated purely. That can only happen artificially (in post processing).

That's how I discovered that your "custom input profile" was artificial. It cannot be "accurate".

Ysarex Loc: St. Louis

And of course I never said that the custom input profile was "accurate." Are you trying to misrepresent what I said?

Blenheim Orange Loc: Michigan

Many, many years ago I worked in the auto body paint and supply field and one of the things our company did was custom color matching. The customer would bring in a fender or some other piece of the car and we would match the color, and the spray some of our match onto a small area of the part the customer had brought us. We got some funny responses from the customers. One customer looked and looked at the part said "I can't tell if it is a good match or not. I can't see where the original color stops and your color starts." Uh, right. That's what we call a match. Our favorite was the guy who said "sure the two colors look the same, but are they really the same?"

selmslie Loc: Fernandina Beach, FL, USA

We know what you said, "The custom input profile used in the middle image most accurately reproduces the color checker."

But we can't confirm your claim without measuring the squares. That's where I discovered that you set the white balance for the three examples (middle gran, third square top row): 1. RGB=129,129,131; #2 129,129,129; #3 137,138,137. Maybe you used a different gray square.

Notice that #3, NX-Studio shows significantly higher vallues

Then I checked the saturation (blue square second row): 1. 38,48,196; 2. 0,50,172; 3. 32,46,213. The same thing happened to the cyan square and the other blue value. Where did the blue values go? They should not be 0 or 1.

Ysarex Loc: St. Louis

And to begin that paragraph I said; "The middle one is considerably more accurate to the actual colors of the subject than the other two."

So I clearly did not claim that the colors produced by the custom input profile were "accurate." You didn't answer my question. Are you trying to misrepresent what I said?

burkphoto Loc: High Point, NC

Unless you’re a photo lab. Then you have to use proper color management. What goes on paper has to look like what’s on a properly calibrated and profiled monitor.

selmslie Loc: Fernandina Beach, FL, USA

My answer is crystal clear.

"The middle one is considerably more accurate to the actual colors of the subject than the other two."
"The custom input profile used in the middle image most accurately reproduces the color checker."

What's the difference? Both are just your opinion. You can't objectively demonstrate either statement.

I had no problem showing the flaws in your demonstration.