Hi @StendProg, thanks for this PR - it looks really interesting! I will try to get this reviewed and merged this weekend shortly

I added an example to the cookbook: https://github.com/swharden/ScottPlot/tree/master/cookbook#color-by-density

Grayscaled looks sad.
To highligh benefits of this mode, its better to use more powerfull noise with stdDev 5 times bigger then signal amplitude like in demo.
For cookbook example may look better:

double[] noise = ScottPlot.DataGen.RandomNormal(rand, pointCount, 0, 5);

It's hard to see any trend without color by density mode.

I prototyped some preprocessing algoritms to make this mode works fast like SignalConst do for Min/Max calculations. But it take a long time to implement this.

Perhaps use Viridis color scale for this application.

Thanks for the suggestions @StendProg and @citizen3942, this new example looks better:

// plot the noisy signal using the traditional method
var plt = new ScottPlot.Plot(width, height);
plt.PlotSignal(data, yOffset: -40, color: Color.Red);

// use a color array for displaying data from low to high density
Color[] colors = new Color[]
{
	ColorTranslator.FromHtml("#73D055"),
	ColorTranslator.FromHtml("#1F968B"),
	ColorTranslator.FromHtml("#39568C"),
	ColorTranslator.FromHtml("#440154")
};
plt.PlotSignal(data, colorByDensity: colors);

However, I found a bug (#194): when using the colormap, the yOffset argument is ignored. This prevents me from being able to see both graphs when I use:

plt.PlotSignal(data, color: Color.Red);
plt.PlotSignal(data, yOffset: -40, colorByDensity: colors);