Hmm, this is really hard to debug without seeing the actual filters. But if I would need to guess, I'd say the implementation of the ROI callback (passed when calling the CIKernel) is wrong. I can also recommend using a MTKView for displaying a CIImage instead of using Core Graphics. There is a sample from Apple showing how to do that.

Filed as FB12901757. Thanks for looking into this!

FB12914857. Thanks for looking into this!

Also filed as FB12966266, including a demo project. Thanks for looking into this!

Filed under FB13121690. Thanks for looking into this!

You can also access the tip through the configuration. I think you can simply add a button to your tip layout that invalidates the tip when tapped: Button(action: { configuration.tip.invalidate(reason: .tipClosed) }, label: { Image(systemName: "xmark") })

Filed as FB13168973. Thanks!

The issues seem to have been fixed on iOS 17.1 beta 2. Thanks a lot!

The issues seem to have been fixed on iOS 17.1 beta 2. Thanks a lot!

The order of the color channels in a CGImage depends on its alphaInfo and bitmapInfo. The default format for 8-bit images is actually BGRA (instead of RGBA) in many places. Where do you get the image from? What's its bitmapInfo and does it differ between iOS 16 and 17?

The -framework CoreImage flag needs to be set in two lines in Xcode, otherwise it will mask the space between the words:

The -framework CoreImage flag needs to be set in two lines in Xcode, otherwise it will mask the space between the words:

There is also an Obj-C constant for it.

From the comment in the header file: A NSNumber that specifies the maximum memory footprint (in megabytes) that the CIContext allocates for render tasks. Larger values could increase memory footprint while smaller values could reduce performance. It basically sets how much memory Core Image is allowed to use (roughly) during rendering. From what I observed, the default seems to be 256 MB.

I was wondering the same thing, and it seems that improvements are coming in later versions.