The identification of synthetic vowels in real vowel frames from different speakers

Each of three synthetic /h–d/ tokens, displaced linearly in the F1–F2 plane in the direction /ε/–/æ/, was presented prefixed by real speaker-specific /h–d/ pairs in a binary forced choice identification task. Results indicated that variation in identification scores for the synthetic tokens could not be accounted for with an “Adaptation Level” model based on simple acoustic contrast. The investigation of the response process as a function of phonetic content of the frame requires controls which take account of discrete vs. continuous processing strategies of the frame and focal stimulus.