Strange WB settings in latest Windows weekly

Software darktable
,
1

I updated to 4.3 +2296 this morning, and just noticed that the white balance module is giving very strange readings:
image
This is set to “as shot”, the image looks fine, but it’s very hard to adjust anything using the temp and tint sliders. Actually, moving the temp slider even a little gives a black image.
I opened the same image in 4.3 +1435 and the image looks the same, i.e. fine but the WB module is giving expected readings, something around 4800K and tint near 1.0

Have I missed anything?
OpenCL is active. Workflow is set to ‘none’.

Edit: Here’s an image, no xmp as it shows with default settings.
Licenced Creative Commons, By-Attribution, Share-Alike.
DSC_5059.NEF (24.1 MB)

2

Just noticed that the camera presets work fine, but the temp and tint sliders don’t move. The individual channel coefficients respond though.

3
1 Like
4

For me all looks normal; in either mode “camera preset” and and “WB as shot” I can move sliders and don’t see strange settings. I am on Windows 11; have a Pentax; use dng file format

1 Like
5

Indeed DNG doesn’t seem to be affected. But all other formats I test (NEF, ARW, CR2, RW2, ORF, RAF…) changed to 25K.

1 Like
6

I can replicate @123sg problem with the latest build using a Pentax raw file and change the setting from as shot, to daylight or tungsten doesn’t shift the slider position even though the color changes.

image

1 Like
7

I just added a note to the issue on Github that it’s not only the camera reference setting affected, but all other values (as used for legacy WB) as well.

8

.pef not .dng?

9

I use the pef extension so that I know which camera was used and that it really is a RAW as programs can create DNGs which are not true RAW.

10

Understand, just wanted to confirm. Thanks!

11

It’s over my head :slightly_smiling_face: but I think it’s been fixed correct inverted validity test · darktable-org/darktable@eefa5c7 · GitHub

Thanks all!