Concatenative Speech Synthesis Using SRELP