I'm now feeling quite stupid: I assumed that the Argyll XYZLUT display profile with color-shifted matrix part would show some kind of obvious colour distortion when the client software CMM used the matrix part rather than the normal LUT part. Howver, both Firefox 3.5.9 and 3.6.3 (with the CMM active for all image data with or without embedded profiles) using that kind of Argyll display profile displayed JPEGs with normal colours, including ordinary JPEGs, as well as special colour-shift test images with corresponding special embedded profiles, as can be found on http://foto.beitinger.de/browser_farbmanagement/farb-testbilder.html and those with v2 embedded profiles on http://color.org/version4html.xalter As Marti Maria says, those same browsers indisputably show "matrix only, discard the LUT" behaviour with his specially prepared image + profile. Why is the use of the Argyll XYZLUT profiles with colour-shifting matrix parts then not showing the shift in Firefox, if Firefox is only using the matrix parts of the profile? If the Firefox CMM were silently rejecting the entire Argyll display profile and reverting to non-CMM display, then the image at the top of the foto.beitinger.de page would be displayed with RGB shifted to BRG--but it displays normal colours in Firefox using the Argyll display profile with the shifted matrix part. The Windows 7 "Photo Viewer" application uses only the matrix part and shows RGB shifted to GRB, while the Windows 7 file manager ("Explorer") uses the LUT part to colour manage its JPEG previews, so those are shown with correct colours, as do other applications that invoke/use the Windows CMM. (Only Microsoft, with its vast software development resources, is able to achieve such counter-intuitive results between its own photo preview application, previews in its file browser, and other applications using the Windows CMM. Bravo Microsoft!) Does anyone have ideas on what's happening here? What am I doing (or thinking) wrong?