Re: RGB to CIELab transformation
Kai-Uwe Behrmann <[email protected]>
| Newsgroups | gmane.comp.lib.cairo |
|---|---|
| Message-ID | <[email protected]> |
Am 20.01.12, 11:55 +0100 schrieb Andrea Canciani: > On Fri, Jan 20, 2012 at 8:01 AM, Necko Nedjoni <[email protected]> wrote: >> As a part of a larger project, I was recently assigned to investigate the >> possibility of colorings >> within our visualizations. >> Namely, I came across the CIELab color space, which seems to be preferable >> choice in the related >> literature. The initial readings suggest that a conversion from >> device-depended RGB is possible, and >> I wonder how this could be achieved with cairo. > > Converting a color in a device-dependent colorspace to a color in a > device-independent colorspace is not actually a meaningful operation. > You might assume that the RGB components are in a specific > device-independent colorspace (given current cairo behavior I would > suggest a linear rgb colorspace with the same primaries as sRGB or > sRGB itself). > You could then use normal color conversion routines, like those > provided by LittleCMS. Agreed, it makes sense to normalise from device colours to a blending colour space, then do compositing in Cairo and convert into something appropriate for output. You suggested sRGB for blending and that is fine like with most single gamma profiles like ROMM/ProPhotoRGB and LStarRGB to name a few. Necko, you stated to want use Lab, so you might choose a bigger gamut than sRGB. Cairo has no support for colour spaces with asymmetric gamma like CIE*Lab during blending. One can expect in Cairo hue changes during blending CIE*Lab content. kind regards Kai-Uwe Behrmann -- developing for colour management www.behrmann.name + www.oyranos.org -- cairo mailing list [email protected] http://lists.cairographics.org/mailman/listinfo/cairo