I have a program that uses GDAL for reading .sid files. It works well for many .sid files, but recently a user sent a file where the colors look incorrect in my program, and they look incorrect in the same way in QGIS, which makes me think it might be a problem in GDAL. The images are linked below. The first image is how a certain part of the .sid file looks in my program and in QGIS. The second image shows how it looks (correct) in the LizardTech GeoViewer program. Because both my program and QGIS show the same issue, I am wondering if it's a problem in GDAL or its .sid plugin. You can see that many of the colors (the red container and the red earth, and the green of the trees) seem washed out in the QGIS/MARPLOT image. But you also see that some of the colors seem to be inverted -- the shadows around the trees are white, and so are the long horizontal black containers. It's perhaps behaving as if it's confusing signed and unsigned values -- values under a certain threshold (such as the near 0,0,0 black values for the tree shadows and the black containers) are being interpreted as actually the highest values, near FF,FF,FF white. And I guess that could give a washed-out look to the other colors like green and red -- instead of appearing as the low valued dark versions, they are somehow being interpreted as the corresponding light versions. I would think they would invert colors (the trees appearing red or blue), but I suppose there are various kinds of signed/unsigned mappings with colors. Is anyone familiar with this phenomenon? (I can provide the .sid file. It's about 800 MB.) IMAGES: GDAL image (my program and QGIS): https://drive.google.com/file/d/11dYIs2BhmgSNmSetO6XNr-Redx9uKYjA/view?usp=sharing GeoViewer image: https://drive.google.com/file/d/1BZqto8vRr4QcKoCS4kEn3AT87yYs9hmi/view?usp=sharing
_______________________________________________ gdal-dev mailing list gdal-dev@lists.osgeo.org https://lists.osgeo.org/mailman/listinfo/gdal-dev
