Bug #8445

Resizing artifacts when exporting files on disk

Added by jetxee - about 9 years ago. Updated almost 7 years ago.

Closed: invalid
Start date:
Due date:
% Done:


Estimated time:
Affected Version:
git development version
hardware architecture:


I've noticed some artifacts in the exported images when exporting files on disk and the max size is specified. In my case the effect is particularly noticeable when exporting 1920x1920 (see the file attached, pixelated lines along tree branches). All file formats seem to be affected (PPM, JPG, PNG). Export with size 0x0 doesn't produce such artifacts.

I suppose the wrong interpolation algorithm is applied (or not applied at all).

DT version: git, 0.7.1+880 + a custom patch for issue #8385


#1 Updated by jetxee - about 9 years ago

Seems to be not an issue in darktable 0.7.1+913~ga0939ea.

(Does 909089 affect also export?)

#2 Updated by Igor Kuzmin almost 9 years ago

Here's one more resizing artifact:
1:1 export:
downscaled version: (png looks the same)
Notice vertical and horizontal patterns in the sky. Rotating the image reduces the effect, enabling green eq removes it completely I think. So basicly the pattern exists in the original raw data but gets way amplified by downscaling algorithm. I don't want to enable green eq until bug #185 is fixed, but in any case there can be some other patterns that would be amplified like that and that's not good, downscaling should be fixed somehow.

#3 Updated by Tobias Ellinghaus almost 8 years ago

Does this still happen?

#4 Updated by jetxee - almost 8 years ago

Replying to [comment:4 houz]:

Does this still happen?

It doesn't happen for me.

#5 Updated by Igor Kuzmin almost 8 years ago

Artifact I described in my previous comment still happens in master. Both with low and high quality export options. Here's raw file: (only medium contrast tone curve was enabled for previous jpeg examples)
Switching demosaic methods doesn't help. Local variants of match greens options ("local average" and "full and local average") seems to help with this particular picture.

#6 Updated by Simon Spannagel almost 8 years ago

so parafin,

works for you or you still have general problems with this?

#7 Updated by Igor Kuzmin almost 8 years ago

Short answer - yes.
Long answer - not sure if this is a bug. As I understood from discussion on IRC, every scaling algorithm has its downsides. It seems that the one used in DT amplify such regular patterns in image more than other scaling implementations. I see this undesirable, but if match greens functionality is fixed (see bug #185) it wouldn't be an issue for me anymore. So decide for yourselves if you want to fix this.

#8 Updated by Igor Kuzmin almost 8 years ago

By "yes" I meant the problem still exists :)

#9 Updated by Simon Spannagel over 7 years ago

  • Target version changed from 1.0.3 to Candidate for next patch release

#10 Updated by Tobias Ellinghaus over 7 years ago

  • Affected Version set to git development version
  • % Done changed from 0 to 20
  • Status changed from New to Incomplete

Any news on this one? Does it still happen or can we close?

#11 Updated by Jose Carlos Garcia Sogo almost 7 years ago

  • System set to unknown
  • % Done changed from 20 to 0
  • Status changed from Incomplete to Closed: invalid

Also available in: Atom PDF

Go to top