Anyone else found this problem with the WB pipette in PL8?

I have an image with pure white bands, more than likely reflective.

Here, I have just reduced the maximum luminance on the L tone curve to 252 to cancel the over-exposure marker…

Then I simply apply the WB pipette to one of the pure white bands…

What on earth is going on here? I have never had this happen before.

I’ve not seen that before.
I tried my 8.1.0.35, reduced the L to 252, activated vinetting and did some WB but it behaved as expected.

edit: whats the RGB-values for those reflective bands?

My guess would be that the reflective strips aren’t actually as white as they appear. Could they have a colour cast hidden within their bright appearance? Or could the reflective material they are made from be tricking the pipette tool into making an incorrect reading?

Have you tried taking a WB from one of the grey areas in the scene? I know Nikon (I can’t say for other camera manufacturers) recommend in the camera user manual to use a grey card to set white balance rather than a white card.

251 after using SmartLighting Spot Mode.

Here’s another one, this time, the whites are definitely not reflective…

Original…

Pipette…

It seems to be fine if I place the pipette on an area that isn’t already equal RVB levels.

Except it doesn’t equalise the RVB levels perfectly

Edit: Yup. It really doesn’t like trying to balance something that is already balanced - it applies a mucky greenish cast.

1 Like

Your tint is doing crazy.
What happens if you correct that?
I’ve no problems.

George

Yes, right down to black. As long as the RVB levels are equal, applying the pipette seems to change the cast to something else.

Well, of course, it changes but it rather negates using the pipette to get it right.

I’ve no problems. But I’m on windows. And you use the Mac. Maybe…?

George

The green tint looks like what I get when I expose with UniWB.
The WB temperature and tint settings look like UniWB too.

PhotoLab 6, 7 and 8 get the Hulk when I WB from a blown highlight…and bringing down the values with the tone curve does not help.

does the same thing with PL7, can’t use white to set the white, it just doesn’t work and shift colours every single time. its been broken for quit some time as far as i can tell.

1 Like

Ah, now he tells me :roll_eyes::kissing_heart::teddy_bear:

White balancing a photo that is overexposed to the max results in a UniWB setting - if the camera accepts to WB from blown white.

Not sure if we can call PhotoLab’s behaviour a bug, but DxO could warn users about WBing such an area.

1 Like

Everything that I have read about WB is that you take your reading from a grey area rather than from a bright white. I actually never use the pipette as it very seldom does a good job.

I mostly use the R, G, and B channels in the curves tool to do my white balance with the added advantage of being able to see how the histogram behaves while making your changes. You can watch the peaks for the individual channels and when they align then your WB is pretty much done. I mostly just adjust the white point on each curve to move the histogram peaks.

This technique works particularly well on underwater photos where you have a strong blue cast and you need to add RED to bring it back into balance.

3 Likes

Interesting method.
What value do you put in white balance tool before adjusting with curves ?

I let PL set the WB to “As Shot” so that I can control initial WB in-camera.

1 Like

Thank you so much for sharing this. I do find that the picker does a serviceable job, if I choose a good gray tone, but using this to fine tune it makes the process very intuitive and very much fun.

1 Like

I’ve seen such wild colour casts often. The trouble with choosing grey is that it is very difficult to judge what will be neutral. Whites and blacks are far more likely to be truly neutral.

So it does, I never realised that until now. Oh well, “Forewarned is forearmed.”

I really dont know, but can it be that your camera / files for that area really is something like: R:255 G:252 B: 255 ? (The wells in your camera sensor are not filled simultaneously in the 3 channels.) In that way it’s not white but looks like it’s white…

I can reproduce it, on windows. Put the eye dropper on a clipping part of the image and the image is getting green. I think that’s a wrong way of using the eye dropper.

George