The user must provide an array of floats specifying how to distribute the colors along the gradient. These values should range from 0.0 to 1.0 and act like keyframes along the gradient (they mark where the gradient should be exactly a particular color).

In the event that the user does not set the first keyframe value equal to 0 and/or the last keyframe value equal to 1, keyframes will be created at these positions and the first and last colors will be replicated there. So, if a user specifies the following arrays to construct a gradient:

     {Color.BLUE, Color.RED}, {.3f, .7f}
 

     {Color.BLUE, Color.BLUE, Color.RED, Color.RED}, {0f, .3f, .7f, 1f}
 

The user may also select what action the LinearGradientPaint object takes when it is filling the space outside the start and end points by setting CycleMethod to either REFLECTION or REPEAT. The distances between any two colors in any of the reflected or repeated copies of the gradient are the same as the distance between those same two colors between the start and end points. Note that some minor variations in distances may occur due to sampling at the granularity of a pixel. If no cycle method is specified, NO_CYCLE will be chosen by default, which means the endpoint colors will be used to fill the remaining area.

The colorSpace parameter allows the user to specify in which colorspace the interpolation should be performed, default sRGB or linearized RGB.

The following code demonstrates typical usage of LinearGradientPaint:

     Point2D start = new Point2D.Float(0, 0);
     Point2D end = new Point2D.Float(50, 50);
     float[] dist = {0.0f, 0.2f, 1.0f};
     Color[] colors = {Color.RED, Color.WHITE, Color.BLUE};
     LinearGradientPaint p =
         new LinearGradientPaint(start, end, dist, colors);
 

This code will create a LinearGradientPaint which interpolates between red and white for the first 20% of the gradient and between white and blue for the remaining 80%.

This image demonstrates the example code above for each of the three cycle methods: