METHODS

1
The 149th Meeting of the ASA in Vancouver, Canada
on May 17, 2005
Downloadable from http//mypage.iu.edu/knagao/res
earch.html
2pSC8. Sociological effects on vocal aging Age
related F0 effects in two languages Kyoko Nagao
Department of Linguistics, Indiana
University Department of Speech and Hearing
Sciences, Indiana University

• METHODS
• Recording experiments
• Talkers
• Sixty talkers participated in the recording.
  They are subdivided into 12 groups by their age
  group (Young, Middle-age, or Elderly), sex
  (Female or Male), and native language (English or
  Japanese). Data of the middle-age talker groups
  will be presented here.

• INTRODUCTION
• Fundamental frequency and aging
• Fundamental frequency (F0) of female talkers
  decreases with advanced aging(Krook, 1988 Xue
  Mueller, 1996 Xue Deliyski, 2001 Xue, Neeley,
  Hagstrom, Hao, 2001), while male f0 tends to
  increase with advanced aging (Hollien Shipp,
  1972 Mysak, 1959 Mysak Hanley, 1958 Xue
  Deliyski, 2001).
• Perception of talkers age
• Listeners can estimate the age of a speaker
  fairly accurately from their speech (Ptacek and
  Sander, 1966 Ryan and Burk, 1974).
• Acoustic factors
• Relationship between F0 and age perception is
  controversial, but most of the studies showed
  that f0 is not significant factor correlated with
  age perception. (Ryan and Burk, 1974 Braun and
  Rieveld, 1995)
• Speech rate has also been documented as either
  a weak or strong cue. (Braun and Rietveld, 1995
  Ptacek and Sander, 1966 ).
• Acoustic properties related with voice
  qualities are strongly correlated with perceived
  age (Ryan and Burk, 1974).
• Socio-cultural factors
• Pitch can convey various linguistic aspects of
  speech. Japanese females tend to employ higher
  portion of their pitch ranges. It is considered
  that this is due to a socio-cultural constraint
  on females in Japanese societies (Loveday, 1981,
  Ohara, 1992).
• Stereotypes of speech associated with a certain
  age group might be different in different
  cultures.
• Listeners familiarity with a certain age group
  might affect accuracy of age perception.
  Cohabitation with grandparent(s) is more common
  in Japan than in the US XX vs. X.
• ? Does these sociological factors cause a
  difference in age perception?
• RESEARCH QUESTIONS

RESULTS Rejuvenation vs. Control Maturation
vs. Control
Figure 3. Mean f0 values (top left) and mean
durations (top right) of each talker group in
each condition.
Figure 1a. Mean F0 values of each talker in the
Rejuvenation condition against the values in the
Control condition.
Figure 1b. Mean F0 values of each talker in the
Maturation condition against the values in the
Control condition.

• When talkers pretend to be 20 years old younger
  than their age, their mean f0 values were
  increased compared with the mean f0 values in the
  control condition.
• Regardless of the conditions, Japanese females
  exhibited higher f0 values than English females.
  Men did not exhibit such differences.
• All the talkers seemed to produce the utterances
  at faster rate in the rejuvenation condition.

4. When talkers pretend to be 20 years old older
than their age, their mean f0 values slightly
decreased for some females, but did not change
much with the mean f0 values of the control
condition.
Table. Sheffes post-hoc test results for F0
values ( indicates significant results, while
n.s. stands for not significant.)
Figure 4. Mean age differences (PA-CA) between
listeners perceived age (PA) and chronological
age (CA).
7. In spite of the small f0 and temporal
differences between Maturation condition and
Control condition, Perceived age for each talker
group was older than the actual age. These
results suggest that there are other acoustic
properties to indicate talkers age than mean f0
values and durational information.

• SUMMARY
• For most of the talkers (especially females),
  higher pitch is related with speech associated
  younger talkers.
• Lower pitch seems associated with older talkers
  for some female talkers, but not for male
  talkers. These relations are similar for both
  Japanese and English talkers.
• Fast speaking rate seems associated with younger
  speech. However, slow rate speech is not
  necessary related with older speech.
• Listeners can perceive the age differences
  introduced by talkers in the age-disguised
  speech.

Figure 2a. Mean durations of each talker in the
Rejuvenation condition against the values in the
Control condition.
Figure 2b. Mean durations of each talker in the
Maturation condition against the values in the
Control condition.
References Loveday, L. (1981). Pitch, politeness
and sexual role An exploratory investigation
into the pitch correlates of English and Japanese
politeness formulae. Language and Speech, 24,
71-89. Ptacek and Sander (1966)
Table. Sheffes post-hoc test results for
duration ( indicates significant results, while
n.s. stands for not significant.)

• When talkers pretend to be 20 years old younger
  than their age, all the talkers seemed to produce
  the utterances at faster rate in the rejuvenation
  condition.
• However, speaking rate was not systematically
  changed in the maturation condition.