Linear scanning, speak every button in a different, quieter voice Split sound -> scanning goes to headphones, speaking goes to speaker