Hi, I have found something very strange. I was using gqview to verify that I had successfully created an ImageMagick resize pipeline that properly took into account the gamma of the images (1). Certain pipelines that should have yielded idential results didn't end up yielding identical results. The differences disappeared when viewing at 1:1 magnification. At first I thought it had something to do with image profiles/color management/linear color space, however I was able to simplify it down to an obscure bug in Geeqie related to handling .tif vs. other formats.
In particular, when opening an image as .tif and then zooming out, the "hyper" scaling algorithm will shift luminance. This shift is not observed when opening the identical image as .jpg or .png. (or possibly vice versa - .tif is handled correctly and .png/.jpg is shifted). Since this is quite subtle, I took screenshots and diffed the screenshots in gimp to verify I wasn't going insane. This can be recreated using the following two images: http://martin.stolle.name/files/geeqie_tif_hyper/gamma_noprof.png http://martin.stolle.name/files/geeqie_tif_hyper/gamma_noprof.tif Note: both images were created from the same jpeg and stripped of all color profile information. They contain the identical pixels (verified by gimp diff) When displaying them in geeque with a 1:2 zoom scale (zoom out), they no longer appear identical: http://martin.stolle.name/files/geeqie_tif_hyper/geeqie_png.png http://martin.stolle.name/files/geeqie_tif_hyper/geeqie_tif.png Note: Color management was turned off (but turning it on doesn't change the fact that there is a difference). Diff and diff+equalize from gimp, verify the difference: http://martin.stolle.name/files/geeqie_tif_hyper/geeqie_diff.png http://martin.stolle.name/files/geeqie_tif_hyper/geeqie_diff_equalize.png I also tried explicitly assigning the sRGB color profile to both images, and the difference is still there, with color management turned off OR on. The effect is also observed when using bilinear interpolation, but it is less pronounced. It would be great if someone familiar with source could look into it... Maybe someone else has noticed this before? I wish I could fix it myself, won't have time in the forseeable future. At very least, others may find this report useful 8-). Although I admit the effect is quite subtle, it can be very distracting when doing color critical work 8-(. Martin References: (1) http://www.4p8.com/eric.brasseur/gamma.html ------------------------------------------------------------------------------ Colocation vs. Managed Hosting A question and answer guide to determining the best fit for your organization - today and in the future. http://p.sf.net/sfu/internap-sfd2d _______________________________________________ Geeqie-devel mailing list Geeqie-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geeqie-devel
