Anyone ever heard of this technique? Apparently its to do with hue manipulation, heres an extract from my proggy I use:
'The Hue, Saturation, and Lightness (HSL) color model defines a color based on its hue (color), saturation (purity of the color), and lightness (brightness).
Hue
Hue is what we usually define as the color of an object. The hue of the sky is blue, the hue of a banana is yellow, and the hue of an apple is red.
The color wheel is a useful tool in understanding hue and the numeric values assigned to hues.
The lightness axis adds a third dimension to a color wheel, which produces the HSL color model.
The hue of a color is assigned a number from 0° to 360°. Red is defined at the 0° point, blue is 120°, cyan is 180°, green is 240°, and so on. Change the hue (H) values on Picture Publisher’s Color Picker to see the hue change. Also, look at the Hue Shift slider in Picture Publisher’s Hue Map dialog box for another way to change the hue of an image. Open the Hue Map dialog box by clicking Hue Map on the Map menu.
Saturation
The saturation of a color describes the purity of the color. The range of saturation is defined in value percentages from 100% (full color intensity) to 0% (no color intensity). A black-and-white photograph has a color saturation value of 0%. Look at the Saturation Shift slider in the Hue Map dialog box to see how changes in saturation affect an image.
Lightness
Lightness refers to the amount of white or black in a color and is defined in percentages from 100% (totally white) to 0% (totally black). Fifty percent lightness is the pure hue. Look at the Lightness Shift slider in the Hue Map dialog box to see how changes in lightness affect an image.'
I've had a bit of a play around, but I'm not sure if theres any advantage in doing it. (more playing is necessary to see if it helps with images)
Attached is a photo (mars) split into the HSL channels: no manipulation, just split to show you.
The trouble with processing the hue channel as a greyscale image is that hue is cyclic (goes round and round the colour wheel without bounds) while the greyness scale is bounded by black and white.
But splitting the image into HSV channels is useful for working on the value ("lightness", HSV=HSL) channel alone (and to a lesser extent saturation) without having to deal with distracting colour artefacts.
