The gamma ramp is a remapping/lookup table. The values of the R,G, and B components of a colour are used as array indices to index into the gamma ramp table; the values read from the gamma ramp table are what gets displayed by the video hardware.


Think of it as working like this (except done by the hardware):

struct INPUT_COLOUR{    BYTE red, green, blue;};struct OUTPUT_COLOUR{    WORD red, green, blue;};OUTPUT_COLOUR ApplyGammaRamp( INPUT_COLOUR input, D3DGAMMARAMP gammaRamp ){    OUTPUT_COLOUR output;    output.red   = gammaRamp.red[ input.red ];    output.green = gammaRamp.green[ input.green ];    output.blue  = gammaRamp.blue[ input.blue ];    return output;}

The input to the gamma ramp is an 8-bit value (i.e. in the range 0 to 255), and the output of the gamma ramp is a 16bit value (i.e. in the range of 0 to 65535); the reason for this is one of the reasons for gamma correction in the first place - human vision isn't linear, the output of a monitor/TV isn't linear, and 8 bits isn't enough to represent the range well enough (you need at least 12, which is what many RAMDACs use).

So to tint the output a certain colour, just tint all the entries in the gamma ramp by that colour, scaled by the position in the table; and to fade, just scale all the values in the table.

If you want to know more about gamma in general, Charles Poynton's site is an excellent resource: http://www.poynton.com/GammaFAQ.html