KenwardV

An amateur/hobbyist here...

I take underwater shots using either a Nikon J2 or J3. A common issue I am having when processing the raw file is the obvious speckles of color, most easily seen in the background water in the example below. I have tried both darktable and Affinity.

Following is one such example. To get the color correct I adjust the WB, usually resulting in a large red influence as expected for UW photos with poor balance (no filters and I couldn't preset WB with a slate). The Nikon-produced image (the 1st one, without the "_01" extension number) does not show this to anything near the same degree.

While I speak of underwater shots, it is also seen in regular pictures (out here where the rest of us breath). The WB adjustment is not the same in those cases, but the effect is still there.

A link to the raw file is https://www.dropbox.com/s/4tj5vopxl36bs42/_DSC0312.NEF?dl=0

My usual approach (in darktable, at least) is to do a normal sharpening, adjust the base curve, do a spot WB (darktable seems quite good at this), and adjust saturation/exposure if I want to brighten things up a bit. This effect does not change if I remove sharpening, change the order of the first few steps, etc.

I'm not including the EXIF information, as that can be found in the files. This does not seem to be based on that. I have not seen such effect with my D750 images (no UW housing for that one yet anyway..).

I'm guessing that this has to do with the nature of the sensor, and some method to smooth that out?

Many thanks in advance for constructive comments! :)

(Download)


(Download)

Linda From Maine Loc: Yakima, Washington

One thought is you're attempting to sharpen what is murky, blurry and otherwise not sharp due to the natural conditions - kind of like attempting to sharpen something shrouded in mist or smoke. Do your editors allow you to make selective edits rather than global?

(added) In rereading your opening, you said you still have the color spots even if you don't sharpen, but I cannot duplicate them in my raw editor.

Also, it sounds like maybe you can't adjust the saturation of individual colors with those editors in order to lessen the red effect after initial adjustments. Is that correct?

Linda From Maine Loc: Yakima, Washington

At first I had the strong blue influence like Hamltonblue's below my entry in this thread.

So I played with WB in my raw editor, a less powerful version of Adobe ACR that comes with PS Elements. Then I upped the red saturation in PS Elements to bring out these colors. #2 is a screenprint of my edits. No additional sharpening beyond the default in ACR + the clarity slider (pushed all the way ).

(Download)


(Download)

Hamltnblue Loc: Springfield PA

Here it is with 2 levels of dehazing
Remember water will filter out different colors depending on depth.

(Download)


(Download)

rmalarz Loc: Tempe, Arizona

I'll take a look at this a bit later when I get home. It might be a simple fix.
--Bob

robertjerl Loc: Corona, California

When I download and enlarge I see a lot of "noise", those colored bits - is that what you are talking about? That is generally a factor in low light, high ISO shots. I will get the RAW and try to dehaze without getting so much noise.

OK, did a quick and sloppy edit:
A. Camera RAW 10.3
1. custom WB
a. upped color temp a lot
b. tint towards magenta
c. + dehaze
d. + clarity
e. + vibrancy
f. + saturation
g. + opened shadows
h. + contrast

Then over to PS
a. under "image" hit "auto color" once
b. under "image" > "adjustments" > shadows/highlights/midtones - upped all three settings a bit
c. under "filters" > " "smart sharpen" +50%

Then to NIK "Dfine 2" hot pixels filter

Saved as JPEG at Maximum quality and file size - here is the result. Oh, I deleted everything including record of adjustments and then emptied my Recycle Bin.
I suppose I could have cropped to get of the black corners.

(Download)

imagemeister Loc: mid east Florida

Here is my take using your first image - using my standard JPEG PP in Elements 9 and tweaking the colors a bit to what I think they should look like under the circumstances.....

..

Linda From Maine Loc: Yakima, Washington

Here is the full size image of my edits - no noise reduction, no de-haze. (the rest is detailed in my entry above, which I ran out of time to change, lol) There is a lot of "gunk" in the water itself.

To the OP: note the difference in the water color (upper corners) of mine compared to your edits. Hopefully someone can address darktable's WB editor or this comment about UW shots and red influence as I had to work hard to get any red: "To get the color correct I adjust the WB, usually resulting in a large red influence as expected for UW photos with poor balance."

(Download)

robertjerl Loc: Corona, California

I added my effort as a major edit on my first post.

Linda From Maine Loc: Yakima, Washington

Have you ever compared the same photo, especially online, using two devices? I wonder what colors each of us really sees in these

robertjerl Loc: Corona, California

I have always wondered if we all see the "same" colors. We learn them by being told "this is green..." so after that it is green to us. If we could transmit what we see into someone else's mind would it be the same that they see? I think I was sick and running a fever when I first thought about that.

Linda From Maine Loc: Yakima, Washington

rmalarz Loc: Tempe, Arizona

As promised, here's my shot at it.
--Bob

(Download)

selmslie Loc: Fernandina Beach, FL, USA

Any underwater photography will be progressively more deficient in red as you go deeper. Eventually yellows start to go as well. To make matters worse, it's not just the distance to the surface but also the subject distance that filters out the red. This makes underwater landscape photography very different from images taken in clear air - you just have to accept that they will appear to be predominantly blue-green.

You can't post process your way out of it and you are not going to get very far removing the haze from suspended particles in the water.

The only way to retain red is to shoot nearby subjects in shallow water or to bring your own lighting. If you add lighting it needs to be off camera by a couple of feet or the haze near the flash will get lit up.

You need to commit some resources. That may not be worth the expense unless you plan to do a lot of underwater photography.

Incidentally, your 10mm lens (27mm equivalent) behaves like 6.7mm (18mm) underwater. That's why the lens hood is cutting across the corners.

I added back some red to your image but you can see that the editor was somewhat indiscriminate.

(Download)

KenwardV

To all who have replied--I thank you for the alternative approaches! I'm in the middle of jury duty right now and have been unable to reply quickly to everyone's thoughts, but perhaps can summarize what I see here.

I'm well aware of the effect on light from shooting underwater. The fact that post production allows one to at least partially recover from that is a change from film to digital that some people should love! ;-) The shots I took that day were all with an INON UWL-100 wet lens with a dome, attached to the front port, effectively creating something close to a fisheye effect at 10mm. I used no strobes on that dive. The subject of this shot (the coral) was likely no more than a foot or so from the front of the camera. The ISO was 400, so not too high as well as I understand things.

When I made the statement that a lot of red was involved in that recovery, the following two screenshots from darktable perhaps show what I mean. The first is after application of a base curve, the second is after spot WB. Notice the shifts in T and colors. A higher relative proportion of green, and distinctly more red than before.

The issue with the spottiness is presumably due to my choice of steps taken, and the algorithms used by the tools. I can rationalize what happened with mine, since haze is whiter spots in a darker background--those spots becomes redder when adjusted since they have more red to begin with than the blue background. Being quite new with PP I am searching for other approaches to achieving the results. The different examples offered by people (thank you! :) show varying levels of achieving that balance.

I REALLY appreciate people sharing with me what they did to get their results!

Plans include attempting WB through a selective mask, and also dehazing first to minimize the haze "high points" in the background. Saturation may be a better way to take the last stage of white balancing, if I want to leave blue in the picture (in the corners). Any sharpening (standard step in the default) may be played with only as needed.