Let's say that I got 120 image files in my project. And I want to animate a sprite with all of them.
I was thinking about this method:
CCSprite *temp = [CCSprite spriteWithFile:#"TheNextSprite.png"];
[sprite setTexture:[temp texture]];
The above code will run like once every 0.03 seconds.
sprite is my animated CCSprite.
temp is simply a temporary CCSprite initialized with the image file for the next animation frame, so I can take its texture later.
Is that way efficient? I heard that I could use CCAnimate with a CCSpriteBatchNode, but this is only for one single sprite. Plus I got 120 (big) frames - they won't fit in a 2048x2048 texture canvas.
Have a look at the CCAnimation, CCAnimate and CCSpriteFrame classes.
Sprite frames are what you are trying to mimic, a 'piece' of a texture to which a sprite is mapped. They are best used when combined with spritesheets, not all of the sprites in the animation need to be in the same sprite sheet.
CCAnimation allows you to create an animation out of a sequence of such frames.
And CCAnimate allows you to run that animation as an action:
[node runAction:[CCAnimate actionWithAnimation:animationInstance restoreOriginal:NO]];
Related
all.
I am now developing an ios game using sprite kit.The game mechanics include: infinitely scroll sprites from top to bottom.
My game scene holds just 3 sprites,top,middle,bottom.
I have dozens of sprite images,so my solution is dynamic create 3 sprites ( so the sprites can fill the whole screen),when the bottom sprite is off screen,then destroy it;when the top sprite
goes down the top frame,create a new sprite node,then exchange the bottom,middle,top sprite.
The pseudo code:
in the interface file:
- GameScene:SKScene
{
...
SKSpriteNode *_topSprite;
SKSpriteNode *_middleSprite;
SKSpriteNode *_bottomSprite;
...
}
in the implementation file:
- (void)update:(CFTimeInterval)currentTime
{
// 1 compute time interval
// 2 update sprite node
_topSprite.position move down 100*timeInterval
_middleSprite.position move down 100*timeInterval
_bottomSprite.position move down 100*timeInterval
// 3 crop the off-screen bottom sprite node
if (_bottomSprite is offScreen)
{
[_bottomSprite removeFromParent]
}
// 4 check the highest sprite position visible
// if the highest position is below the frame top,then create new sprite
_bottomSprite = _middleSprite;
_middleSprite = _topSprite;
_topSprite = [self createNewNode]; // random a sprite image
}
the game fps is 60,and the nodes counts is not increasing.
All seems good,but i found sometimes the game just suddenly has a very shot choke causing the moving frame not continuous.
In my instinct,I thought the main reason is caused by step 4.
But I don't have any idea to solve it.
Any suggestions will be appreciated.
Here is the createNewNode method:
- (SKSpriteNode *)createNewNode {
NSString *blockName = [self randomImageName];
SKSpriteNode *sprite = [[SKSpriteNode alloc] initWithImageNamed:blockName];
sprite.position = _highestPosition;
sprite.anchorPoint = CGPointMake(0.5, 0);
NSLog(#"putAnotherNodeOnTop %#", blockName);
[self addChild:sprite];
return sprite;
}
A couple of suggestions.
Since you are using the same three images, you don't need to re-create the sprites each time. It should be less expensive to set the hidden value to YES, move the sprite somewhere, set the texture and then show it again with hidden to NO. SpriteKit doesn't draw hidden nodes, or ones off screen.
Your code basically asks for a PNG on disk every frame in update when createNewNode: is called. It should be more performant to load the textures in memory and assign them to sprites as needed. If you have just a few, making them SKTexture instance variables would be an easy way to do it.
Use spritesheets (texture atlases), it is much faster to draw an image from memory at mapped coordinates than to load it from disk each time. For example with three images in a spritesheet this can be a single draw call. From disk this is three separate draw calls.
You are returning a sprite and adding it to the scene in the createNewNode: method. Maybe that is what you want for some reason. It stood out as possible duplicate effort.
can you create a new CCSprite from a part of a CCSprite within a CCSpriteBatchNode?
For a long time, I've used SpriteFrameCache and BatchNode without a 100% understanding of the two, in particular how they relate to the textureCache. I could use some clear advice to accomplish the following:
Currently, I load a texture atlas into CCSpriteBatchNode and the frame list into the CCSpriteFrameCache and generate a sprite in what I think is basic standard fasion
CCSpriteBatchNode *batchNode = [CCSpriteBatchNode batchNodeWithFile:#"textureAtlasImage.png"];
[[CCSpriteFrameCache sharedSpriteFrameCache] addSpriteFramesWithFile:#"textureAtlasFrames.plist"];
CCSprite * gameObject = [CCSprite spriteWithSpriteFrameName:#"gameObject.png"];
[self addChild:batchNode];
[batchNodeaddChild:gameObject];
For the sake of simplicity with the question, What I'd like to do is divide gameObject into 4 pieces programatically (rather than divide the original image into four pieces and adding each into the textureAtlasImage.png individually).
From reading, I'm thinking something like:
CCSpriteFrame * gameObjectFrame = [[CCSpriteFrameCache sharedSpriteFrameCache] spriteFrameByName:#"textureAtlasFrames.plist"];
CCTexture2D * gameObjectIndividualTexture = [[gameObjectFrame] texture];
CCSpriteFrame * gameObjectPartFrame = [CCSpriteFrame frameWithTexture:gameObjectIndividualTexture offset: rectInPixels: ] ;
[[CCSpriteFrameCache sharedSpriteFrameCache] addSpriteFrame:gameObjectPartFrame name:#"gameObjectPart1"];
But my questions then are:
Is this already in the batchNode? If not, how do I actually create the sprite out of the gameObjectPart1 using batchNodes?
Is it wasteful to add another spriteframe to the cache that duplicates data elsewhere?
You should be able to adjust the texture rect after creating the sprite. Create four sprites using the same sprite frame, then set each sprite's texture rect to one of the 4 smaller regions you want them to use.
Use the sprite.textureRect property to get the CGRect with the original size and change the size respectively origin. For example the lower left rectangle can be created by setting textureRect to the same rect but with size.width and size.height halved.
I have two ImageViews, one called imageView and the other called subView (which is a subview of imageView).
I want to blend the images on these views together, with the user being able to switch the alpha of the blend with a pan. My code works, but right now, the code is slow as we are redrawing the image each time the pan gesture is moved. Is there a faster/more efficient way of doing this?
BONUS Q: I want to allow for my subView image to drawn zoomed in. Currently I've set my subView to be UIViewContentModeCenter, however I can't seem to draw a zoomed in part of my image with this content mode. Is there any way around this?
My drawrect:
- (void)drawRect:(CGRect)rect
{
float xCenter = self.center.x - self.currentImage1.size.width/2.0;
float yCenter = self.center.y - self.currentImage1.size.height/2.0;
subView.alpha = self.blendAmount; // Customize the opacity of the top image.
UIGraphicsBeginImageContext(self.currentImage1.size);
CGContextRef c = UIGraphicsGetCurrentContext();
CGContextSetBlendMode(c, kCGBlendModeColorBurn);
[imageView.layer renderInContext:c];
self.blendedImage = UIGraphicsGetImageFromCurrentImageContext();
UIGraphicsEndImageContext();
[self.blendedImage drawAtPoint:CGPointMake(xCenter,yCenter)];
}
You need to use GPU for image processing which is far faster than using CPU (as you're doing right now).
You can use Core Image framework which is very fast and easy to use but requires iOS 5, or you can use Open GL directly but you need to be experienced and have some knowledge about Open GL Shading.
FUTURE VIEWERS:
I have managed to finish this rotation animation and code with description can be found on tho question. NSImage rotation in NSView is not working
Before you proceed please up vote Duncan C 's answer. As I manage to achieve this rotation from his answer.
I have an image like this,
I want to keep rotating this sync icon, On a different thread. Now I tried using Quartz composer and add the animation to QCView but it is has very crazy effect and very slow too.
Question :
How do I rotate this image continuously with very less processing expense?
Effort
I read CoreAnimation, Quartz2D documentation but I failed to find the way to make it work. The only thing I know so far is, I have to use
CALayer
CAImageRef
AffineTransform
NSAnimationContext
Now, I am not expecting code, but an understanding with pseudo code will be great!
Getting an object to rotate more than 180 degrees is actually a little bit tricky. The problem is that you specify a transformation matrix for the ending rotation, and the system decides to rotate in the other direction.
What I've done is to create a CABasicAnimation of less than 180 degrees, set up to be additive , and with a repeat count. Each step in the animation animates the object more.
The following code is taken from an iOS application, but the technique is identical in Mac OS.
CABasicAnimation* rotate = [CABasicAnimation animationWithKeyPath: #"transform.rotation.z"];
rotate.removedOnCompletion = FALSE;
rotate.fillMode = kCAFillModeForwards;
//Do a series of 5 quarter turns for a total of a 1.25 turns
//(2PI is a full turn, so pi/2 is a quarter turn)
[rotate setToValue: [NSNumber numberWithFloat: -M_PI / 2]];
rotate.repeatCount = 11;
rotate.duration = duration/2;
rotate.beginTime = start;
rotate.cumulative = TRUE;
rotate.timingFunction = [CAMediaTimingFunction functionWithName:kCAMediaTimingFunctionLinear];
CAAnimation objects operate on layers, so for Mac OS, you'll need to set the "wants layer" property in interface builder, and then add the animation to your view's layer.
To make your view rotate forever, you'd set repeat count to some very large number like 1e100.
Once you've created your animation, you'd add it to your view's layer with code something like this:
[myView.layer addAnimation: rotate forKey: #"rotateAnimation"];
That's about all there is to it.
Update:
I've recently learned of another way to handle rotations of greater than 180 degrees, or continuous rotations.
There is a special object called a CAValueFunction that lets you apply a change to your layer's transform using an arbitrary value, including values that specify multiple full rotations.
You create a CABasicAnimation of your layer's transform property, but then instead of providing a transform, the value you supply is an NSNumber that gives the new rotation angle. If you provide a new angle like 20pi, your layer will rotate 10 full rotations (2pi/rotation). The code looks like this:
//Create a CABasicAnimation object to manage our rotation.
CABasicAnimation *rotation = [CABasicAnimation animationWithKeyPath:#"transform"];
rotation.duration = 10.0;
CGFLOAT angle = 20*M_PI;
//Set the ending value of the rotation to the new angle.
rotation.toValue = #(angle);
//Have the rotation use linear timing.
rotation.timingFunction =
[CAMediaTimingFunction functionWithName:kCAMediaTimingFunctionLinear];
/*
This is the magic bit. We add a CAValueFunction that tells the CAAnimation we are
modifying the transform's rotation around the Z axis.
Without this, we would supply a transform as the fromValue and toValue, and
for rotations > a half-turn, we could not control the rotation direction.
By using a value function, we can specify arbitrary rotation amounts and
directions and even rotations greater than 360 degrees.
*/
rotation.valueFunction =
[CAValueFunction functionWithName: kCAValueFunctionRotateZ];
/*
Set the layer's transform to it's final state before submitting the animation, so
it is in it's final state once the animation completes.
*/
imageViewToAnimate.layer.transform =
CATransform3DRotate(imageViewToAnimate.layer.transform, angle, 0, 0, 1.0);
[imageViewToAnimate.layer addAnimation:rotation forKey:#"transform.rotation.z"];
(I Extracted the code above from a working example application, and took out some things that weren't directly related to the subject. You can see this code in use in the project KeyframeViewAnimations (link) on github. The code that does the rotation is in a method called `handleRotate'
I have two circles which move around the screen. The circles are both UIViews which contain other UIViews. The area outside each circle is transparent.
I have written a function to create a CGPath which connects the two circles with a quadrilateral shape. I fill this path in a transparent CALayer which spans the entire screen. Since the layer is behind the two circular UIViews, it appears to connect them.
Finally, the two UIViews are animated using Core Animation. The position and size of both circles change during this animation.
So far the only method that I have had any success with is to interrupt the animation at regular intervals using an NSTimer, then recompute and draw the beam based on the location of the circle's presentationLayer. However, the quadrilateral lags behind the circles when the animation speeds up.
Is there a better way to accomplish this using Core Animation? Or should I avoid Core Animation and implement my own animation using an NSTimer?
I faced a similar problem. I used layers instead of views for the animation. You could try something like this.
Draw each element as a CALayer and include them as sublayers for your container UIVIew's layer. UIViews are easier to animate, but you will have less control. Notice that for any view you can get it's layer with [view layer];
Create a custom sublayer for your quadrilateral. This layer should have a property or several of properties you want to animate for this layer. Let's call this property "customprop". Because it is a custom layer, you want to redraw on each frame of the animation. For the properties you plan to animate, your custom layer class should return YES needsDisplayForKey:. That way you ensure -(void)drawInContext:(CGContextRef)theContext gets called on every frame.
Put all animations (both circles and the quad) in the same transaction;
For the circles you can probably use CALayers and set the content, if it is an image, the standard way:
layer.contents = [UIImage imageNamed:#"circle_image.png"].CGImage;
Now, for the quad layer, subclass CALayer and implement this way:
- (void)drawInContext:(CGContextRef)theContext{
//Custom draw code here
}
+ (BOOL)needsDisplayForKey:(NSString *)key{
if ([key isEqualToString:#"customprop"])
return YES;
return [super needsDisplayForKey:key];
}
The transaction would look like:
[CATransaction begin];
CABasicAnimation *theAnimation=[CABasicAnimation animationWithKeyPath:#"customprop"];
theAnimation.toValue = [NSValue valueWithCGPoint:CGPointMake(1000, 1000)];
theAnimation.duration=1.0;
theAnimation.repeatCount=4;
theAnimation.autoreverses=YES;
theAnimation.timingFunction = [CAMediaTimingFunction functionWithName:kCAMediaTimingFunctionEaseIn];
theAnimation.delegate = self;
[lay addAnimation:theAnimation forKey:#"selecting"];
[CATransaction setValue:[NSNumber numberWithFloat:10.0f]
forKey:kCATransactionAnimationDuration];
circ1.position=CGPointMake(1000, 1000);
circ2.position=CGPointMake(1000, 1000);
[CATransaction commit];
Now all the draw routines will happen at the same time. Make sure your drawInContext: implementation is fast. Otherwise the animation will lag.
After adding each sublayer to the UIViews's layer, rememeber to call [layer setNeedsDisplay]. It does not get called automatically.
I know this is a bit complicated. However, the resulting animations are better than using a NSTimer and redrawing on each call.
If you need to find the current visible state of the layers, you can call -presentationLayer on the CALayer in question, and this will give you a layer that approximates the one used for rendering. Note I said approximates - it's not guaranteed to be fully accurate. However it may be good enough for your purposes.