Project

General

Profile

Bug #11258

Darktable renders (exports) black image for Fuji RAWs

Added by Adam Z over 3 years ago. Updated over 3 years ago.

Status:
Fixed
Priority:
Low
Assignee:
-
Category:
-
Target version:
Start date:
10/26/2016
Due date:
% Done:

100%

Estimated time:
Affected Version:
git master branch
System:
all
bitness:
64-bit
hardware architecture:
amd64/x86

Description

Hello, some of my images come out of "export" as pitch black images with low size. It's easily reproducible with attached RAF and XMP file. The problem occurs with about 1 in 20 images.

Workaround is to crop the image slightly differently (by few pixels) - most often this fixes the problem and the image is exported correctly. If not, re-cropping again will fix the problem.

Camera used: Fuji XT-10, lens: Samyang 12mm f/2

DSCF0013.RAF.xmp (4.08 KB) DSCF0013.RAF.xmp Adam Z, 10/26/2016 05:50 AM
DSCF0013.jpg (98.1 KB) DSCF0013.jpg Adam Z, 10/26/2016 05:50 AM
DSCF0013.RAF (32.1 MB) DSCF0013.RAF Adam Z, 10/26/2016 05:51 AM

Related issues

Related to darktable - Bug #11272: Module `demosaic' outputs NaNs! [preview]Fixed10/29/2016

Associated revisions

Revision 37b3ba58 (diff)
Added by Dan Torop over 3 years ago

demosaic: don't produce NaN in Markesteijn

The TRANSLATE macro which mirrors the image when looking beyond its edge
had an off-by-one in its calculation: the first column beyond the right
edge of the image duplicated the rightmost column of the image, rather
than the second-to-rightmost (and similarly for the first row beyond the
bottom of the image). This caused interpolation code, which looks in a
3x3 area of mosaiced data, to fail, in the case of some red/blue pixels
at the image edge, to find any pixels from which to interpolate. This
produced a division by zero for these pixels, which resulted in an
output of NaN at the edge of Markesteijn demosaiced data.

This bug should only affect the CPU variant of Markesteijn. The OpenCL
version does not mirror the edges, but instead uses VNG
interpolation. Note also that this bug is not in the original dcraw
code.

This fixes #11272. This fixes #11258.

History

#1 Updated by Tobias Ellinghaus over 3 years ago

Are you using OpenCL? If so, could you try if disabling it helps?

#2 Updated by Adam Z over 3 years ago

Tobias Ellinghaus wrote:

Are you using OpenCL? If so, could you try if disabling it helps?

"activate OpenCL support" is checked in the preferences, but can't be unchecked and hover says "not available on this system". It's a bit inconsistent, but I guess OpenCL is not active then ... I don't have dedicated GPU, just Intel integrated one (Intel i5 6200U).

#3 Updated by Tobias Ellinghaus over 3 years ago

  • System changed from Ubuntu to all
  • % Done changed from 0 to 20
  • Status changed from New to Triaged

Ok, now I can reproduce it.

#4 Updated by Tobias Ellinghaus over 3 years ago

It seems to be triggered by shadhi in gaussian mode.

#5 Updated by Roman Lebedev over 3 years ago

  • Related to Bug #11272: Module `demosaic' outputs NaNs! [preview] added

#6 Updated by Roman Lebedev over 3 years ago

  • Target version set to 2.2.0

#7 Updated by Dan Torop over 3 years ago

  • % Done changed from 20 to 100
  • Status changed from Triaged to Fixed

Also available in: Atom PDF

Go to top