Edit: Still a problem?: DT 4.3 master : Vectorscope, CIELUV : Missing point in hue ring?

Hello all,

If if cycle the different vectorscopes, the hue rings have six points, except CIELUV, which shows 5 points and a cusp where the 6:th point ought to be. Is this correct?

I have had this hue ring for quite a while using darktable master source. The picture attached shows a vectorscope CIELUV example with logarithmic scale. The cusp is on the blue side of the hue ring.

Darktable git master release-4.3.0-450-gf5a79a849, Linux Debian 11.6

Screenshot from 2023-02-04 15-04-56

I cannot reproduce so maybe due to actual values in the picture you’re editing. It looks like there is no information in the blue for your image. Worth sharing it!

Hello @Pascal_Obry, there is plenty of blue in my raw files. The rendering is perfect. It just the hue ring for CIELUV that is warped, when I compile darktable source from master for dt 4.3. The other hue rings AzBz and RYB are fine.

I can not share that particular picture (Canon 5D mk IV CR2), from first example in thread, but attached is another example (also Canon 5D mk IV CR2), with more blue colors from the sky.

I get the same hue ring cusp for all files displayed, also for raw files from other camera makers, besides Canon.

Screenshot from 2023-02-04 16-17-33

CR2 example. Free to use for all.
_Y8A0003.CR2 (42.0 MB)

Hello again,

Works for me! Sorry for the noise.

I restarted with an empty configdir: ./darktable --configdir and now the CIELUV vectorscope color ring is fine.

Hm, this is the second time I can fix an issue with a new config file. I wonder what I am doing to end up here?

On my side:
image

Cannot help with that :slight_smile:

Thank you for the test. I get the same plot and the same color ring.

I now understand why a new config directory for darktable application made CIELUV vectorscope hue ring complete.

If I set histogram profile = linear prophoto RGB then I get the vectorscope hue ring problem originally reported for vectorscope CIELUV, i.e., one blue point is missing.

If I set histogram profile = linear rec2020 RGB then I get a complete hue ring for vectorscope CIELUV.

I had historgram profile = working profile and working profile = linear prophoto RGB . When I created a new fresh config directory for darktable then all settings were reset to default and the default setting works.

EDIT: I think this is a bug. histogram profile = linear prophoto RGB should render with a complete 6 point hue ring, not a 5 point hue ring with a cusp.

1 Like

Can you start an Issue in GitHub? The information you have here is great but having the Issue in GitHub helps track the resolution.

Hello @g-man and all.

I have filed an issue here: GIT Master : Vectorscope, CIELUV : Missing point in hue ring · Issue #13532 · darktable-org/darktable · GitHub.

This problem (CIELUV, vectorscope hue ring) exists in release-4.0.1 as well. Is this a really old issue?

EDIT: Yes it is old. Same thing in git checkout release-3.8.1.

1 Like