1887
banner image
No data available.
Please log in to see this content.
You have no subscription access to this content.
No metrics data to plot.
The attempt to load metrics for this article has failed.
The attempt to plot a graph for these metrics has failed.
The interaction of glottal-pulse rate and vocal-tract length in judgements of speaker size, sex, and agea)
a)Portions of this work were presented in “The perception of sex and size in vowel sounds,” British Society of Audiology, UCL London, United Kingdom, 2004, and “Perception of speaker size and sex of vowel sounds,” Acoustical Society of America, Vancouver, Canada, 2005.
Rent:
Rent this article for
USD
10.1121/1.2047107
/content/asa/journal/jasa/118/5/10.1121/1.2047107
http://aip.metastore.ingenta.com/content/asa/journal/jasa/118/5/10.1121/1.2047107
View: Figures

Figures

Image of FIG. 1.
FIG. 1.

The open circles show the GPR and VTL combinations of the stimuli used in the speaker size and sex/age categorization experiments. The circles in the top panel show the “narrower” range of sample points (GPRs of 80, 105, 137, 179, 234, 306, and ; VTLs of 7.8, 9.3, 11.0, 13.2, 15.7, 18.7, and ). The bottom panel shows the “wider” range (GPRs of 61, 87, 125, 179, 256, 366, and ; VTLs of 6.5, 8.2, 10.4, 13.2, 16.7, 21.3, and ). The four ellipses show the normal range of GPR and VTL values in speech for men (M), women (W), boys (B), and girls (G), derived from the data of Peterson and Barney (1952). Each ellipse contains 99% of the individuals from the respective category.

Image of FIG. 2.
FIG. 2.

Speaker size judgements collapsed across VTL (upper panel) and GPR (lower panel), separately for the narrower and the wider ranges (cf. Fig. 1). The arrows on the inset show the dimension over which the data were collapsed. The open circles show the data from the narrower range and the solid circles from the wider range. The dotted line is the best fitting line for the wider range; the dashed line is the best fitting line for the narrower range, and the solid thick line is the best fitting line for the combined data. The error bars are one standard error of the mean (calculated from the average of the eight listeners, where each listener’s average is based on the seven values per point over which the data were collapsed). Each datum point is based on 1400 trials.

Image of FIG. 3.
FIG. 3.

Perceived size (in color) as a function of GPR and VTL on logarithmic axes. The size scale from “very short” to “very tall,” is represented by the spectrum of colors from dark-blue (1) to brown-red (7). The points where speaker size ratings were measured are shown by the open circles; between the data points, the surface was derived by interpolation. The data were averaged across all five vowels and eight listeners, so each point is based on 200 trials. The four ellipses show the range of GPR and VTL in speech for men (M), women (W), boys (B), and girls (G), as derived from the data set of Peterson and Barney (1952).

Image of FIG. 4.
FIG. 4.

Sex and age categorizations. The data are presented as 2D surface plots with color showing probability of assigning a given GPR-VTL combination to one of four categories (man, woman, boy, or girl). The points where sex/age judgements were collected are shown by the open circles; between the data points the surface was derived by interpolation. At each GPR-VTL point, the probabilities from the four panels sum to 1 (imagine the four separate 2D maps stacked vertically and aligned over each other). The data is averaged across all five vowels and eight listeners (each sample point probability based on 200 trials). The dotted black contour line marks the classification threshold, that is, a probability of consistently choosing one category out of the four available. The region of GPR-VTL values enclosed by this line defines a region categorized as one particular sex or age. The four ellipses show the range of GPR and VTL in speech for men (M), women (W), boys (B), and girls (G), as derived from the data set of Peterson and Barney (1952).

Image of FIG. 5.
FIG. 5.

Perceived speaker size as a function of GPR, for VTLs of 6.5, 13.2, and . The open and solid circles show data from the narrower and wider stimulus ranges, respectively. The solid lines show the best-fitting regression lines for perceived speaker size rating as a function of the natural logarithm of GPR. The error bars are one standard error of the mean (calculated from the average of the eight listeners). Each datum point is based on 200 trials.

Image of FIG. 6.
FIG. 6.

Perceived speaker size as a function of VTL, for GPRs of 61, 179, and . The solid lines show the best-fitting regression lines for speaker size rating as a function of the natural logarithm of VTL. For all other details see Fig. 5.

Loading

Article metrics loading...

/content/asa/journal/jasa/118/5/10.1121/1.2047107
2005-11-01
2014-04-20
Loading

Full text loading...

This is a required field
Please enter a valid email address
752b84549af89a08dbdd7fdb8b9568b5 journal.articlezxybnytfddd
Scitation: The interaction of glottal-pulse rate and vocal-tract length in judgements of speaker size, sex, and agea)
http://aip.metastore.ingenta.com/content/asa/journal/jasa/118/5/10.1121/1.2047107
10.1121/1.2047107
SEARCH_EXPAND_ITEM