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.
Speech recognition with varying numbers and types of competing talkers by normal-hearing, cochlear-implant, and implant simulation subjectsa)
a)Portions of this work were presented in “Two’s company; three’s a crowd. Speech recognition with competing talkers: normally-hearing, cochlear implant and CI simulation subjects,” American Auditory Society meeting, Arizona, March 2006.
Rent:
Rent this article for
USD
10.1121/1.2805617
/content/asa/journal/jasa/123/1/10.1121/1.2805617
http://aip.metastore.ingenta.com/content/asa/journal/jasa/123/1/10.1121/1.2805617

Figures

Image of FIG. 1.
FIG. 1.

Mean SRT as a function of masker type in seven normal-hearing, seven cochlear-implant subjects, and seven normal-hearing subjects using an eight-channel sine-carrier cochlear-implant simulation. Target material was HINT sentences spoken by a male. The squares represent the mean SRT for normal-hearing subjects. The circles represent the mean SRT for cochlear-implant subjects. The triangles represent the mean SRT for cochlear-implant simulation subjects. The hexagons and inverted triangles represent the SRTs for the right and left ears, respectively, for subject CINH001, who has a cochlear implant in the right ear and virtually normal hearing in the left ear. Error bars represent one standard deviation. For clarity, only the upward bar is shown for the cochlear-implant users, and only the downward bar for the normal-hearing and simulation subjects. Lines joining the points are purely for clarity, and are not suggesting a functional relation, due to the maskers being categorically different. The asterisks represent a SRT value significantly different from the SRT with a steady-state noise masker.

Image of FIG. 2.
FIG. 2.

Mean SRT as a function of number of talker maskers in seven normal-hearing, seven cochlear-implant subjects, and seven normal-hearing subjects using an eight-channel sine-carrier cochlear-implant simulation. Target material was HINT sentences spoken by a male. The squares represent the mean SRT for normal-hearing subjects. The circles represent the mean SRT for cochlear-implant subjects. The triangles represent the mean SRT for cochlear-implant simulation subjects. Error bars represent ± one standard deviation.

Image of FIG. 3.
FIG. 3.

Mean fundamental frequency (F0) for the target and each of the masker voices. The target HINT sentences were spoken by a male. Error bars represent ± one standard deviation. As expected, the child has the highest F0, followed by the female, then the two males.

Image of FIG. 4.
FIG. 4.

Mean SRT as a function of masker type and masker reversal in normal-hearing subjects. Target material was HINT sentences spoken by a male. Maskers were one or two female or male talkers played either forward or reversed. The squares represent the mean SRT when masker sentence was played forward. The triangles represent the mean SRT when the masker segment was reversed. Sixteen subjects were evaluated for the single-talker maskers (f and m), seven subjects for the two-talker maskers (f2, m1f1, and m2). Error bars represent one standard deviation. For clarity, only the upward bar is shown for the masker-forward condition, and only the downward bar for the masker-reversed condition.

Image of FIG. 5.
FIG. 5.

Mean SRT for six different child maskers in eight normal-hearing subjects. Target material was HINT sentences spoken by a male. Error bars represent ± one standard deviation. Four female and two male children were used, with ages ranging from seven to nine years, with voice fundamental frequencies from . The mean SRT, however, was approximately the same for all child maskers.

Image of FIG. 6.
FIG. 6.

Mean SRT for one, two, three, four, and six child maskers in eight normal-hearing subjects. Target material was HINT sentences spoken by a male. Error bars represent ± one standard deviation. The mean SRT gradually increases as the number of interfering talkers increases.

Image of FIG. 7.
FIG. 7.

Mean F0 and syllabic rate for all the single-talker maskers. The bars represent the mean F0 for each talker, using the left-hand axis. The circles represent the means syllabic rate for each talker, using the right-hand axis. Error bars represent ± one standard deviation. The male F0 is significantly lower than all the other maskers. All child maskers except child-A have a significantly higher F0 than the female. There is much variation in syllabic rate, with some children having higher rate than the female, and some comparable.

Image of FIG. 8.
FIG. 8.

Temporal envelope modulation data. Graphs show envelope spectra differences between masker and target for seven octave bands, depicted by modulation index difference (MI diff) as a function of third octave band modulation frequency. Data are shown for three maskers: female, male, and child (child-E).

Tables

Generic image for table
TABLE I.

Masking material used in the three experiments, including the abbreviations utilized in this paper. Each experiment used only a subset of the maskers, due to the limited target material available; the conditions used are indicated by ‘X’. All talker maskers spoke sentences. The adults spoke IEEE sentences; the children spoke sentences from the CMU Kids Corpus.

Generic image for table
TABLE II.

Details of child maskers. Child talkers were obtained from the CMU Kids Corpus. They read aloud from grade-appropriate Weekly Reader Stories. Only sentences spoken fluently without mistakes, hesitation, or background noise were included.

Generic image for table
TABLE III.

t and p values from the paired t tests performed on the mean SRT with female, child, and male masker in normal-hearing, cochlear-implant, and cochlear-implant simulation subjects. Values shown are two tailed, with six degrees of freedom. The p values were compared to 0.017 using the Bonferroni correction for multiple comparisons. All three groups showed a significantly better mean SRT for the female masker than both the child and male maskers. There was no significant difference between the mean SRT for the child and male maskers.

Loading

Article metrics loading...

/content/asa/journal/jasa/123/1/10.1121/1.2805617
2008-01-01
2014-04-18
Loading

Full text loading...

This is a required field
Please enter a valid email address
752b84549af89a08dbdd7fdb8b9568b5 journal.articlezxybnytfddd
Scitation: Speech recognition with varying numbers and types of competing talkers by normal-hearing, cochlear-implant, and implant simulation subjectsa)
http://aip.metastore.ingenta.com/content/asa/journal/jasa/123/1/10.1121/1.2805617
10.1121/1.2805617
SEARCH_EXPAND_ITEM