<div dir="ltr">Michael,<br><br>I cannot reproduce the problem here.<br><br>I have tested to import your RAW file using digiKam for KDE4 (0.10.0-beta4) or digiKam for KDE3 (0.9.5-svn) using new RAWImport tool for image editor, and all work fine with highlight recovery settings.<br>
<br>For RAW file decoding, i use libkdcraw based on LibRaw (<a href="http://www.libraw.org">http://www.libraw.org</a>). LibRaw currently use the last dcraw code for demosaicing. Now look my screenshots:<br><br>* This the preview of raw import tool with RAW demosaicing settings. Postprocessing use default settings (no gamma/curve/exposure adjustment) . Of course like it's a preview with a divided by 4 image size, huge zooming give artifact. But it's enough to set demosaicing configuration. Also, working on a reduced raw decoded image speedup computation. Note that i use 16 bits color depth, not 8 (it's a non sence to use 8 bits color depth with RAW)<br>
<br><a href="http://digikam3rdparty.free.fr/Screenshots/temp/rawimport/pef-rawimport-preview.png">http://digikam3rdparty.free.fr/Screenshots/temp/rawimport/pef-rawimport-preview.png</a><br><br>And now, the best: the full raw image in editor. Look like the color (white balance) is better than RAW studio !<br>
<br><a href="http://digikam3rdparty.free.fr/Screenshots/temp/rawimport/pef-editor.png">http://digikam3rdparty.free.fr/Screenshots/temp/rawimport/pef-editor.png</a><br><br>For me, it's a good example how to use properly RAW files in digiKam.<br>
<br>Note : the preview from digiKam icon view is rendered using JPEG preview image from RAW file. It's not RAW image data decoded. In fact the preview JPEG is adjusted using camera firmware. It's still JPEG : only 8 bits color depth ! this is why highlight are broken, and it's a good example to explain why RAW is always better than JPEG.<br>
<br>Unforget : jpeg is fine to take shot and publish as well the image on the web (fats workflow). Never use JPEG if you want to retouch photo, especially color settings ! This file format is not adapted and use lossy compression algorithm.<br>
<br>Best<br><br>Gilles Caulier<br></div>