How to make CGImageCreateWithImageInRect reference transformed image? - ios

Question

Xcode 5, iOS 7
I am loading an image into a UIImage, and then copy it into another UIImage with a Mirror transform, i.e.:
self.imageB.image=[UIImage imageWithCGImage:[self.imageA.image CGImage] scale:1.0 orientation:UIImageOrientationUpMirrored];
Next, I'm trying to combine the two images into one when saving (originalImage refers to the loaded image prior to copying/transforming into imageA and imageB):
newSize=CGSizeMake(originalImage.size.width*2,originalImage.size.height);
UIGraphicsBeginImageContext(newSize);
UIImage *leftImage=self.imageA.image;
UIImage *rightImage=self.imageB.image;
[self.imageA.image drawInRect:CGRectMake(0,0,newSize.width/2,newSize.height)];
[self.imageB.image drawInRect:CGRectMake(newSize.width/2,0,newSize.width/2,newSize.height) blendMode:kCGBlendModeNormal alpha:1.0];
UIImage *newImage = UIGraphicsGetImageFromCurrentImageContext();
UIGraphicsEndImageContext();
UIImageWriteToSavedPhotosAlbum(newImage, nil, nil, nil);
This works and gives me a single saved mirrored image.
However, I'm trying to do this using only a sub-region of mageA and imageB,
and the result is that the portion from imageB loses it's Mirrored transformation.
I end up with both sides of the final having the same orientation!(Note: visRatio is the percentage of the image I want to keep)
newSize=CGSizeMake(originalImage.size.width*2,originalImage.size.height);
UIGraphicsBeginImageContext(newSize);
UIImage *leftImage=self.imageA.image;
UIImage *rightImage=self.imageB.image;
CGRect clippedRectA = CGRectMake(0,0,originalImage.size.width*visRatio,originalImage.size.height);
CGImageRef imageRefA = CGImageCreateWithImageInRect([self.imageA.image CGImage], clippedRectA);
UIImage *leftImageA = [UIImage imageWithCGImage:imageRefA];
[leftImageA drawInRect:CGRectMake(0,0,newSize.width/2,newSize.height)];
CGRect clippedRectB = CGRectMake(originalImage.size.width-(originalImage.size.width*visRatio),0,originalImage.size.width*visRatio,originalImage.size.height);
CGImageRef imageRefB = CGImageCreateWithImageInRect([self.imageB.image CGImage], clippedRectB);
UIImage *rightImageB = [UIImage imageWithCGImage:imageRefB];
[rightImageB drawInRect:CGRectMake(newSize.width/2,0,newSize.width/2,newSize.height) blendMode:kCGBlendModeNormal alpha:1.0];
UIImage *newImage = UIGraphicsGetImageFromCurrentImageContext();
UIGraphicsEndImageContext();
UIImageWriteToSavedPhotosAlbum(newImage, nil, nil, nil);
It's as though "CGImageCreateWithImageInRect" copies the image from the original, not the transformed, data.
How can I accomplish this without losing the mirrored transformations of imageB ?

Answer 1

I figured it out - this will use the original, full resolution data opened into originalImage and saved a clipped, transformed version (I left the mirroring out of this post for simplicity):
//capture the transformed version
CGSize tmpSize=CGSizeMake(originalImage.size.width, originalImage.size.height);
UIGraphicsBeginImageContext(tmpSize);
[self.imageA.image drawInRect:CGRectMake(0,0, tmpSize.width, tmpSize.height)];
UIImage *tmpImageA=UIGraphicsGetImageFromCurrentImageContext();
visRatio=0.5;
//clip it
clippedRectA=CGRectMake(0,0,roundf(originalImage.size.width*visRatio),originalImage.size.height);
imageRefA=CGImageCreateWithImageInRect([tmpImageA CGImage],clippedRectA);
leftImageA=[UIImage imageWithCGImage:imageRefA];
UIGraphicsEndImageContext();
savedImage=UIGraphicsGetImageFromCurrentImageContext();
UIImageWriteToSavedPhotosAlbum(savedImage, nil, nil, nil);
I believe the key is to draw the image into the context, then use the context for cropping/saving, instead of the main image (imageA).
If you find a more efficient way, please post it!
Otherwise, I hope this helps someone else....