I have one this scenario:
I add to a layer CAAnimation that transforms it to a specific frame. with a starting time of 0.
Then I add another CAAnimation that transforms it to a different frame. with a starting time of 0.5.
what happens is that the layer immediately gets the second frame (with no animation) and after the first animation time passes the second animation is completed correctly.
This is the animation creation code:
+ (CAAnimation *)transformAnimation:(CALayer *)layer
fromFrame:(CGRect)fromFrame
toFrame:(CGRect)toFrame
fromAngle:(CGFloat)fromAngle
toAngle:(CGFloat)toAngle
anchor:(CGPoint)anchor
vertical:(BOOL)vertical
begin:(CFTimeInterval)begin
duration:(CFTimeInterval)duration {
CATransform3D fromTransform = makeTransform(layer, fromFrame, fromAngle, anchor, vertical);
CATransform3D midTransform1 = makeTransformLerp(layer, fromFrame, toFrame, fromAngle, toAngle, anchor, 0.33, vertical);
CATransform3D midTransform2 = makeTransformLerp(layer, fromFrame, toFrame, fromAngle, toAngle, anchor, 0.66, vertical);
CATransform3D toTransform = makeTransform(layer, toFrame, toAngle, anchor, vertical);
CAKeyframeAnimation *animation = [CAKeyframeAnimation animationWithKeyPath:#"transform"];
animation.values = #[[NSValue valueWithCATransform3D:fromTransform],
[NSValue valueWithCATransform3D:midTransform1],
[NSValue valueWithCATransform3D:midTransform2],
[NSValue valueWithCATransform3D:toTransform]
];
animation.beginTime = begin;
animation.duration = duration;
animation.fillMode = kCAFillModeBoth;
animation.calculationMode = kCAAnimationPaced;
animation.removedOnCompletion = NO;
return animation;
}
EDIT
in most scenarios, this code works well and the animations are sequenced correctly. But if I set 1 transform animation to start after 2 seconds and then set another transform to start after 4 seconds. the first transform is applied immediately to the layer and the second animation starts from there.
Any Idea how can I separate the animation to run one after the other?
(I prefer not using a completion block)
Thanks
The easiest and most glaring early fix would be to change the fill mode so that the second animation is not clamped on both ends overriding the previous animation.
animation.fillMode = kCAFillModeForwards;
Also I would adjust the begin time to be
animation.beginTime = CACurrentMediaTime() + begin;
If it is a matter of overlapping begin times and durations and not this let me know and I can provide that as well.
I have the strangest problem with animating two identical UIImageViews. A separate event triggers these two UIImageViews being moved closer to another, which I animate in the following manner:
CGFloat originalX = self.firstMatchImage.layer.position.x;
self.firstMatchImage.layer.position = CGPointMake(originalX + displacement, self.firstMatchImage.layer.position.y);
CABasicAnimation *animation = [CABasicAnimation animationWithKeyPath:#"position.x"];
animation.fromValue = #(originalX);
animation.duration = .18;
[self.firstMatchImage.layer addAnimation:animation forKey:#"position.x"];
CGFloat originalXsecond = self.secondMatchImage.layer.position.x;
self.secondMatchImage.layer.position = CGPointMake(originalXsecond - displacement, self.secondMatchImage.layer.position.y);
CABasicAnimation *animationSecond = [CABasicAnimation animationWithKeyPath:nil];
animationSecond.fromValue = #(originalXsecond);
animationSecond.duration = .18;
[self.secondMatchImage.layer addAnimation:animationSecond forKey:nil];
And this works just fine. The images move toward one another quite readily. However, I also have an event trigger an animation to make these UIImageviews move back to their original location. To do this, on the viewDidLoad method of my storyboard controller, I capture their base locations thusly:
firstImageOrigin = self.firstMatchImage.center;
secondImageOrigin = self.secondMatchImage.center;
So in the method call to move these UIImageViews back I do the following:
CGFloat originalX = self.firstMatchImage.layer.position.x;
self.firstMatchImage.layer.position = firstImageOrigin;
CABasicAnimation *animation = [CABasicAnimation animationWithKeyPath:nil];
animation.fromValue = #(originalX);
animation.duration = .18;
[self.firstMatchImage.layer addAnimation:animation forKey:nil];
CGFloat originalXsecond = self.secondMatchImage.layer.position.x;
self.secondMatchImage.layer.position = secondImageOrigin;
CABasicAnimation *animationSecond = [CABasicAnimation animationWithKeyPath:nil];
animationSecond.fromValue = #(originalXsecond);
animationSecond.duration = .19;
[self.secondMatchImage.layer addAnimation:animationSecond forKey:nil];
Much to my displeasure, only the UIImageView called firstImageView moves back correctly. It works perfectly for the first image, but the second UIImageView(self.secondMatchImage), doesn't behave the same way, it animates and moves a little, but not back to its original location. here is what the result looks like:
I've played with naming the keyPaths something other than nil, but when I trigger a tap gesture, it crashed the app if I have named the keyPaths. I have tested this on an iPhone and the simulator - same behavior. I am super stumped here.
Any insight on how this is happening would be super great! Thanks!
So, turns out, grabbing the location of the images in the controller's viewDidLoad method was not the ideal solution. Instead, I set the original coordinates to
CGPointMake(0,0)in the viewDidLoad method and use this login in my animation method:
if (firstImageOrigin.x == 0 && firstImageOrigin.y == 0) {
firstImageOrigin = self.firstMatchImage.center;
secondImageOrigin = self.secondMatchImage.center;
}
And now it works perfectly!
I was just wondering if this is the correct way to animate a CALayer with a CABasicAnimation.
On Stack Overflow I have been taught how to animate UI objects by setting a new position before running a CABasicAnimation:
Animating a UI Object Example
gameTypeControl.center = CGPointMake(gameTypeControl.center.x, -slidingUpValue/2);
CABasicAnimation *removeGameTypeControl = [CABasicAnimation animationWithKeyPath:#"transform.translation.y"];
[removeGameTypeControl setFromValue:[NSNumber numberWithFloat:slidingUpValue]];
[removeGameTypeControl setToValue:[NSNumber numberWithFloat:0]];
[removeGameTypeControl setDuration:1.0];
[removeGameTypeControl setTimingFunction:[CAMediaTimingFunction functionWithControlPoints:0.8 :-0.8 :1.0 :1.0]];
[[gameTypeControl layer] addAnimation:removeGameTypeControl forKey:#"removeGameTypeControl"];
Now I've been tried this method on a CALayer but it seems to work differently. For me to get the same result. I have the set the ToValue to the new y position instead of using the value 0 like I've done with my UI object animations.
Animating a CALayer Example
serveBlock2.position = CGPointMake((screenBounds.size.height/4)*3, -screenBounds.size.width/2);
CABasicAnimation *updateCurrentServe2 = [CABasicAnimation animationWithKeyPath:#"position.y"];
updateCurrentServe2.fromValue = [NSNumber numberWithFloat:slidingUpValue/2];
[updateCurrentServe2 setToValue:[NSNumber numberWithFloat:-screenBounds.size.width/2]];
[updateCurrentServe2 setDuration:1.0];
[serveBlock2 addAnimation:updateCurrentServe2 forKey:#"serveBlock2 updateCurrentServe2"];
Is this correct? Am I doing this right?
The problem is that if serveBlock2 is not a view's immediately underlying layer, setting its position, as you do in the first line of the second example, starts a different animation (an implicit animation). The way to prevent that is by turning off implicit animations. Thus this example from my book:
CompassLayer* c = (CompassLayer*)self.compass.layer;
[CATransaction setDisableActions:YES]; // <=== this is important
c.arrow.transform = CATransform3DRotate(c.arrow.transform, M_PI/4.0, 0, 0, 1);
CABasicAnimation* anim = [CABasicAnimation animationWithKeyPath:#"transform"];
anim.duration = 0.8;
[c.arrow addAnimation:anim forKey:nil];
That way, I don't have to have a fromValue or a toValue! The old value and the new value are known automatically from the presentation layer and the model layer.
I'm trying to chain the animation of a layer and its sublayer. However, the problem I'm having is that the model update of the sublayer's position for its animation is visually apparent during its superlayer's animation. Here's the code I'm using to chain animations:
// Superlayer.
CFTimeInterval now = [self.artworkContainer.layer convertTime:CACurrentMediaTime() fromLayer:nil];
CABasicAnimation *slideDown = [CABasicAnimation animationWithKeyPath:#"position"];
slideDown.duration = SLIDE_DOWN_DURATION; //SLIDE_DOWN_DURATION;
slideDown.beginTime = now;
slideDown.timingFunction = [CAMediaTimingFunction functionWithName:kCAMediaTimingFunctionLinear];
slideDown.fromValue = [self.artworkContainer.layer valueForKey:#"position"];
CGPoint finalPointOfContainer = CGPointMake(self.artworkContainer.layer.position.x, self.artworkContainer.layer.position.y+verticalDistanceOfFirstFall);
slideDown.toValue = [NSValue valueWithCGPoint:finalPointOfContainer];
self.artworkContainer.layer.position = finalPointOfContainer;
[self.artworkContainer.layer addAnimation:slideDown forKey:#"position"];
// Sublayer
CABasicAnimation *moveActualArtwork = [CABasicAnimation animationWithKeyPath:#"position"];
moveActualArtwork.duration = timeOfSecondFall;
moveActualArtwork.beginTime = now + SLIDE_DOWN_DURATION;
moveActualArtwork.timingFunction = [CAMediaTimingFunction functionWithName:kCAMediaTimingFunctionLinear];
moveActualArtwork.fromValue = [self.artWork.layer valueForKey:#"position"];
CGPoint finalPointOfArtwork = CGPointMake(self.artWork.layer.position.x, self.artWork.layer.position.y+verticalDistanceOfSecondFall);
moveActualArtwork.toValue = [NSValue valueWithCGPoint:finalPointOfArtwork];
self.artWork.layer.position = finalPointOfArtwork; // If this is commented out, the superlayer animation looks correct (of course this isn't a true fix because the sublayer snaps back to its original position without a model update).
moveActualArtwork.delegate = self;
[self.artWork.layer addAnimation:moveActualArtwork forKey:#"position"];
I found the answer as I was writing the question. I needed to set the fillMode property of the second animation: moveActualArtwork.fillMode = kCAFillModeBackwards;
If one were to think of the fromValue and toValue as positions on a timeline, the problem was that before the sublayer hit the fromValue spot on the timeline, it was working off of its model layer to determine its position. By using kCAFillModeBackwards, the fromValue is effectively extended to cover the timeline previous to it so iOS knows what to render before the animation actually begins.
I found the answer reviewing the WWDC 2011 video Session 421 - Core Animation Essentials, at 45:31.
I am rotating a CALayer and trying to stop it at its final position after animation is completed.
But after animation completes it resets to its initial position.
(xcode docs explicitly say that the animation will not update the value of the property.)
any suggestions how to achieve this.
Here's the answer, it's a combination of my answer and Krishnan's.
cabasicanimation.fillMode = kCAFillModeForwards;
cabasicanimation.removedOnCompletion = NO;
The default value is kCAFillModeRemoved. (Which is the reset behavior you're seeing.)
The problem with removedOnCompletion is the UI element does not allow user interaction.
I technique is to set the FROM value in the animation and the TO value on the object.
The animation will auto fill the TO value before it starts, and when it's removed will leave the object at it's correct state.
// fade in
CABasicAnimation *alphaAnimation = [CABasicAnimation animationWithKeyPath: #"opacity"];
alphaAnimation.fillMode = kCAFillModeForwards;
alphaAnimation.fromValue = NUM_FLOAT(0);
self.view.layer.opacity = 1;
[self.view.layer addAnimation: alphaAnimation forKey: #"fade"];
Core animation maintains two layer hierarchies: the model layer and the presentation layer. When the animation is in progress, the model layer is actually intact and keeps it initial value. By default, the animation is removed once the it's completed. Then the presentation layer falls back to the value of the model layer.
Simply setting removedOnCompletion to NO means the animation won't be removed and wastes memory. In addition, the model layer and the presentation layer won't be synchronous any more, which may lead to potential bugs.
So it would be a better solution to update the property directly on the model layer to the final value.
self.view.layer.opacity = 1;
CABasicAnimation *animation = [CABasicAnimation animationWithKeyPath:#"opacity"];
animation.fromValue = 0;
animation.toValue = 1;
[self.view.layer addAnimation:animation forKey:nil];
If there's any implicit animation caused by the first line of above code, try to turn if off:
[CATransaction begin];
[CATransaction setDisableActions:YES];
self.view.layer.opacity = 1;
[CATransaction commit];
CABasicAnimation *animation = [CABasicAnimation animationWithKeyPath:#"opacity"];
animation.fromValue = 0;
animation.toValue = 1;
[self.view.layer addAnimation:animation forKey:nil];
Reference:
Animations Explained by objc.io.
"iOS 7 Programming Pushing the Limits" by Rob Napier and Mugunth Kumar.
Set the following property:
animationObject.removedOnCompletion = NO;
You can simply set the key of CABasicAnimation to position when you add it to the layer. By doing this, it will override implicit animation done on the position for the current pass in the run loop.
CGFloat yOffset = 30;
CGPoint endPosition = CGPointMake(someLayer.position.x,someLayer.position.y + yOffset);
someLayer.position = endPosition; // Implicit animation for position
CABasicAnimation * animation =[CABasicAnimation animationWithKeyPath:#"position.y"];
animation.fromValue = #(someLayer.position.y);
animation.toValue = #(someLayer.position.y + yOffset);
[someLayer addAnimation:animation forKey:#"position"]; // The explicit animation 'animation' override implicit animation
You can have more information on 2011 Apple WWDC Video Session 421 - Core Animation Essentials (middle of the video)
just put it inside your code
CAAnimationGroup *theGroup = [CAAnimationGroup animation];
theGroup.fillMode = kCAFillModeForwards;
theGroup.removedOnCompletion = NO;
A CALayer has a model layer and a presentation layer. During an animation, the presentation layer updates independently of the model. When the animation is complete, the presentation layer is updated with the value from the model. If you want to avoid a jarring jump after the animation ends, the key is to keep the two layers in sync.
If you know the end value, you can just set the model directly.
self.view.layer.opacity = 1;
But if you have an animation where you don't know the end position (e.g. a slow fade that the user can pause and then reverse), then you can query the presentation layer directly to find the current value, and then update the model.
NSNumber *opacity = [self.layer.presentationLayer valueForKeyPath:#"opacity"];
[self.layer setValue:opacity forKeyPath:#"opacity"];
Pulling the value from the presentation layer is also particularly useful for scaling or rotation keypaths. (e.g. transform.scale, transform.rotation)
So my problem was that I was trying to rotate an object on pan gesture and so I had multiple identical animations on each move. I had both fillMode = kCAFillModeForwards and isRemovedOnCompletion = false but it didn't help. In my case, I had to make sure that the animation key is different each time I add a new animation:
let angle = // here is my computed angle
let rotate = CABasicAnimation(keyPath: "transform.rotation.z")
rotate.toValue = angle
rotate.duration = 0.1
rotate.isRemovedOnCompletion = false
rotate.fillMode = CAMediaTimingFillMode.forwards
head.layer.add(rotate, forKey: "rotate\(angle)")
This works:
let animation = CABasicAnimation(keyPath: "opacity")
animation.fromValue = 0
animation.toValue = 1
animation.duration = 0.3
someLayer.opacity = 1 // important, this is the state you want visible after the animation finishes.
someLayer.addAnimation(animation, forKey: "myAnimation")
Core animation shows a 'presentation layer' atop your normal layer during the animation. So set the opacity (or whatever) to what you want to be seen when the animation finishes and the presentation layer goes away. Do this on the line before you add the animation to avoid a flicker when it completes.
If you want to have a delay, do the following:
let animation = CABasicAnimation(keyPath: "opacity")
animation.fromValue = 0
animation.toValue = 1
animation.duration = 0.3
animation.beginTime = someLayer.convertTime(CACurrentMediaTime(), fromLayer: nil) + 1
animation.fillMode = kCAFillModeBackwards // So the opacity is 0 while the animation waits to start.
someLayer.opacity = 1 // <- important, this is the state you want visible after the animation finishes.
someLayer.addAnimation(animation, forKey: "myAnimation")
Finally, if you use 'removedOnCompletion = false' it'll leak CAAnimations until the layer is eventually disposed - avoid.
Without using the removedOnCompletion
You can try this technique:
self.animateOnX(item: shapeLayer)
func animateOnX(item:CAShapeLayer)
{
let endPostion = CGPoint(x: 200, y: 0)
let pathAnimation = CABasicAnimation(keyPath: "position")
//
pathAnimation.duration = 20
pathAnimation.fromValue = CGPoint(x: 0, y: 0)//comment this line and notice the difference
pathAnimation.toValue = endPostion
pathAnimation.fillMode = kCAFillModeBoth
item.position = endPostion//prevent the CABasicAnimation from resetting item's position when the animation finishes
item.add(pathAnimation, forKey: nil)
}
Simply setting fillMode and removedOnCompletion didn't work for me. I solved the problem by setting all of the properties below to the CABasicAnimation object:
CABasicAnimation* ba = [CABasicAnimation animationWithKeyPath:#"transform"];
ba.duration = 0.38f;
ba.fillMode = kCAFillModeForwards;
ba.removedOnCompletion = NO;
ba.autoreverses = NO;
ba.repeatCount = 0;
ba.toValue = [NSValue valueWithCATransform3D:CATransform3DMakeScale(0.85f, 0.85f, 1.0f)];
[myView.layer addAnimation:ba forKey:nil];
This code transforms myView to 85% of its size (3rd dimension unaltered).
#Leslie Godwin's answer is not really good, "self.view.layer.opacity = 1;" is done immediately (it takes about one second), please fix alphaAnimation.duration to 10.0, if you have doubts.
You have to remove this line.
So, when you fix fillMode to kCAFillModeForwards and removedOnCompletion to NO, you let the animation remains in the layer. If you fix the animation delegate and try something like:
- (void)animationDidStop:(CAAnimation *)anim finished:(BOOL)flag
{
[theLayer removeAllAnimations];
}
...the layer restores immediately at the moment you execute this line. It's what we wanted to avoid.
You must fix the layer property before remove the animation from it. Try this:
- (void)animationDidStop:(CAAnimation *)anim finished:(BOOL)flag
{
if([anim isKindOfClass:[CABasicAnimation class] ]) // check, because of the cast
{
CALayer *theLayer = 0;
if(anim==[_b1 animationForKey:#"opacity"])
theLayer = _b1; // I have two layers
else
if(anim==[_b2 animationForKey:#"opacity"])
theLayer = _b2;
if(theLayer)
{
CGFloat toValue = [((CABasicAnimation*)anim).toValue floatValue];
[theLayer setOpacity:toValue];
[theLayer removeAllAnimations];
}
}
}
The easiest solution is to use implicit animations. This will handle all of that trouble for you:
self.layer?.backgroundColor = NSColor.red.cgColor;
If you want to customize e.g. the duration, you can use NSAnimationContext:
NSAnimationContext.beginGrouping();
NSAnimationContext.current.duration = 0.5;
self.layer?.backgroundColor = NSColor.red.cgColor;
NSAnimationContext.endGrouping();
Note: This is only tested on macOS.
I initially did not see any animation when doing this. The problem is that the layer of a view-backed layer does not implicit animate. To solve this, make sure you add a layer yourself (before setting the view to layer-backed).
An example how to do this would be:
override func awakeFromNib() {
self.layer = CALayer();
//self.wantsLayer = true;
}
Using self.wantsLayer did not make any difference in my testing, but it could have some side effects that I do not know of.
It seems that removedOnCompletion flag set to false and fillMode set to kCAFillModeForwards doesn't work for me either.
After I apply new animation on a layer, an animating object resets to its initial state and then animates from that state.
What has to be done additionally is to set the model layer's desired property according to its presentation layer's property before setting new animation like so:
someLayer.path = ((CAShapeLayer *)[someLayer presentationLayer]).path;
[someLayer addAnimation:someAnimation forKey:#"someAnimation"];
Here is a sample from playground:
import PlaygroundSupport
import UIKit
let resultRotation = CGFloat.pi / 2
let view = UIView(frame: CGRect(x: 0.0, y: 0.0, width: 200.0, height: 300.0))
view.backgroundColor = .red
//--------------------------------------------------------------------------------
let rotate = CABasicAnimation(keyPath: "transform.rotation.z") // 1
rotate.fromValue = CGFloat.pi / 3 // 2
rotate.toValue = resultRotation // 3
rotate.duration = 5.0 // 4
rotate.beginTime = CACurrentMediaTime() + 1.0 // 5
// rotate.isRemovedOnCompletion = false // 6
rotate.fillMode = .backwards // 7
view.layer.add(rotate, forKey: nil) // 8
view.layer.setAffineTransform(CGAffineTransform(rotationAngle: resultRotation)) // 9
//--------------------------------------------------------------------------------
PlaygroundPage.current.liveView = view
Create an animation model
Set start position of the animation (could be skipped, it depends on your current layer layout)
Set end position of the animation
Set animation duration
Delay animation for a second
Do not set false to isRemovedOnCompletion - let Core Animation clean after the animation is finished
Here is the first part of the trick - you say to Core Animation to place your animation to the start position (you set in step #2) before the animation has even been started - extend it backwards in time
Copy prepared animation object, add it to the layer and start the animation after the delay (you set in step #5)
The second part is to set correct end position of the layer - after the animation is deleted your layer will be shown at the correct place