CR2 files from a Canon Powershot S100 are not properly recognized and treated: they are heavily magenta-tinted, and show a strong lens distorsion. Here is an example of both CR2 and JPG files produced by the camera: http://www.marie-noelle-augendre.com/fichiers/Darktable/IMG_0008.CR2 http://www.marie-noelle-augendre.com/fichiers/Darktable/IMG_0008.JPG
Marie Noelle, This is a typical case where JPEG is better than RAW, due to post processing done by camera to render JPEG. Here you can see the comparison of Raw (on left side) and JPEG (on right side): http://www.flickr.com/photos/digikam/6964153119/sizes/o/in/photostream/ RAW is loaded in light table as half size preview. We don't use JPEG embedded. JPEG is full size preview. If we load RAW JPEG embedded, both image are the same. Why ? It's simple : you camera, to process JPEG, it apply Lens distortion corrections. You have used a short focal (as a fish eyes lens), and camera fix image using lens profile. Probably that chromatic aberration correction are processed. This is why the magenta-tinted area disappear in JPEG. RAW file, is... raw image, unprocessed by camera. It's not possible (or it's complex) to reproduce camera fixes with RAW. Alex can you confirm my analyse ?
Created attachment 69370 [details] Your sample processed with dcraw_emu -w Cannot reproduce with LibRaw: dcraw_emu -w works fine on your sample (sure, no barrel correction). See attached result
Marie Noelle This file still valid using last digiKam 3.50 and a most recent version of libkdcraw ? Gilles Caulier
Marie Noelle, What's about this file using last digiKam 4.2.0 AND last Libraw 0.16 ? Gilles Caulier
Marie Noëlle, Problem still valid with last digiKam 4.10.0 ? Gilles Caulier
New digiKam 4.11.0 is available with official PKG installer for OSX. https://www.digikam.org/node/740 Can you reproduce the problem with this release ? Gilles Caulier
digiKam 4.12.0 is out : https://www.digikam.org/node/741 We need a fresh feedback using this release please... Thanks in advance.
With digiKam 5.0.0, this problem is not reproducible. I close this file now. Don't hesitate to re-open if necessary. Gilles Caulier