Processing takes the log (natural algorithm), then the first derivative and then removes the average. It removes gain, but more importantly , transforms the equation to be linear with concentration (assuming the Beer-Lambert model applies).