I want to apply series of affine transformations to an image using Core Image, but I have one confusion. Can series of transformations cause loss in quality of image that is preventable, and does Core Image rendering engine actually optimises it to avoid loss in quality? Example, I start with a 1920×1080 image which is rotated by 90 degrees and scaled down to fit in a 1920×1080 rectangle. Next if I rotate the image back by 90 degrees, I scale it up by applying an affine transform to fit the original 1920×1080 rectangle. If scaling down and scale up are done serially, this will cause loss in quality of image. But if Core Image concatenates the matrices, this could be avoided. Can we say that Core Image rendering engine always takes care of such scenarios?