Human and Automatic Speaker Recognition over Telecommunication Channels