Auto-Matched Tone curve in RawTherapee_dev_release_5.8-3126 appimage on Linux Mint (Solved)

The behaviour of RawTherapee_dev_release_5.8-3126 & RawTherapee_dev_release_5.8-3124 is markedly different from its previous releases (5.8-3120). The Auto-Matched tone curves and Histogram are very much dissimilar. the screenshots and the common pp3 file is attached. Am I doing something wrong or is it a bug.

B&W Conversion Basic.pp3 (14.0 KB)

Do you have the possibility to share an affected raw file?
Does this affect all your files, or only from a particular camera?

Hello, perhaps the profile you used, B&W Basic Conversion.pp3, isn’t working like it should in this version of RT (just a guess). When I use this profile on a negative (also with RT v5.8-3126), it shows up way too dark!

I suggest to make a new profile using the Film Negative module.

This is happening with all images of converted BW film negatives. This image is a film negative copied with Canon 1000D/50mm 2.8 enlarging lens setup. I had copied about 4K negatives exposed during 1980 decade with this setup and have archived them, each roll in a separate folder. I have had no problem with earlier versions of Rawtherapee. The profile I have used is very basic for initial conversion. After this I select the images edit them further. Now the problem is that when I open any image in the latest RT 5.8-3126, which I have edited with earlier builds, the edits go haywire. Now I have a serious problem with my earlier edited images.

The module used.
Demosicing RCD-Bilinear
White Balance
Film Negative
Black and White
Auto-Matched Tone Curve

These are the minimum basic edits which I have applied. I am attaching the raw file. You can have a look.

BW-UL-01_01.CR2 (9.3 MB)

I’m pretty sure your problem is the combination of Film negative and the Auto Tone curve. Film negative works its own magic for the exposure while inverting. Could you experiment with that?

There is something odd, even without any effect (profile: neutral)

I have some old raws of the same camera, and the new dev version renders them darker, even with neutral profile. Therefore the more extreme auto tonecurve.

Seems to be limited to the 1000D for me. 600D, Sonys and Nikons are fine.

Is there a change in how RT get black and white points?? or some other metadata?? I know this is showing up recently in DT for a few camera’s …not this specific issue but issues impacted by a value not the same as the one in the camera meta data… its mostly WB and highlights that have shown some impact at times there… just wondering…

Thanks for your suggestion @Thanatomanic. I am aware of that. I feel you have not tried my profile on negative with both the versions. My specific question was, that prior to RawTherapee_dev_release_5.8-3124/3126 every thing was working fine with my setup. RawTherapee_dev_release_5.8-3120 still works pretty fine, with the same initial basic profile. Has some bug creeped in within a week or so. You squash one bug and two generate like mythological Hydra. I have got my full respect for developers for their tireless work on this great software. Also I know that they are under tremendous pressure to get a regular version released at the earliest. I will keep on trying the new dev versions for this bug or whatever. Till then I will keep on using 5.8-3120. I can’t afford to open my edited negatives for refinements or to export them through these new versions as the edits go haywire and then I have to start over again.

Why am I using Rawtherapee as this does not happen in ART, Darktable, Lightroom or any other raw developer software which I have tried?
Well that is a different question altogether and for another time.

Too little time, and too many negatives to edit :thinking:

@nnanda, point taken. You are correct that I did not look into it properly and have not tested your image. I will try to see what is going on somewhat later today!

I have opened an issue on GitHub for the darkened rendering shown by @apostel338. The issue with auto-match tone curve may be a side-effect.

There’s definitely something wrong with the white point. Notice the magenta specular highlights on the goose(?).

@nnanda I am not seeing the same effects as you are on the 5.8-3120 build. I loaded your CR2 file, applied the Neutral profile, then loaded your PP3. This is the result:

Looks pretty okay to me…

Could the problem be that you are applying your profile in “partial” mode?

1 Like

@Thanatomanic I am not complaining about ver 3120. Your screenshot of edit with 3120 and mine (the first image) are quite similar except your one is a bit enlarged. Processing profile with fill mode “on and off” is very clear if you read the description with mouse-over hint. Either of them don’t impact my image with my profile in any way as no data is missing to the best of my knowledge. My concern is with version 3126 which is not behaving properly. I hope I am not ranting and annoying you all.

RawTherapee_dev_release_5.8-3128 has been released. I will get back after testing it.

I find it similar to 5.8-3126. Hope for the best in future releases.

@nnanda This is most likely the culprit of your issue as well, it affects more Canon camera’s. We will either revert the change or come up with a proper solution. Sorry about the trouble.

Thanks @Thanatomanic for initiating the action. Nothing to be sorry about.
Thanks @Lawrence37 for opening the issue.

@Thanatomanic @Lawrence37
Tested the latest version 5.8-3137 of RawTherapee. The issue which I had opened regarding Auto-Matched tone curves remains unresolved. I hope in the release version this will be taken care of.

Yes, but there is a solution on the way:

Bugs don’t solve themselves after a rebuild, but wouldn’t that be nice. Anyway, I merged the pull request mentioned by @Thanatomanic and closed the issue. According to my tests, the auto-matched tone curve issue should be fixed as well.


@Lawrence37, @Thanatomanic, @apostel338, @priort and other developers. Thanks for testing and squashing the bug.

Today I downloaded and tested RawTherapee_dev_release_5.8-3139. Auto-Matched Tone Curve seems to work fine with this release, on my film negatives copied on Canon 1000D.

Thanks once again.