This is a new audio edge memorization method I came up with. The aim of the method is to produce a sequence of audio pairs that is as short and easy to remember as possible on average, without being too hard to learn. I also wanted some properties, such as edge orientation, to be immediately...