Settings to accurately represent true raw histogram in output image

Software RawTherapee
1

I’m looking to use RawTherapee to output a TIFF/PNG/JPEG that most accurately represents the true exposure / histogram of the raw data. This would be similar in concept to UniWB. Does anyone have any setting recommendations / a profile that I could use to achieve this?

I also would love to be able to illustrate areas that are clipping in the rendered output, which I believe unchecking “Clip out-of-gamut colors” allows me to do, but I just wanted to confirm that’s the best approach. Thanks!

2

Hi @dustinkerstein, and welcome aboard! :smiley:

I’m not sure I understand what you need, but I’m going to guess a bit:

  • if you want to see the raw histogram straight away in RT, click the raw button ( rawhist ). You will get a raw histogram, scaled in a linear, log or log-log scale ( rawloghist )
  • if you want a file that shows data the same as they come out from the sensor (after demosaicing), I’m not sure if you will need to use the same profile as an input, working and output profiles, so the data is not converted at any step of the process. But in any case, you will need a linear output profile: you can create one yourself with the ICC Profile Creator rawicchist ) (this is for advanced users), or select one from Elle Stone’s repository of high quality profiles, with a linear TRC (called whateverg10.icc)
  • in either case, I don’t think that has anything to do with UniWB…
  • for the clipping indications as stated in RawPedia:
    The histogram shows the channels listed above using the gamma-corrected output profile when the gamut button Gamut-hist.png is disabled (default), or using the working profile when the button is enabled. The status of this button also affects the values shown in the Navigator panel, as well as the clipped shadow Warning-shadows.png and Warning-highlights.png highlight indicators. It does not affect the raw histogram.
    It has nothing to do with the Clip OOG colors option (which refers to an entirely different thing)

Hope it helps

1 Like
3

Thanks for the reply. Sorry if I wasn’t entirely clear. I’m building a Raspberry pi-based system that uses images created by RawTherapee as the “viewfinder” (rather than relying on the camera’s own JPEG output). I want these images to represent as close to the true exposure of the raw data so I can see where clipping is occurring (useful for me as I’m looking to achieve perfect white for product photos). It sounds like a linear profile might be what I’m looking for. Are there any other settings that would need to be turned off / modified in RawTherapee, or will the ICC profile, once imported, set things correctly?

As for showing clipping in the output file, from what I’ve seen it does appear that Clip OOG does what I’m looking for. I don’t need to see clipping in the histogram in RawTherapee; I need it visually represented (ie. stamped) into the output file. Does that make sense? Thanks for the help.

4

So, guessing again:

  • I understand that your Rasp-Pi system doesn’t have any means of showing a live preview, so you’re effectively blind shooting
  • to know when you are clipping your shots you need the clipping indicators to create a mask over the image, and that mask to be included in the output file: I don’t think that’s possible in RT. Maybe a developer could chime in and say otherwise
  • you need a correct white in your shots, and that means that your sensor needs to be properly recognized by RT, and that you need a proper input profile, specific to your sensor. I think that RT don’t have any such profile, so the shortest path to get your own camera profile would be following Elle Stone’s articles about creating camera profiles
  • sorry, but are you sure that the Clip OOG colors is what you really need? That setting won’t allow you to clip your images, or more precisely, won’t allow your images to have clipped channels. That means in general photography you won’t have to worry (much) about clipping while processing images, but you will effectively lose the most saturated colors that you are in fact clipping, although you won’t realize about it. Any dev may correct me if I’m wrong here
  • if you need correct colors and wish to check a preview of your lighting setup, you may follow this:
    • use the correct, custom input profile
    • load the Neutral profile in RT to cancel any default setting different from 0
    • use a wide gamut working profile (Rec2020 as a bare minimum). It would be even better if you use your own input profile as working profile (although there are a lot of people saying this isn’t advised)
    • option 1: export with the same input camera profile: this would give you a linear rendering of your shot (in many programs this would give a dark rendering)
    • option 2: export with an sRGB output profile (AdobeRGB if you have a good wide gamut display) to look at your image in your display with an image viewer (if your viewer has good color management capabilities, both this image and the previous one should look the same)
  • if you’re shooting product photos, I dare to say you have control over the shot and the ilumination, so you can set up lights, shot a good white card and analyze the RGB channels in RT or an image editor: if you nail the camera settings with the white card, your shots with that lighting won’t clip

Hope it helps

5

As far as profiling - RT’s official documentation on using dcamprof to generate DCP profiles is at How to create DCP color profiles - RawPedia

There’s a bit of discussion on improving the workflow a bit, among other things Reference TIFFs are not tagged with an ICC profile that indicates linear data · Issue #5575 · Beep6581/RawTherapee · GitHub (I still need to take a look at the issues @Thanatomanic mentioned, I’ve been kind of busy with non-photography stuff lately…)

6

Awesome! Thanks for all the info. I’ll play around with those suggestions. Clip OOG does seem like it’s helpful as it visibly shows areas that are clipped (whites show up pink), but it is very possible it’s misleading me…