The study focuses on the effects of using word-final units in word-internal positions in synthesized speech, which are often perceived negatively by listeners due to disruptions in local timing.