iOS How to perform rotation on a cgimage efficiently (up to 125 frames)? - iphone

Question

I have two images, let's call them image A and image B. I'm trying to rotate and scale image A and to draw it on top of image B at specific location.
I use the function and method below to :
1. rotate and scale my image
2. Draw Image A into Image B at a specific location.
This code works fine but is rather slow, is there a way to improve it?
Function to rotate and scale :
CGImageRef rotateAndScaleImageCreate(const CGImageRef cgImage, const CGFloat radians,const CGFloat scalefactor){
CGImageRef rotatedImageRef = NULL;
const CGFloat originalWidth = CGImageGetWidth(cgImage)*scalefactor;
const CGFloat originalHeight = CGImageGetHeight(cgImage)*scalefactor;
const CGRect imgRect = (CGRect){.origin.x = 0.0f, .origin.y = 0.0f, .size.width = originalWidth, .size.height = originalHeight};
const CGRect rotatedRect = CGRectApplyAffineTransform(imgRect, CGAffineTransformMakeRotation(radians));
/// Create an ARGB bitmap context
CGContextRef bmContext = NYXImageCreateARGBBitmapContext(rotatedRect.size.width, rotatedRect.size.height, 0);
if (!bmContext)
return nil;
/// Rotation happen here
CGContextTranslateCTM(bmContext, +(rotatedRect.size.width * 0.5f), +(rotatedRect.size.height * 0.5f));
CGContextRotateCTM(bmContext, radians);
/// Draw the image in the bitmap context
CGContextDrawImage(bmContext, (CGRect){.origin.x = -originalWidth * 0.5f, .origin.y = -originalHeight * 0.5f, .size.width = originalWidth, .size.height = originalHeight}, cgImage);
/// Create an image object from the context
rotatedImageRef = CGBitmapContextCreateImage(bmContext);
/// Cleanup
//CGImageRelease(rotatedImageRef);
CGContextRelease(bmContext);
return rotatedImageRef;
}
Method to draw image at specific location within another image:
-(UIImage*) PositionImage:(CGImageRef)image backgroundImage:(CGImageRef)backgroundImage atPointX:(float)X pointY:(float)Y withScale:(float)scale andRotation:(float)rotation{
//////////////////////////////////////////
CGLayerRef layer;
CGImageRef resultImage;
CGContextRef context, layerContext;
void *bitmapData;
CGColorSpaceRef colorSpace;
CGSize canvasSize;
int bitmapByteCount;
int bitmapBytesPerRow;
//Get the background image
//UIImage * backgroundImg = [UIImage imageNamed:#"Mask-Inverted.png"];
//Initialize the canvas size!
canvasSize = CGSizeMake(480,640);//[backgroundImg size];
//
bitmapBytesPerRow = (canvasSize.width * 4);
bitmapByteCount = (bitmapBytesPerRow * canvasSize.height);
//Create the color space
colorSpace = CGColorSpaceCreateDeviceRGB();
bitmapData = malloc( bitmapByteCount );
//Check the the buffer is alloc'd
if( bitmapData == NULL ){
NSLog(#"Buffer could not be alloc'd");
}
//Create the context
context = CGBitmapContextCreate(bitmapData, canvasSize.width, canvasSize.height, 8, bitmapBytesPerRow, colorSpace, kCGImageAlphaPremultipliedLast);
if( context == NULL ){
NSLog(#"Context could not be created");
}
///Create the Layer
layer = CGLayerCreateWithContext(context, canvasSize, NULL);
if( layer == NULL ){
NSLog(#"Layer could not be created");
}
///Create the layer Context
layerContext = CGLayerGetContext(layer);
if( layerContext == NULL){
NSLog(#"No Layer context");
}
//Draw the image background into the bitmap context
CGContextDrawImage(context, CGRectMake(0.0f,0.0f,480,640),backgroundImage);
//Rotate and Scale Image
CGImageRef rotatedImage =rotateAndScaleImageCreate(image,rotation,scale);
//Draw a image on the layer
CGContextDrawImage(layerContext,CGRectMake(0,0,CGImageGetWidth(rotatedImage),CGImageGetHeight(rotatedImage)),rotatedImage);
CGImageRelease(rotatedImage);
//Draw the layer in the context at the coordinate
CGContextDrawLayerAtPoint( context, CGPointMake(X-CGImageGetWidth(rotatedImage)*0.5f, Y-CGImageGetHeight(rotatedImage)*0.5f), layer);
CGLayerRelease(layer);
//Get the result image
resultImage = CGBitmapContextCreateImage(context);
CGContextRelease(context);
//Cleanup
free(bitmapData);
CGColorSpaceRelease(colorSpace);
//Create UIImage
UIImage *imageReturned = [UIImage imageWithCGImage:resultImage];
CGImageRelease(resultImage);
return imageReturned;
}

Answer 1

If you're creating lots of images at the same time, reuse the same CGContext. Also, don't create a new rotated image; rotate the image within the existing context.
Basically, you should only have to create a CGContext once, no matter how many images you end up making with it. That should speed things up significantly.