Professional Documents
Culture Documents
INTRODUCTION
Research has established that qualities of auditory stimuli and innate abilities influence perceptual
accuracy, and that perceptual accuracy is highly
correlated with the accuracy of vocal pitch or fundamental frequency (F0) control.1,2 One method
used to evaluate the accuracy of F0 control is a pitch
matching task. Those individuals who have accurate pitch discrimination abilities tend to be accurate pitch matchers. This relationship has been
substantiated in numerous populations, including
school-age children, musicians, and accurate
singers.1,3,4 In most populations, a continuum exists
560
561
562
ROBERT E. MOORE ET AL
Interfering Tones
(Hz)
Comparison Tone
(Hz)
110
107/113/104/117
104/117/113/107
104/117/107/113
105
110
115
131
123/127/139/135
139/135/127/123
139/135/123/127
125
131
137
156
147/165/152/161
165/161/152/147
165/147/161/152
149
156
163
186
175/197/181/191
197/191/181/175
197/191/175/181
178
186
194
220
228/235/209/191
215/235/228/209
235/209/228/215
211
220
230
FIGURE 1. Stimuli timing pattern used for pitch discrimination task (top) and pitch discrimination with memory interference task (bottom).
563
Procedures
Pitch matching task
Each of eight stimulus complex tones was presented randomly at 75 dB SPL, one at a time via
a speaker. A total of 24 stimulus tones were presented in three trials. Subjects were instructed to
listen to the complex tone presented, and then to
vocally match the pitch of each target tone with
the vowel ah. Subjects were asked to maintain
the ah for 5 seconds, which was timed by the investigator. Subjects pitch matching responses were
recorded using a head-mounted microphone and the
Computerized Speech Lab (CSL; Kay Elemetrics
Corporation, Lincoln Park, NJ), with sampling
rate set at 44.1 kHz. Each subjects response was
saved in an individual file to the hard drive of the
computer for further analysis.
Pitch discrimination task
For the pitch discrimination task, participants
were seated at a desk inside the sound-attenuated
room, with a color monitor and computer mouse.
The investigator monitored the participants responses and progress on a networked computer
from outside the room. The investigator was also
able to see the subject using a closed-circuit camera
and monitor system. Presentation of each reference
and comparison tone condition was controlled by
ECos/Win software. The stimuli were presented
binaurally through E-A-R 5A insert earphones at
75 dB SPL.
Participants were instructed to listen for two
tones and to indicate whether the tones were
same or different in pitch. Participants indicated their judgment of the two tones by using
the computer mouse to click on the appropriate
icon displayed on the monitor. A tone pair was
not presented until the participant had made a judgment regarding the preceding pair. One block of 15
stimuli was presented to each participant as practice. Once it was shown that the participant understood and could complete the task, two further
experimental blocks were presented. The order of
presentation of the tone pairs was randomized individually for each block by the ECos/Win software.
All participant responses were saved to the computer hard drive for further analysis. Only the last
two blocks were scored for the experiment.
Journal of Voice, Vol. 21, No. 5, 2007
564
ROBERT E. MOORE ET AL
Analysis
Three dependent variables were measured in
this study: pitch matching accuracy, pitch discrimination accuracy, and pitch discrimination accuracy with memory interference. Pitch matching
ability was calculated by analyzing the middle
4 seconds of each participants vocal match. This
time frame was chosen to avoid measuring onset
and offset time, which could lead to variability
in the measurement. The fundamental frequency
of each sample was calculated by the CSL using
the Multi Dimensional Voice Profile software. An
average of the three vocal attempts for each target
tone was calculated. The frequency of the target
tone and average frequency of the pitch matching
attempts was used to calculate a difference in
semitones. Finally, the semitone differences for
each of the eight target tones were averaged to arrive at an overall semitone difference for each participant. A smaller score indicated more precise
pitch matching accuracy. Performance accuracy
for the pitch discrimination and pitch discrimination with memory interference tasks was measured
by calculating the percentage correct score for
each participant from a total of 30 stimulus
presentations.
A one-way repeated measures analysis of variance (ANOVA) was performed to compare the
means of the pitch discrimination and pitch discrimination with memory interference task to ensure that the interference tones had a significant
effect on pitch discrimination ability. The presence
and strength of the relationship between pitch
matching and pitch discrimination, as well as between pitch matching and pitch discrimination
with memory interference, was also evaluated.
This was done using Pearsons product-moment
correlations.
Journal of Voice, Vol. 21, No. 5, 2007
RESULTS
The descriptive statistics for the pitch matching,
pitch discrimination, and pitch discrimination with
memory interference tasks are shown in Table 2.
Standard deviations are shown with the pitch
matching results. In general, variance was greatest
with less precise pitch matching accuracy. As expected, the scores for the pitch discrimination
task were better than the scores for the pitch discrimination with memory interference task for
most of the participants. One participant scored
the same for both the pitch discrimination and pitch
discrimination with memory interference tasks,
PM (Semitones)
0.09
0.12
3.73
3.53
0.09
0.16
0.19
0.78
0.15
0.07
1.37
0.18
0.13
0.15
0.08
0.29
1.78
1.94
0.09
2.31
0.36
1.96
1.01
1.32
1.14
2.41
2.75
4.75
0.53
0.18
(0.08)
(0.09)
(3.89)
(3.37)
(0.07)
(0.09)
(0.10)
(0.90)
(0.09)
(0.06)
(0.41)
(0.07)
(0.65)
(0.13)
(0.05)
(0.11)
(1.16)
(2.01)
(0.09)
(1.37)
(0.08)
(1.23)
(1.72)
(1.91)
(1.55)
(2.44)
(2.06)
(3.74)
(0.89)
(0.03)
PD (%)
PDI (%)
100.0
70.0
37.7
46.7
93.3
93.3
66.7
43.3
96.7
96.7
63.3
83.3
86.6
80.0
100.0
93.3
50.0
53.3
96.7
93.3
100.0
43.3
46.7
53.3
86.6
73.3
56.7
70.0
76.7
96.7
66.7
66.7
50.0
60.0
60.0
93.3
60.0
50.0
43.3
63.3
43.3
43.3
60.0
66.7
46.7
36.7
40.0
36.7
63.3
60.0
70.0
40.0
43.3
40.0
83.3
60.0
53.3
60.0
66.7
66.7
565
DISCUSSION
Perceptual accuracy is needed for accurate F0
control. One method used to study fundamental frequency control is a pitch matching paradigm. Previous research has demonstrated a relationship
between pitch matching ability and pitch discrimination ability. There is variability in the ability of
individuals to match and discriminate pitch. To
complete a vocal pitch matching task, an individual
must hear the stimulus and process the pitch of the
stimulus. In other words, pitch discrimination is
a part of pitch matching. Based on these first two
functions, the individual then programs motor neurons and reproduces the sound as closely as possible. During the time in which the individual is
preparing to produce the sound, the processed pitch
must be held in memory. Previous researchers have
suggested that short-term memory for pitch is
unique in place and pattern of storage.
The purpose of this study was to examine the
role of pitch memory as it relates to ones ability
to match and discriminate pitch. Specifically, we
were interested in investigating the relationship between pitch matching and pitch discrimination and
how that relationship is affected by interfering with
pitch memory. In agreement with previous studies,
the results of the present study demonstrated a relationship between ones ability to match pitch and
ones ability to discriminate pitch. Most individuals
with increased pitch discrimination accuracy were
able to match pitch more accurately. Pitch discrimination ability explained 44% of the variance in
pitch matching ability. Because pitch discrimination accuracy appears to have a significant influence on an individuals pitch matching accuracy,
any change in pitch discrimination accuracy caused
by pitch memory interference should be greater in
those who are the most precise pitch matchers, if
one assumes that those individuals typically make
better use of pitch memory in the pitch discrimination phase of pitch matching. Other factors that
may influence pitch matching accuracy are motor
control and, of course, vocal accuracy.
To test the role of pitch memory, four interference
tones were placed between the two tones used in the
pitch discrimination portion of the study. Previous
research has indicated that such a paradigm results
Journal of Voice, Vol. 21, No. 5, 2007
566
ROBERT E. MOORE ET AL
in decreased pitch discrimination ability. The decline in pitch discrimination has been interpreted
as pitch memory interference.7 In the present study,
a decline in pitch discrimination ability was seen for
most individuals, and overall, there was a significant
decrease in pitch discrimination ability. Under this
condition, correlational analysis indicated that there
was no significant relationship between pitch
matching and pitch discrimination with memory interference. Pitch discrimination ability in the presence of memory interference explained only 6% of
the variance in pitch matching ability. This would
suggest that pitch memory plays a significant role
in both pitch discrimination and pitch matching. It
appears that an important part of fundamental frequency control, as measured by pitch matching, is
pitch discrimination ability with its inherent component, memory for pitch.
No attempt was made in participant identification
to preselect the participants based on vocal musical
ability. However, previous research results would
suggest that those with superior pitch matching
ability could be predicted to have superior vocal accuracy than those with poorer pitch matching accuracy. Accordingly, as shown in Table 3, the
participants were divided into three groups based
on their pitch matching abilities. One group was
made up of those participants with the ability to
match pitch within 0.1 semitones. These are termed
the superior pitch matchers. A second group was
composed of participants who were unable to
match pitch within 1.0 semitone. These were
Group
Superior (5)
Intermediate (12)
Poor (13)
PD (%)
PDI (%)
Difference (%)
97.34
80.55
59.55
60.00
60.28
51.53
37.34
20.27
8.02
567