6

The clipping indicators in Lightroom show when a channel or all channels are clipping in either the shadows or the highlights, but I've seen them used in two ways:

  1. Use them as an indicator of when you have a pure white / black so that for a full tonal range both shadow and highlight clipping indicators will be showing (although perhaps only a tiny number of pixels will actually be clipped.

  2. Use them as an indicator of when you have gone too far, so when a clipping indicator shows, the user backs off slightly, so for a full tonal range the white and black points will be just short of triggering the clipping indicators.

So which is correct? Do they indicate a value has been pushed beyond maximum (or minimum) or that a value is at maximum or minimum?

Note: This isn't a question about what clipping is or whether in some circumstances clipping is OK. All I want to know is exactly what the meaning of the clipping indicators are in Lightroom. Do they mean at or over.

3

When looking to set a maximum black or white point, both scenarios you mention can be regarded as correct or incorrect as they both provide you with an indication only.

If developing to export in jpeg for displaying on monitor only, then the triangle in the histogram turning to white will not have any noticeable affect on your image.

However, if you were developing to print, then any pixel that is blown out, will not be printed. The printer cannot print white and this will be immediately noticeable on paper.

Going back to the histogram, the clipping happens in stages, and if you look at the indicator, before turning to white, it can turn, Pink, Yellow, Cyan etc indicating which channel clips first.

with certain types of images, EG, fine art prints, textiles or jewellery, it is imperative to understand that this process of the first channel clipping can start when the RGB pixel value reaches 242/245 but the triangle will not turn white till 255, at which point, there is a loss of detail.

The way to correctly judge the pixel values of the image is to turn on Soft Proofing in LR. this will then allow you to hover over your image and correctly see the separate RGB values under the Histogram chart and as a result, help you manage your black and white points more effectively to ensure there is no clipping.

Therefore, the answer to your question which is correct?

Scenario 1 can be used when developing for viewing on monitor only

Scenario 2 is better suited when developing for print.

And finally, the answer to your question, do they mean at or over, the answer is over. If you wait for the triangle to turn white, then you have already lost some detail.

2

I finally got what I think is a solid answer from trshaner on the Lightroom Forum which I'm adding here:

Well you can test this for yourself. Here's raw image uniWB test shot with no raw clipping pushed to show clipping inside LR. The RGGB levels of the UniWB raw file are all equal, but after the camera profile is applied the RGB values inside LR are at different levels (pink color). The first screenshot shows the levels inside the clipped area and below outside the clipped area. The actual "trigger level" appears to be at 99.8% for any one channel, which is a 254.49 value for 8-bit data (255 max.). Edit in PS shows 255 in the clipped area and 254 at the border of the clipped area.

So it looks like to me the Adobe Engineers want the Highlight indicator to trigger when any one channel hits the wall.

enter image description here

  • Are these values in ProPhoto RGB color space? (the working color space of Lightroom) – Ben-Uri Oct 31 '16 at 21:16
1

Maximum is the highest value possible. There is no way to record any information beyond the maximum value. Minimum is the lowest value possible. There is no way to go beyond either of them. For a 24-bit JPEG there are 8-bits per color channel. Eight bits allow 256 discreet numerical values which we refer to as 0-255. When a value is at 255 we say it is saturated.

For pure white the value of all three color channels must be at 255. For pure black they are all at 0. Various shades of "pure" gray have values that are the same in every channel: (85,85,85) is darker than (175,175,175), but they are both a shade of neutral gray.

One reason exposing for the highlights is critical with digital photography is that there is a hard line at whatever the maximum value is. For an 8-bit image the maximum value is (2^8)-1=255. For a 14-bit file the maximum value is (2^14)-1=16,383. If a point in the scene contains twice as much red as it does green or blue, then properly exposed it may be expressed as something like (240,120,120). But if the scene is exposed with a higher Exposure Value so that green and blue are fully saturated as well the value will be (255,255,255) and not (510,255,255) because the camera can not differentiate anything brighter than 255 from anything as bright as 255.

  • Thanks for replying, but interesting though your answer is, it is really just generic advice about clipping and you fail to answer my question. Which is the correct way to use the slider? 1) or 2)? – Undistraction Oct 16 '15 at 20:20
  • Which way to use the slider depends on how you want the image to appear. There are times, such as when the sun is in the frame, that you probably want a fairly significant portion of the pixels in the image to be totally saturated (clipped). At other times you may not want any pixels to be near a fully saturated value. – Michael C Oct 17 '15 at 3:59
  • I'm not asking about creative use of clipping, just what the clipping indicator showing actually means. It feels like you haven't actually read my question. Only the (original) title. – Undistraction Oct 17 '15 at 8:29
  • "f a point in the scene contains twice as much red as it does green or blue, then properly exposed it may be expressed as something like (240,120,120)." For the purpose of your answer this statement is ok, but in reality, you need to consider gamma, which is "compressing" the low and high values to prevent clipping and capture a higher dynamic range. – agtoever Oct 17 '15 at 9:43
  • Yes, it is over simplified because the output from most sensors is linear and either 12-bit or 14-bit. But it does make the point that if the least represented color is over saturated the other colors are not going to be saturated even more - they'll all be equally saturated. – Michael C Oct 17 '15 at 16:36
0

The answer is that they are at the max or min.

They are right at the clipping boundary, but for digital imagery that edge happens to be inside the last bin, so to speak. All the program can really do is mark for you where the pixels are that sit in that last bin, either at fully on or fully off.

In other words if we just consider one color channel there is a bin for fully off and there is a bin for fully on and there several bins in between (254 for 8 bit images). When you perform some sort of an operation on the image the values that go beyond that bin, get stuck in the bin.

So then, when a particular pixel is black or zero the question is: was it calculated as black after shift or was it calculated as beyond black? That's why you've reported on two possible techniques. How confident are you that those "clipped" pixels have only just been put in that limiting bin? Odds are, at least some of them are actually clipped. Well personally, I would be okay with a few pixels "clipped", but others like to pull back just a tad so that none are.

-1

They indicate that a value has been pushed beyond a maximum or minimum. However, that extreme value is not the actual maximum 100% or minimum 0% value, but slightly above/below that.

After some tests in Lightroom I think it works as the following for highlights:

  • LR checks if the average value (maybe the RMS value?) of the channels is beyond 99.5% of the maximum range (i.e. 255)
  • It then marks that area as "clipped"

So, for highlights LR marks everything that exceeds the 99.5% threshold, for shadows everything below the 0.5% threshold.

  • So you believe that if the clipping indicators aren't showing at either end, then 1% of the tonal range is not being used? – Undistraction Oct 16 '15 at 13:39
  • My simple tests in LR seem to show that. – Saaru Lindestøkke Oct 16 '15 at 13:47
  • I haven't used LR in several versions. Can you not set the clipping indicators to whatever values you wish? You can with Canon's Digital Photo Professional. – Michael C Oct 16 '15 at 18:00
  • 1
    @MichaelClark I did not find such a setting in LR. – Saaru Lindestøkke Oct 16 '15 at 19:53
  • 1
    @Bart, assuming an image with 8 bit color channels: 255 (full on) ÷ 256 (the full range of values) is actually 99.6% (greater than 99.5%) while 254 ÷ 256 = 99.2% (less than 99.5%). So its really only showing the pixels that are fully on. – Octopus Oct 16 '15 at 20:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.