Masks
If you create transparent areas in a layer using a transparent gradient, the shape of the transparent areas is controlled by the selected coloring algorithms. Certain colors in the gradient are transparent, so those colors will become transparent in the layer, too. This means that you cannot use this method to create arbitrarily shaped transparent areas.
If you use transformations instead and set the opacity of the solid color to less than 255, you can create arbitrarily shaped transparent areas in the layer. You do need a transformation that will output the area you need, but you could write it yourself. Still, a limitation of this technique is that pixels are either set to the solid color (transparent), or they are colored according to the gradient and the selected coloring algorithms. You can only create sharp edges, not smooth transitions.
Masks overcome these problems. A mask is an invisible layer that is attached to the layer that needs transparent areas. The mask contains transparent areas, that are created with an ordinary transparent gradient. Since the mask is invisible, these transparent areas are invisible as well.
Instead, the layer owning the mask “borrows” the transparent areas of the mask. The shape of these areas is defined by the selected fractal formula, the selected coloring algorithms, and the gradient of the mask. The shape is independent from the layer owning the mask.
 |  |
 | |
 |
| | ||||
|
Layer that needs transparent areas
|
Mask with transparent areas
|
Layer with the mask applied
|
As you can see, the layer uses the transparent areas of the mask. The colors in the mask layer are ignored.
Layers can have multiple masks to add differently shaped transparent areas. If the layer contains transparent areas itself (for example created with a transparent gradient), these are also taken into account.
Next: Working with masks
See Also
Tutorial: Masking
Layers