Is the histogram that's shown on a DSLR different from the histogram of the same image shown in Lightroom?

Question

I was taking some images and experimenting with the ETTR approach. I was trying to take pictures with the exposure setting that would have the right end of the histogram just touching the border on the right. But the histogram shown in DSLR and the one of the same image in Lightroom seem to be different in terms of whether the rightmost end is touching the border or not.

Is this normal? Why does it happen? Which one is more accurate?

I used a Nikon d5600. Also, i am shooting in RAW, if that makes a difference.
Also i am not talking about the LR histogram after making adjustments. That would obviously be different. I am talking about the in camera histogram vs the histogram in LR BEFORE making any LR adjustments

Answer 1

There are many reasons why they may be different.

• Many cameras will show luminance histogram clipping if any one of the color channels is clipped.
• Many cameras will show a color channel as clipped at a level of around 245-250 (i.e. early).
• Some camera's histograms just lie pretty badly (my D800 did)...
• The camera's histogram is based on the camera processed jpeg and is highly unlikely to match the raw file data (if recording/editing raw).

The histogram in LR is almost certainly more accurate than the one in your camera.

---

Edit

And interestingly, LR's histogram may not match the camera's histogram for a jpeg image either (reasons 1-3 above).

Here is a D850 jpeg image review on the camera.

It shows all three channels and the combined luminance as touching the right side (clipping) and the highlight warning shows a large area on the upper chest as clipped for all four (R/G/B/L).

This is the same jpeg w/o any edits in LR.

The histogram doesn't show any luminance clipping (short of right edge), and if I hold alt while selecting the exposure slider it shows me that the small clipping warnings are for the blue channel only.

And the raw file histogram looks somewhat different as well (as expected). You can do some things so that the camera histogram(s) and image review more closely matches the raw file histogram(s); at the expense of less usable jpegs SOOC. I use the "neutral" camera profile with reduced contrast (minimum) and brightness (-1) settings so the image review is closer to a raw image in LR with the same "neutral" profile applied. You can also go down the path of using uni-wb in camera; for totally useless jpegs, a lot of hassle, and a bit better histogram match.

But none of that is absolutely necessary... all you really need is to develop a good idea of the difference between what your camera typically shows vs the same raw/jpeg file opened with defaults (and it's highlight recoverability).

---

Many who shoot ETTR record raw files; push the in-camera histogram to where it is just showing clipping (slightly climbing right side; at least touching), and rely on the ~1stop of recovery capability. But you have to experiment with your camera to know how far you can push it. Also, do not ETTR by increasing the ISO... there's no point to doing that (ISO is not light/exposure/data).

Answer 2

Is this normal?

Yes, this is to be expected.

Why does it happen?

The JPEG preview image produced by the camera (when saving raw data) is one among a near countless possible interpretations of the raw data. The in-camera histogram is based upon the JPEG preview.

The image displayed on your screen when LR opens a raw file is one among a near countless possible interpretations of the raw data. The histogram displayed by LR is based on the current processed interpretation of the raw data that is displayed on your monitor.

They're not the same interpretation, therefore they won't have the same histogram.

Which one is more accurate?

Neither is necessarily more or less accurate than the other. They just reflect the results of two different interpretations of the same raw data. Neither is a representation of the linear, unprocessed values in the raw file. Both are representations of two different interpretations among a near countless number of possible interpretations of the raw data.

Answer 3

They're both "correct" for a given value of "correct".

The one in the camera knows the formula of the conversion between the RAW file & the presented image. Lightroom has only guesswork, a reverse-engineered interpretation of what Nikon intended for those settings on that RAW file.

So, both are 'correct' but each is measuring something different.

If you want to see what Nikon thinks it ought to look like, use ViewNX-i or the newer NX Studio.

Answer 4

A histogram displays the frequency at which values fall into an arbitrary set of ranges.

Arbitrary in the sense that two histograms of the same data can divide a range possible values into a different quantity of ranges.

For example, one histogram might divide the possible values of the blue channel into 10 ranges and another might divide the same possible values into 100 ranges.

How many ranges to use is entirely dependent on intent. Lightroom’s histogram cannot prevent over exposure (tethering excepted). It is too late. The camera histogram cannot prevent clipping in post processing (in-camera edits excepted). It is too soon.

Ultimately, the photographer just has to learn from experience what the histogram in each tool means.