Hi,
MilkDrop's "per-pixel" and Flickernoise's "per-vertex" equations are
exactly the same.
First versions of MD improperly called them "per-pixel", but they also
changed the name to "per-vertex" later. Both MD and FN use the same
linear interpolation between vertices.
To get an idea of what these equations do, you can imagine that the
current picture is mapped to a surface made of rubber, on which you have
a grid of handles (the "vertices") that you can move to distort the
picture. You can move a "handle" using dx (in the horizontal direction)
or dy (in the vertical direction) in a per-vertex equation. The input
variables x and y (between 0 and 1) give the position of the current
"handle".
For example, this gives a (repeated) sinewave distortion of the picture:
per_vertex=dy=0.1*sin(12*x)
You can write to other variables (zoom, rot, ...) in per-vertex
equations to produce local effects. Also, besides x and y, there are two
other input variables that depend on the current vertex:
* rad, which gives the distance from the center of the screen
* ang, which gives the angle between the current vertex/"handle" and a
horizontal line at the center of the screen. ang is not supported by FN yet.
S.
_______________________________________________
http://lists.milkymist.org/listinfo.cgi/devel-milkymist.org
IRC: #milkymist@Freenode
