Histogram reading excersice

Hello,

I was wondering what conclusions would you make looking at the histogram below. I will post my thoughts on the image as well as an info on what processing tools were applied to the image after getting 3-5 replies here.

Original histogram (neutral profile in RawTherapee):

Histogram after some edits applied:

You clipped some highlights of red channel

1 Like

Looks like it may have been slightly under-exposed with some leeway on the highlights.

Possibly a tad contrasty in what’s left?

Looks underexposed in the original. Could’ve used another stop or so in exposure.

In the processed one I’d maybe apply some light tonemapping to push the first peak into the last bright quarter.

Exciting challenge! Let’s see who can be most wrong.

I conclude that the photo shows a scene which could be split into two main areas - the light source (lamp or sky) and the subject. It is not a landscape. It is not a portrait. It is not a winter scene. It’s probably also not a daytime scene. If it is a daytime scene then it shows a red bricked building. But more likely it’s night or a scene lit by man-made light. It’s possibly a city, perhaps includes a bright sign, or something shiny. Or it’s a close-up of something man-made.
The prevalence of red in the highlights could be explained by there being incandescent light in the scene.
The shape of the histogram immediately following the second spike is interesting. Whatever it is, it’s probably man-made, i.e. not a photo of a real animal or plant. Could be a photo of some electronic device, a raspberry pi circuit board for example.
The hues in the photo are quite neutral.
The photo probably has a warmed-up tone to it. You probably pushed the white balance temperature, or you used the RGB curves.
The photograph has a small field of view, i.e. not wide-angle.
The end result is still underexposed and you lost detail in the reds.
The areas which are black have a warm tint to them, further indicating the use of RGB curves or film simulation.
Furthermore you use KDE and have a short moustache.

12 Likes

Well, yeah, of course. I just figured it was too obvious to bother pointing out…

4 Likes

@Andrius Excellent food for thought. I especially liked the fact that you have given the processed histogram too. I will strongly use this additional fact provided by you while analysing your first histogram.

From your second histogram, it is clear that you have not tried to lift shadows at all. Also you seem to have dampened the mid tones (in RGB as well value) a bit and only enhanced the red channel without caring for it being clipped.

To my mind, this means that the photo will have a dark background/foreground and a subject that is lit only partially. Since the red colour is dominant in fair skin and during sunset/sunrise time, my guess will be either a long shot of a landscape during sunset/sunrise or a close up of a partially lit face of a Caucasian person with dark background. Of course, considering the extra fact that this is winter time now, the photo could be of some red foliage too!!

Another thing I remembered is that if you apply Fuji Velvia film simulation, the red channel gets enhanced in a similar way. So, second guess will be, you liked the image as it was and just applied a film simulation to it. :slight_smile:

you guys crack me up :stuck_out_tongue:

1 Like

My conclusion is the photo is of a hot pink or fuscia flower.

Thank you all for the comments here.
Here is the source of the red color. As you can see it is neither a Raspberry Pi board nor a fuscia flower. LOL

You are correct. I am wondering if there is a way to boost the red channel of the whole image excluding the parts that are going to get clipped.

What RawTherapee module would that be? I found Color Toning only.

Great thoughts! You were right for most part.
I don’t have any moustache though :slight_smile:

You could use a parametric R-Curve for example

Tonemapping is on the first tab, under retinex.

I am able to reduce the highlights clipping on the red channel via Lab Adjustments → a curve → Parametric → moving Red Saturated slider to the left.
The T shirt losing its very bright color but the whole image is not getting very greenish in that case.

I could not push the second pike to the left yet though. Still playing