17.05.2014 Views

FRIDAY MORNING, 20 MAY 2005 REGENCY E, 8:30 A.M. TO 12:00 ...

FRIDAY MORNING, 20 MAY 2005 REGENCY E, 8:30 A.M. TO 12:00 ...

FRIDAY MORNING, 20 MAY 2005 REGENCY E, 8:30 A.M. TO 12:00 ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

2:10–2:15 Break<br />

2:15–4:15<br />

Panel Discussion<br />

<strong>FRIDAY</strong> AFTERNOON, <strong>20</strong> <strong>MAY</strong> <strong>20</strong>05<br />

<strong>REGENCY</strong> D, 1:<strong>00</strong> <strong>TO</strong> 5:<strong>00</strong> P.M.<br />

Session 5pSC<br />

Speech Communication: Speech Production and Perception II „Poster Session…<br />

Bryan W. Gick, Chair<br />

Dept. of Linguistics, Univ. of British Columbia, 1866 Main Mall E 270, Vancouver, BC V6T 1Z1, Canada<br />

Contributed Papers<br />

All posters will be on display from 1:<strong>00</strong> p.m. to 5:<strong>00</strong> p.m. To allow contributors an opportunity to see other posters, contributors of<br />

odd-numbered papers will be at their posters from 1:<strong>00</strong> p.m. to 3:<strong>00</strong> p.m. and contributors of even-numbered papers will be at their<br />

posters from 3:<strong>00</strong> p.m. to 5:<strong>00</strong> p.m.<br />

5pSC1. Perceptual development of phonotactic features in Japanese<br />

infants. Ryoko Mugitani NTT Commun. Sci. Labs., NTT Corp., 2-4<br />

Hikari-dai, Seika-cho, Soraku-gun, Kyoto, 619-0237, Japan, Laurel Fais<br />

Univ. of British Columbia, Vancouver, BC, Canada V6T 1Z4, Sachiyo<br />

Kajikawa, Shigeaki Amano NTT Corp., Seika-cho, Soraku-gun, Kyoto,<br />

619-0237, Japan, and Janet Werker Univ. of British Columbia,<br />

Vancouver, BC, Canada V6T 1Z4<br />

Acceptable phonotactics differ among languages. Japanese does not<br />

allow consonant clusters except in special contexts and this phonotactic<br />

constraint has a strong effect on adults speech perception system. When<br />

two consonants follow one another in nonsense words, adult Japanese<br />

listeners hear illusory epenthetic vowels between the consonants. The current<br />

study is aimed at investigating the influence of language-specific phonotactic<br />

rules on infants’ speech perception development. Six-, <strong>12</strong>-, and<br />

18-month-old infants were tested on their sensitivity to phonotactic<br />

changes in words using a habituation-switch paradigm. The stimuli were<br />

three nonsense words: ‘‘keet /ki:t/,’’ ‘‘keets /ki:ts/,’’ and ‘‘keetsu<br />

/ki:tsu/.’’ ‘‘Keetsu’’ perfectly follows Japanese phonotactic rules.<br />

‘‘Keets’’ is also possible in devoicing contexts in fluent speech, but the<br />

acceptability of ‘‘keets’’ for adult native Japanese speakers is much less<br />

than ‘‘keetsu.’’ ‘‘Keet’’ is phonotactically impossible as a Japanese word.<br />

The results indicate the existence of a developmental change. Twelve<br />

months and older infants detected a change from the acceptable Japanese<br />

word ‘‘keetsu’’ to the possible but less acceptable word ‘‘keets.’’ However,<br />

discrimination between ‘‘keets’’ and the non-Japanese ‘‘keet’’ is seen only<br />

in 18-month infants. Implications for infants’ speech perception with relation<br />

to language specific phonotactical regularities will be discussed.<br />

5pSC2. Temporal evidence against the production of word-internal<br />

syllable structure. Melissa A. Redford Dept. of Linguist., <strong>12</strong>90 Univ.<br />

of Oregon, Eugene, OR 97403, redford@darkwing.uoregon.edu<br />

A production study examined whether the temporal pattern that cues<br />

onset cluster syllabification in English should be attributed to syllable<br />

structure per se or to coarticulatory processes that depend on segment<br />

identity and sequencing, but not on syllable structure. Experiment 1 tested<br />

for independent effects of syllable structure on intervocalic stop-liquid<br />

consonant duration. Three speakers produced stimuli in which sequence<br />

order, boundary strength, and stress were manipulated. The results showed<br />

that stimuli with word-internal sequences were produced with similar<br />

stop-liquid duration (S:L) ratios regardless of sequence order or stress.<br />

When a word boundary split the consonants, S:L ratios were much<br />

smaller, primarily because liquids were longer. Experiment 2 was aimed at<br />

distinguishing between alternate explanations for the pattern observed in<br />

Experiment 1. Data from three Russian speakers suggested that the English<br />

pattern was due to word-internal liquid reduction rather than to<br />

boundary lengthening. The S:L ratios in Russian were not affected by<br />

sequence order or boundary strength. Further, the ratios were significantly<br />

different from the word-internal S:L ratios of English, but not from the<br />

word-peripheral ones. Overall, the results argue against a causal relationship<br />

between syllable structure and articulatory timing, and for word-sized<br />

articulatory units in English.<br />

5pSC3. Perception of coarticulated speech with contrastively<br />

enhanced spectrotemporal patterns. Travis Wade and Lori Holt Dept.<br />

Psych., Carnegie Mellon Univ., 5<strong>00</strong>0 Forbes Ave., Pittsburgh, PA 15213,<br />

twade@andrew.cmu.edu<br />

High-level contrastive mechanisms cause perception of auditory<br />

events to be influenced by spectral and temporal properties of surrounding<br />

acoustic context, and may play a role in perceptual compensation for<br />

coarticulation in human speech. However, it is unknown whether auditory<br />

contrast is incorporated optimally to compensate for different speakers,<br />

languages and situations or whether amplification of the processes involved<br />

would provide additional benefit, for example, in the perception of<br />

hypoarticulated speech, under adverse listening conditions, or in an incompletely<br />

acquired language. This study examines effects of artificial<br />

contrastive modification of spectrotemporal trajectories on the intelligibility<br />

of connected speech in noise by native and non-native listeners. Adopting<br />

methods known to improve automatic classification of speech sounds,<br />

we model contrast-providing context as an averaged estimated vocal tract<br />

function LPC-derived log area ratio coefficient vector over a Gaussianweighted<br />

temporal window. Local coefficient values are adjusted from this<br />

context based on previously observed contrastive perceptual tendencies,<br />

and the intelligibility of the resulting speech is compared with that of<br />

unmodified trajectories across listener language backgrounds. Results are<br />

discussed with respect to implementation and applicability of general auditory<br />

processes.<br />

2617 J. Acoust. Soc. Am., Vol. 117, No. 4, Pt. 2, April <strong>20</strong>05 149th Meeting: Acoustical Society of America 2617<br />

5p FRI. PM

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!