Not really. Imagine taking a rectangular region of interest from an image. With offset arrays, you can use the original indices if that suits you. I’d say that’s strictly better than using always zero based offsets.