When mounting, when should I use a loop device?

According to the following question:
What is a “loop device” when mounting?
A loop device is a file that acts as a block-based device. While I can understand that conceptually, how is this different from mounting any other file? For example, if I take a dd image of an ext2 filesystem and then mount it, I don’t have to use the loop option- mount -t ext2 file.dd /mnt/mount-point succeeds.