Inspiring

Answered

Trouble (and confusion) regarding shadow and highlight clipping

Forum|Forum|2 years ago
March 14, 2023
7 replies
3824 views

I am using LrC 12.1 on an M1-based MacBook Pro. I have a photo of birds on a beach (a .tif file produced by Topaz Photo AI from a .jpg file). In-camera exposure is just a bit off, but very little is "blown". I have both shadow and highlight clippnig indicators turned on. The shadow clipping indicator is gray (no shadow clipping?) and the highlight clipping indicator is white (highlight clipping on all channels?). Looking at the entire photo ("fit" in Navigator), it is easy to see areas of red pixel indicating highlights are clipped; this is expected. However, again looking at the entire photo, I can see no blue pixels to indicate shadow clipping; this is what I would expect.

In the Basic panel, I can adjust highlights left and get rid of highlight clipping. The indicator goes to gray, and no red areas are visible when looking at the entire photo. However, when I zoom to look at one area in the photo (a bird head) I find a bunch of blue pixels indicating shadow clipping! Not what I expected! I took saturation to -100 to make sure the pixels were really indicators rather than produced by the photo. I do not understand why these indicators occur in the photo, but the overall shadow clipping indicator indicates nothing! Do I misunderstand the meaning of the overall indicator?

I tried to get rid of the alleged shadow clipping by moving the shadow slider to +100. As far as I could tell, nothing changed; all the blue pixel indicators remained! I then also moved the black slider to +100 and still nothing changed! I've only processed around 1000 photos in LrC, but I've never seen this behavior before. Are my expectations wrong? Could this be a bug, or am I doing something wrong?

I experimented; I increased dehaze a bit and increased exposure a bit. A a result, the shadow clipping indicator turned blue, and more blue pixels appeared near the bird head. Once again, nothing I did with shadows or blacks made the inidicator or pixels go away.

I experimented some more, in this case attempting to get a more pleasing result. I reached a state where the highlight clipping indicator was gray, but I found red pixels in the photo indicating highlight clipping! I was able eliminate the red pixels by futher reducing highlights and whites.

I have to add that at one point, I created a combination of tone and presence sliders with which I could not get rid of the red pixels in the photo indicating highlight clipping no matter what values I used for the highlights and whites sliders!

Again, all this is unexpected behavior! Am I expecting the wrong thing, or is there a problem with my photo, my computer, or is there a possible bug, and I found a corner case that exposes the bug?

Any help appreciated!

macOS

This topic has been closed for replies.

Correct answer TheDigitalDog

In-camera exposure is just a bit off, but very little is "blown". I have both shadow and highlight clippnig indicators turned on.

By @Gregory25556071f00h

If you're shooting raw, nothing in Lightroom Classic or any Adobe converter tells you about what actual exposure has clipped, how far you are from clipping etc. It's only showing you the current rendering using LR's tools. Plus, there is some minor clipping reconstruction in the engine (if one or two channels is clipped but not the 3rd). So the first question is, do you really want to know what's happening with the raw data based on actual exposure, or are you only concerned with clipping the processed data?

Also, there's blown channel(s) due to saturation and fully blown data due to either exposure or processing. One report is based on the color space you'll end up with; the other again depends on the rendering, not necessarily the captured data.

I see zero reasons to be concerned with clipped shadows (sometimes, a true black makes an image look better). Clipping highlights is kind of up to the image creator. Some images need a spectacular highlight or it looks 'dull'.

S

Shanice319616537sd6

Participating Frequently

Found it! You have to turn off the GPU: https://community.adobe.com/t5/lightroom-classic-discussions/p-artifacts-resembling-clipping-odd-pixels-in-some-areas/m-p/13877381/page/2

Go to Lightroom classic - preferences - performance - Use graphics processor: turn off atleast until you need to turn it back on

Rob_Cullen

Community Expert

@Shanice319616537sd6 FYI Your problem is "GPU" related- causing colour artifacts. The theme of this forum thread is about "Clipping"- an entirely different topic.

Regards. My System: Windows-11, Lightroom-Classic 15.1.1, Photoshop 27.3.1, ACR 18.1.1, Lightroom 9.0, Lr-iOS 10.4.0, Bridge 16.0.2 .

S

Shanice319616537sd6

Participating Frequently

Hi @Rob_Cullen, Yes mine is to do with a bug affecting the GPU, I may have got that wrong but based on the last part of the post I had thought maybe she was having the same issue. Anyway glad its not been an issue since!

S

Shanice319616537sd6

Participating Frequently

Just to say, your not crazy! I'm an experienced Lightroom user and this has been an issue for me since the last update, I found a setting in a forum that removed the issue but it also affected the images so I changed it back but it's now become so much of an issue I need to change it again so I'm trying to refind the forum. I'll let you know if/when I find it!

TheDigitalDogCorrect answer

Inspiring

In-camera exposure is just a bit off, but very little is "blown". I have both shadow and highlight clippnig indicators turned on.

By @Gregory25556071f00h

If you're shooting raw, nothing in Lightroom Classic or any Adobe converter tells you about what actual exposure has clipped, how far you are from clipping etc. It's only showing you the current rendering using LR's tools. Plus, there is some minor clipping reconstruction in the engine (if one or two channels is clipped but not the 3rd). So the first question is, do you really want to know what's happening with the raw data based on actual exposure, or are you only concerned with clipping the processed data?

Also, there's blown channel(s) due to saturation and fully blown data due to either exposure or processing. One report is based on the color space you'll end up with; the other again depends on the rendering, not necessarily the captured data.

I see zero reasons to be concerned with clipped shadows (sometimes, a true black makes an image look better). Clipping highlights is kind of up to the image creator. Some images need a spectacular highlight or it looks 'dull'.

Author “Color Management for Photographers" & "Photoshop CC Color Management/pluralsight"

G

Gregory25556071f00hAuthor

Inspiring

Quite interesting. Makes it clear how little I know about photo processing.

I think the bottom line is that I really don't care much about the gory details of RAW exposure, etc. or how the tools represent the data. What I really care about is how a processed photo will look when it is "presented", be that electronically on a web page or (mostly) printed on some medium. It seems that all the great observations/advice in the responses to my original post suggest roughly the same thing: don't worry so much about what the tool indicates, worry about the results you see, or "don't sweat over the small stuff", instead "look at the big picture".

Thanks!

Conrad_C

Community Expert

In the Basic panel, I can adjust highlights left and get rid of highlight clipping. The indicator goes to gray, and no red areas are visible when looking at the entire photo. However, when I zoom to look at one area in the photo (a bird head) I find a bunch of blue pixels indicating shadow clipping! Not what I expected! I took saturation to -100 to make sure the pixels were really indicators rather than produced by the photo. I do not understand why these indicators occur in the photo, but the overall shadow clipping indicator indicates nothing! Do I misunderstand the meaning of the overall indicator?
By @Gregory25556071f00h

If by “overall indicator” you mean the clipping indicator when the picture is at Fit magnification, and the weirdness is about seeing clipping indicated at one magnification and not in another, I think that might be a consequence of how any view other than 100% has to do some interpolation of pixels if enlarging magnification, or averaging of pixels if reducing magnification. If that’s what you’re seeing, I have seen that variation in clipping displays too, and my theory (since I don’t know the exact programming answer) is that small areas of clipping that are visible when zoomed in, might average out to no clipping when zoomed out.

For example, if a dark 4 by 4 pixel square of the image includes 1 pixel at RGB (0,0,0) (clipped to black), 2 pixels at RGB (2,10,3), and 1 pixel at RGB (3,10,4), then when you are zoomed in to 100% magnification or higher, the one RGB (0,0,0) pixel will be displayed blue to indicate clipping; but when zoomed out to 50%, the four pixels must be represented on screen by one pixel, so the display value of that one pixel might be calculated as the average of the four (I am not sure what math is actually used), and that average RGB value is above clipping, so naturally, it isn’t displayed as clipping.

If I’m right about how this works, then the cause and consequence is based on the same dilemma (having to merge display of multiple pixels when zooming out) that leads to the advice that it’s better to evaluate sharpening and other detail-oriented options at 100% magnification instead of zoomed out. Because again, zooming out forces multiple pixels to be merged together on the display, preventing you from seeing detail edits at full resolution.

And also if I’m right, then maybe you want the application to display a clipping indicator if any pixels are clipped in a group of pixels that is represented as one pixel at a zoomed-out magnification. That might be a feature request to file in the Ideas section of this forum.

G

Gregory25556071f00hAuthor

Inspiring

What you describe sounds plausable. It certainly could be an explanation for what I've seen. Thanks!

GoldingD

Legend

Hold down the Alt key while moving shadow or blacks. This will give a visual (mask)

might be getting my WIN vs MAC Alt key wrong.

And I agree, do not get rid of every bit of clipping at least on the black end (some might on purpose if printing as to avoid pure white)

https://lightroomkillertips.com/three-things-use-option-alt-key-lightroom/

G

Gregory25556071f00hAuthor

Inspiring

I did use the 'Alt'/'Option' technique before the original post; same results. Now I know that knowledgable folks don't see my aledged problems as real problems. Thanks!

J

Jao vdL

Community Expert

You really shouldn't try to eliminate every single clipped pixel. That's a bit of a fool's errand. You generally don't want huge swaths of blown out areas but some pixels here and there or even some little blocks is not a problem. In the end all that matters is what the image looks like, not that the algorithm says it is blown. You can have blown areas that look just perfectly fine.

G

Gregory25556071f00hAuthor

Inspiring

Your response seems consistent with an earlier response, which is gratifying.

I seem to remember a few months ago reading something about blown areas causing problems when printing photos. Is that too something about which I should not worry?

Thanks!

Conrad_C

Community Expert

I seem to remember a few months ago reading something about blown areas causing problems when printing photos. Is that too something about which I should not worry?
By @Gregory25556071f00h

It can be the opposite. Areas that are obviously clipped on screen, where every pixel can be seen clearly especially when magnified, might be less obvious in print, because of various reasons such as the lower dynamic range of print, paper texture that can hide the loss of detail, and the fact that prints are not normally inspected as closely as a magnified screen image.

The Photoshop expert Blake Rudis covers this in one of his videos:

Stop obsessing over noise and blowouts

Before commercial printing started using ICC color profiles, it was common to edit images by setting a white point and black point that were slightly in from the ends, which resulted in slight clipping. I think this was to make sure there was enough contrast within print dynamic range, and at the high end to let specular highlights blow out. Even today, the default values for auto correction in Photoshop, Lightroom Classic, and scanning software result in slight clipping.

Because (you might know this already) that’s another thing to watch out for. If there are true specular highlights in the original/raw image, they should be allowed to blow out because they have no detail to reproduce. Unnecessarily preventing specular highlights from clipping would lower image contrast, in some cases possibly enough to drain the “snap” out of the image.

And even though I like to preserve shadow detail, I do let some small dark areas clip if I think they really are supposed to be absolute solid black. But I sometimes choose to not clip shadows at all in, for example, a landscape where I don’t think any of the plants or dirt should look absolute solid black.

JohanElzenga

Community Expert

Dehaze sometimes does clip shadows, and because it is an algorithm that is recalculated when you move other sliders, it is possible that you cannot get rid of these clipped pixels no matter what you try with these other sliders. If it's just a few pixels I wouldn't bother. Look at your photo and if you like what you see. That is way more important than some blue clipping indicators. But if you do not like what you see, then you'll probably have to dial down dehaze.

-- Johan W. Elzenga

G

Gregory25556071f00hAuthor

Inspiring

Ah, great advice. As a relatively new LrC user, I'm still learning. My problems appear to fall into the categories of "incorrect expectation" and "lack of experience". Thanks!

Sign up

To post, reply, or follow discussions, please sign in with your Adobe ID.

Sign in to Adobe Community

To post, reply, or follow discussions, please sign in with your Adobe ID.

Scanning file for viruses.

This file cannot be downloaded