I'd like to color the line of a 2D-Plot. I want the coloring to be a function of x. Say
f(x)=x^2
I want to produce a plot where the color of f(x) ranges from red to blue,
depending on a function g(x), say
g(x)=x
If g(x)=10, f(x) should be blue, if g(x)=0, f should be colored red. In between I need a smooth transition between the two colors.
Any help, idea or query would be greatly appreciated, thanks!
When plotting points, Gnuplot supports the lc variable option which allows to specify the linetypes (and thus colors) for individual points. However, with the lines plotting style (and a color gradient along the line being plotted), one has to most likely digress to an alternative strategy. For example, one could divide the x-range into independent segments, for each segment calculate the corresponding color (in terms of your function g(x)) and finally plot f(x) on each of the segments independently:
set terminal pngcairo rounded enhanced font ",16"
set output 'test.png'
set size ratio 2/(1+sqrt(5.))
unset key
xMin = 0.0
xMax = 10.0
set xr [xMin:xMax]
set yr [0:100]
#coloring function, gMin/gMax has to be set accordingly
#so that the scaled function h(x) does not overflow the [0,1] interval
gMin = 0.
gMax = +10.
g(x) = x
#function to be plotted
f(x) = x*x
#coloring function rescaled into interval [0,1]
#for each g(x), gMin/gMax should be adjusted correspondingly
h(x) = (g(x) - gMin)/(gMax - gMin)
#generate RGB representation of the interpolated color
#corresponding to the red (gMin) -> blue (gMax) transition
color(x) = sprintf('#%02X00%02X', (1-h(x))*255, h(x)*255);
#divide the x range into N segments
N = 200
dx = (xMax - xMin)/N
set samples 1000
#make the segments overlap a bit so that we don't need too high "samples"
eps = dx/10
#plot each segment with corresponding color
binLeftBorder(i) = xMin + i*dx
binMidPoint(i) = (binLeftBorder(i) + binLeftBorder(i+1))/2
isInBin(i, x) = (x >= (binLeftBorder(i) - eps) && x < (binLeftBorder(i+1) + eps))
plot for [i=0:N-1] isInBin(i, x)?f(x):(1/0) w l \
lw 4 lc rgb color(binMidPoint(i))
This would produce:
