Marine Fish Tank Problems

[Redcrown]

My wife has marine aquariums. She got new lights, some exotic set up. I've photographed the old setup for her in the past. Lot's of problems back then, but I could usually come close to a color match. With the new lights, no can do.

Setting WB on a neutral gray in ACR drives the temp to the max 50000 and the tint to a large negative number, usually -150. And even then, the grays are far from neutral. More troublesome is that the grays are all over the place on their distance from neutral.

I shot the colorchecker under the new lights and tried to create a profile with both Adobe DNG Profile Editor and Xrite ColorChecker Passport software. Both programs fail, and refuse to generate a profile. DNG PE says it can't do white balance. Xrite says "Illuminant not supported."

I've tried to build a curve in ACR on top of various combinations of profiles and WB, and drive the RGB values to be equal. Usually such a curve clips off because the values are so far off. But when I do get one to work without clipping on one gray patch, the other gray patches are still way off.

What to do?

Here's a jpeg sample




Here's a DNG of the colorchecker shot.
https://www.dropbox.com/s/3ixmmg352cxh40t/FishFlrorescent.dng?dl=0

P.S. I haven't used Dropbox in a while and looks like they changed things. If that link does not work, please let me know.

[GWGill]

I shot the colorchecker under the new lights and tried to create a profile with both Adobe DNG Profile Editor and Xrite ColorChecker Passport software. Both programs fail, and refuse to generate a profile. DNG PE says it can't do white balance. Xrite says "Illuminant not supported."

Here's what I get using ArgyllCMS to make a low res cLUT profile, and then process the CC Passport to sRGB with it.



Profile is here.

[Redcrown]

GWGill,

Thanks for that effort. Looks great. Unfortunately, I'm not experienced with ArgyllCMS and I can't use the icm profile in my Adobe workflow (ACR). I did muddle through the steps to make a DCamProf profile (dcp), which uses ArgyllCMS in one step. But that profile failed, not any better than Adobe profiles.

What raw converter did you use to convert the DNG using your icm profile?

[TonyW]

While it would be good to get a camera profile sorted for this type of shoot perhaps if you cannot get other methods to work satisfactorily you may want to have a look at Raw Therapee as the colour temp adjustments have a much wider range than ACR/LR

Just brought your image into RT and maxed out on colour temp at 60,000 - not perfect I know but should get you somewhere in the ballpark I think

Thought it may be of interest

[daicehawk]

My wife has marine aquariums. She got new lights, some exotic set up. I've photographed the old setup for her in the past. Lot's of problems back then, but I could usually come close to a color match. With the new lights, no can do.

Setting WB on a neutral gray in ACR drives the temp to the max 50000 and the tint to a large negative number, usually -150. And even then, the grays are far from neutral. More troublesome is that the grays are all over the place on their distance from neutral.

I shot the colorchecker under the new lights and tried to create a profile with both Adobe DNG Profile Editor and Xrite ColorChecker Passport software. Both programs fail, and refuse to generate a profile. DNG PE says it can't do white balance. Xrite says "Illuminant not supported."

I've tried to build a curve in ACR on top of various combinations of profiles and WB, and drive the RGB values to be equal. Usually such a curve clips off because the values are so far off. But when I do get one to work without clipping on one gray patch, the other gray patches are still way off.

What to do?

Here's a jpeg sample




Here's a DNG of the colorchecker shot.
https://www.dropbox.com/s/3ixmmg352cxh40t/FishFlrorescent.dng?dl=0

P.S. I haven't used Dropbox in a while and looks like they changed things. If that link does not work, please let me know.

There's mixed light sources or a HUGE peak in the blue part of the spectrum in the setup I bet $100. Something like this: https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcQfU6wkoVPDEHNmvI2MBPTsOzhRflCaQECmuiIP6KIZA4rrM3am
Cannot do much about it, sorry.

[Redcrown]

Here's my solution, so far...

I tried my least favorite converter, Canon's DPP. Lo and behold, it made a fairly good white balance when clicking on gray patches of the colorchecker. ACR creates bizzare results.

The "neutral" DPP version, however, did not satisfy the fish people, and I agreed. How a marine aquarium looks to the human eye is not neutral. So I created 6 versions of an image. First was neutral, then 5 with increasing levels of blue tint.

I transferred these to a Samsung tablet so we could hold it up to the actual aquarium. My wife and a fish buddy of hers cycled through these versions until they agreed on one that best matched. That one is shown below. An RGB sample on the gray card in that image reads 89/134/174 !!

Strangely, all but 2 of the corals in the tank matched the image. The two that did not match were in the blue-magenta range. Yet to the human eye they looked very close to other corals that did match. Something about the fluorescence of corals comes into play, I guess.

[Jack Hogan]

You may want to contact Jim Kasson, he did a fair amount of underwater photography and he is very helpful.

Jack