Hi :) Kevin Gale wrote:
> Our application scales RGB digital camera images to specific sizes using > Bilinear scaling. We are investigating adding support for the embedded > profiles using lcms. > If I am making the image smaller I will be able to get better performance by > applying the destination profile after the scale. If I am making the image > larger I will be able to get better performance by applying the destination > profile first. Yes, and yes. > Are there any rules for this (such as always transforming first no matter if > I am making the picture larger or smaller) or does it not really matter? One thing to bear in mind is that the interpolation (or averaging) involved in scaling an image is affected by the gamma of the colourspace in which you're scaling. For work in which no error at all can be tolerated, scaling (and most other manipulations) should be done in linear space, with gamma removed. In practice, however, the effects of this consideration are relatively minor, and since colour-transforms are computationally expensive it generally makes sense to apply them to as few pixels as possible. All the best, -- Alastair M. Robinson ------------------------------------------------------------------------- Check out the new SourceForge.net Marketplace. It's the best place to buy or sell services for just about anything Open Source. http://sourceforge.net/services/buy/index.php _______________________________________________ Lcms-user mailing list Lcms-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lcms-user
