|BLOT_MULTISHADE||Image data is treated as 3 independant grey scale channels.
Each channel is applied as a scalar to a color component. 3 CDATA type values
are required when using this. The first is RED, second GREEN, third BLUE.
These colors are scaled according to the appropriate (RGB) channel, and
summed. This may seem like a silly thing to do, and I have yet to explain
it to anyone such that it makes sense.
I usually use only 2 channels, one application used all 3.
Suppose you have an image, a blue circle, and a green shade on it resembling
a specular. The circle is entirely the same, uniform level of blue.
The green specular is a varying shade of green. Together, the image looks
like a blue circule with a semi-cyan specular - as each channel is indepenantly
added together. Now, apply to the blue channel (which is for argument a constant 192)
some other color. Scale this color between black and full intensity according
to the blue channel. Pick this color as brown. Now the blue circle is brown,
constant. Next, apply the green channel to some other color, say white. Now
adding a white specular upon the brown circle causes a shiny stone like thing
to be present. The circle color could easily be adjusted to red, blue,
pink, brown, grey, and maintaining a white specular, which quickly uses
one image to be many different colors (peices in a game?).
This may seem like a lot of work to do, but in actuality, doesn't cause
as much overhead as one might expect, it's a rather simple operation to do,
nearly approaching a simple shift and add.