Here is the iPad Simulator with four nested UIViews, drawing a custom background, and an inner UILabel. I am rotating the top UIView's CALayer, by getting it's layer and setting transform with a rotateY CATransform3D, animated on a separate thread (but the changes to transform are being sent on the main thread, naturally):
Note- this animation does not loop correctly, hence it appears to bounce.
The layers do animate as a whole, but curiously, the first child and it's descendants appear to be floating above the UIView with the transform applied!
The UIViews themselves are children in another UIView, which has a red background. There are no other transformations being applied anywhere else.
The positions for each UIView were set using setFrame initially at the start.
What is causing this strange behaviour, and how can I ensure the child UIViews transform with their parent, giving a flat appearance to the surface as a whole?
Well. Perhaps unsurprisingly I was doing something silly, but since I'd not used CALayer transforms before, I didn't know if things were acting up. I had overridden layoutSubviews on the UIViews I was creating, and the act of rotating the CALayer was triggering this call and then pushing the child components frame around, due to a bug.
The problem is that CALayers don't actually do 3D perspective by default. In order to do that you need to make a minor change to the layer's transform (which is of type CATransform3D)
You want to change the .m34 field of the transform to a small negative value. Try -1/200 to -1/500 as a starting range. If I remember correctly it should be the negative of 1 over the image height/width.
Change the .m34 property of the layer that you want to appear to "come off the page" and rotate in 3D. When you do that the Z setting of the layer does matter, and will both make closer layers bigger and also make layers that are further away disappear behind other things.
I suggest you do a Google search on "CATransform3D m34" for more information. There is a fair amount of information on the net about it.
Related
I'm animating an ImageView using CABasicAnimation.
I move its layer to left, right, up and down and sometimes I'd scale it bigger then reset it to its original size etc.
I'm doing all this to its layer so I thought I might have to move & scale the real thing along with its layer as well but when I tested it with tap gesture to see if it really was just staying where it started, it wasn't. Therefore I no longer need to change its view's frames as far as I'm concerned.
Is changing a view's layer's values also change its view's values?
A UIView is no more than a fancy wrapper for a CALayer – bringing UIResponder events & animation conveniences among many other things.
Many UIView properties are actually just forwarded versions of the underlying CALayer properties, defined purely for convenience.
A view's frame & bounds properties should always reflect the layer equivalents.
transform is slightly more complex, as for the view it's of type CGAffineTransform – whereas on the layer it's CATransform3D. If the layer's transform can be represented as a CGAffineTransform, then you'll be able to access it from the view after setting it on the layer. If it can't be represented, then its value is undefined.
Therefore yes, you are right in saying you don't need to update the frame or transform on the UIView when changing it on its CALayer. Although note that these properties won't reflect the 'in-flight' values of the animation – you'll need to access the layer's presentationLayer for that.
Also note that as #par & #jrturton mention, if a layer's transform is not the identity transform, then the frame is undefined and you therefore shouldn't use it.
I'm working on an iPhone app which will be displaying large scrollable and zoomable surface containing a grid of pictures and text labels over them. I need to be able to change the position of pictures individually and control the opacity level of labels. I tried to accomplish these goals by using UIScrollView and Core Animation.
The subview of UIScrollView contains two main sublayers: one for displaying pictures and one for the labels. Pictures are CALayers with their contents property set to CGImage and they are added as sublayers to the pictures layer. Labels are CATextLayers and they are sublayers of the second layer. Transparency of the labels layer changes depending on the zoom scale of the scroll view.
And here is the problem: everything works fine when the labels are fully opaque or fully transparent, but when they are semitransparent scrolling starts to be jerky and FPS drops to about 35. Obviously blending of these layers slows everything down, but I couldn't find a way to fix it. I will appreciate any ideas on how to improve the performance in this situation. Maybe there is a better way to draw text labels than using CATextLayer?
Is it possible for you to merge the two "main layers" of your UIScrollView into one? Also, is it possible for you to add layers directly to UIScrollView's layer instead of adding additional ones?
I find that I get huge performance wins by reducing the number of layers that exist for the sole purpose of containing other layers.
One solution is to add a shadow to the most background layers of both image layer and text layer.
There are number of shadow properties that you can tweak i.e. shadowPath, shadowColor, shadowOffset and shadowRadius - set each of them, don't miss any. Also set yourlayer.masksToBounds = NO.
Do not forget to add yourlayer.shouldRasterize = YES because this will have better performance impact.
I want to build an animated transition between two view controllers in iOS, resembling the "Box" transition in PowerPoint or the "Reflection" transition in Keynote.
You can see it here, at 2:10:
http://youtu.be/1fLQg5hFQQg?t=2m10s
What's the best way to do this?
Thanks!
That would be a complex animation to recreate. You'd need to use a CAAnimationGroup that grouped several different animations running at once. You'd want to animate a rotation around the y axis with the center of rotation lifted off the screen, on both the view controller that is animating away and the view that your are animating into place.
You would have to tweak the transform to make it draw with perspective (you add a small value to the .m34 record in the transform). That's because CA animations are orthographic by default (they don't show perspective.)
The reflections could be created using a special subclass of CALayer that lets you create duplicates of a layer. I'm blanking on the name of that layer subclass at the moment. You'd set up 1 duplicate with a scale of -1 on the y axis to flip it upside down, and a darkening effect. I've never done it myself, but I've seen several examples in books and online.
I have multiple views with many UILabels on the views. (all constructed in Interface Builder).
I am then trying to create a "smaller" replica of my view when you pinch the screen.
To do this I apply:
view.transform = CGAffineTransformMakeScale(.5, .5);
and then I also adjust the frame of view.
The problem is that after the transformation, the text in all of my UILabels becomes "blurry". It doesn't stay pixel perfect as it is in full-scale view.
Is there a way to increase the pixelation of the labels after the transformation?
Applying a transform to a UIView or CALayer merely scales the rasterized bitmap of that layer or view. This can lead to blurriness of the resulting UI element, because they aren't re-rendered at that new scale.
If you really want your text or images to be crisp at the new scale factor, you're going to need to manually resize them and cause them to redraw instead of applying a transform. I described one way that I did this with a UIView hosted in a UIScrollView in this answer.
You might be able to create a single method that traverses your view hierarchy for your one main view, recursively reads each subview's frame, scales that down, and then forces a redraw of its contents. Transforms are still great to use for interactive manipulation or animation, but you can then trigger a full manual scaling and redraw at the end of the manipulation or animation.
I'm attempting to optimize my app. It's quite visually rich, so has quite a lot of layered UIViews with large images and blending etc.
I've been experimenting with the shouldRasterize property on CALayers. In one case in particular, I have a UIView that consists of lots of sub views including a table. As part of a transition where the entire screen scrolls, this UIView also scales and rotates (using transforms).
The content of the UIView remains static, so I thought it would make sense to set view.layer.shouldRasterize = YES. However, I didn't see an increase in performance. Could it be that it's re-rasterizing every frame at the new scale and rotation? I was hoping that it would rasterize at the beginning when it has an identity transform matrix, and then cache that as it scales and rotates during the transition?
If not, is there a way I could force it to happen? Short of adding a redundant extra super-view/layer that does nothing but scale and rotate its rasterized contents...
You can answer your own question by profiling your application using the CoreAnimation instrument. Note that this one is only available in a device.
You can enable "Color hits in Green and Misses Red". If your layer remains red then it means that it is indeed rasterizing it every frame.