Han, Heekyung2019-12-112019-12-112019-08https://hdl.handle.net/11299/209072University of Minnesota Ph.D. dissertation. August 2019. Major: Speech-Language-Hearing Sciences. Advisor: Robert Schlauch. 1 computer file (PDF); ix, 105 pages.Clear speech, which is characterized by specific acoustic changes that are distinguished from ordinary conversational speech, is a speaking strategy that enhances a talker’s intelligibility in adverse listening conditions. An increase in the range of a talker’s fundamental frequency (F0) is known as one of the several acoustic changes that is observed when clear speech is produced. Although an increase in F0 range is often seen in clear speech, its contribution to the clear speech benefit is unknown. Experiment 1 in this dissertation examined whether an increase in F0 variation contributes significantly to the clear speech benefit in native speakers of American English. Experiment 2 evaluated clear speech effects in native speakers of Seoul Korean who started to learn English after the age of six and also examined the role of F0 variation. The clear speech benefit was measured by having talkers produce sentences in a conversational and a clear speaking style. The stimuli for these experiments were produced by several talkers and were recorded digitally. At the time of recording, participants were instructed to read aloud low-context sentences in conversational (Experiments 1 & 2), clear (Experiments 1 & 2), and exaggerated-F0 (only Experiment 2) speaking styles. The exaggerated speaking style, which is similar to infant-directed speech with a wide range of F0, was a condition given to the native Korean talkers because Koreans do not typically vary their F0 much in various speaking styles. To characterize acoustic-phonetic changes at the sentence level in talkers’ productions, five acoustic changes were measured: speech rates, long-term spectra, F0 distribution, vowel formant frequencies, and vocal intensity levels. Sentences from the talkers were presented to native listeners of American English in a perceptual study. F0-manipulated speech was synthesized from the clear speech (Experiment 1) and from the exaggerated-F0 speech (Experiment 2) to examine whether F0 variation is a contributing factor in the intelligibility benefits in native English speakers and in native Korean speakers. This was accomplished by compressing the F0 contours of clear speech to match those of conversational speech in Experiment 1 and by compressing the F0 contours of exaggerated-F0 speech to match those of conversational speech in Experiment 2. Listeners were randomly presented with sentences in different speaking styles and asked to type in the sentence after orally repeating each sentence that they heard. The percentage of correct keywords was calculated for each speaking style. The data revealed that F0 range did not contribute to the clear speech benefit. The exaggerated-F0 speech condition for the Korean talkers showed slightly poorer intelligibility benefit than the clear speech condition. A follow-up study of speech naturalness revealed that clear speech is more natural than exaggerated-F0 speech. However, a significant correlation between intelligibility and speech naturalness was not found. Although the experiments were designed to examine directly the role of F0 range on the clear speech benefit, the recordings and perceptual data provided opportunities to study other perceptual correlates of this phenomenon. The primary acoustic factor contributing to the clear speech benefit for native English and native Korean talkers was an increase in the intensity of high-frequency speech sounds.enclear speech benefitexaggerated-F0 speechfundamental frequency (F0) rangehigh-frequency emphasisnative Korean talkersperceptual correlatesThesis or Dissertation