The purpose of this paper is to investigate the relationship between the manual segmental evaluation and the automatic segmental evaluation of a set of English utterances from the same reference sentence. A model utterance from a native speaker of English was chosen based on which all the other utterances were evaluated. Two groups were chosen for evaluation. Fifty eight native speakers of English and one hundred fifty Korean female university students participated in the evaluation. The prosody of the model utterance was cloned to all the other utterances of the two groups. Since the prosodic aspect of the utterances was controlled, only the segmental aspect was then evaluated both by a human evaluator and by a Praat script. The script compared the sequential spectra between the model utterance and the others. The scores from the two evaluations were correlated with each other and this supports the idea that a manual segmental evaluation can be replaced by computer by performing automatic spectral comparisons.