The Backing Image - iOS Core Animation - page 15

If you want to simply zoom the layer contents image, you can do so using the layer's
transform or affineTransform properties (see Chapter 5, “Transforms,” for an
explanation of transforms), but that's not the purpose of contentsScale .
The contentsScale property is actually part of the mechanism by which support for
high-resolution (a.k.a. Hi-DPI or Retina ) screens is implemented. It is used to determine the
size of the backing image that the layer should automatically create when drawing, and the
scale at which the contents image should be displayed (assuming that it isn't already
being scaled by the contentsGravity setting). UIView has an equivalent-but-little-
used property called contentScaleFactor .
If contentsScale is set to 1.0, drawing will be done at a resolution of 1 pixel per point.
If it is set to 2.0, drawing will be done at 2 pixels per point, a.k.a. Retina resolution. (In case
you are unclear on the distinction between pixels and points, this is explained later in the
chapter.)
This doesn't actually make any difference when using kCAGravityResizeAspect
because it scales the image to fit the layer, regardless of its resolution. But if we switch our
contentsGravity to kCAGravityCenter instead (which doesn't scale the image),
the difference will be more apparent (see Figure 2.3).
Figure 2.3 A Retina image displayed with the wrong contentsScale by default
As you can see, our snowman is huge and pixelated. That's because CGImage (unlike
UIImage ) has no internal concept of scale. When we used the UIImage class to load our
snowman image, it correctly loaded the high-quality Retina version. But when we used the
CGImage representation to set that image as our layer contents , the scale factor was
lost in translation. We can fix that by manually setting the contentsScale to match our
UIImage scale property (see Listing 2.2). Figure 2.4 shows the result.