Download - Galster.net

Speech Recognition in Noise with Four Remote Microphone Technologies
Krishna S. Rodemerk, Au.D. & Jason A. Galster, Ph.D.
Starkey Hearing Technologies
Test Conditions (24 Total)
HINT 12 Feet Remote Microphone Only
1) Unaided (6 Feet and 12 Feet)
15
the hearing aid, overcoming the negative effects of talker distance, ambient noise, and
2) Omnidirectional Hearing Aid Microphone Only (6 Feet and 12 Feet; 3 different
reverberation (Chisolm et al., 2007). Traditionally, remote microphone technology was most
sets of hearing aids)
Wireless Bluetooth® paired with near-field magnetic induction (NFMI)
systems were evaluated at each visit: hearing aid only, remote microphone only and remote
2.4 GHz
microphone plus hearing aid microphone. The HINT sessions were at least two weeks apart
regular hearing aid wearers and none had remote microphone experience. All participants
talker (p <0.001).
HINT SNR-50 (dB)
Box plot whiskers represent the 5th and 95th percentiles and the shaded box shows the 25th
0
*
*
*
-10
participants received. However, it was unexpected that performance was similar when
ic
ic
+
90
0
2.
4
M
Hz
GH
z
M
M
ic
ic
M
oo
th
would decrease with activation of the hearing aid microphone (Boothroyd & Iglehart, 1998).
We speculate that the noise levels used in this study were not sufficiently high to cause the
HINT 6 Feet Remote Microphone Only
Figure 4. HINT SNR-50 is shown as a function of microphone condition. Participants were
seated 6 feet from the talker. The remote microphones were set such that there was equal
contribution from the hearing aid microphone and the remote microphone.
15
HINT 12 Feet Remote Microphone + Hearing Aid Mic
was presented from four speakers surrounding the participant (45°, 135°, 225°, and 315°).
Participants were tested at 6 and 12 feet from the talker loudspeaker. The order of remote
microphone conditions was counterbalanced across participants.
presentations levels may reveal this effect.
Though not discussed in detail here, the observed benefits were acoustically predictable in
affects the quality and level of the transmitted signal. Successful use of any remote
10
microphone technology should include counseling on proper use and orientation.
-5
*
*
-15
-20
0
*
-5
*
-10
*
ic
Chisolm, T.H., Noe, C.M., McArdle, R., & Abrams, H. (2007). Evidence for the use of hearing assistive technology by adults: The role of the FM system. Trends in Amplification,
11(2), 73-89
M
ot
e
Re
m
-15
M
ic
+
ic
M
GH
z
2.
4
90
0
M
Hz
oo
th
M
M
ic
ic
+
+
HA
HA
M
ic
M
ic
HA
M
ic
HA
+
FM
Bl
ue
t
2.
4
GH
z
HA
HA
Hz
M
0
90
O
nl
y
O
nl
y
O
nl
y
HA
Un
ai
de
d
-20
oo
th
/F
M
Figure 2. HINT SNR-50 is shown as a function of microphone condition. In the figure
above, participants were seated 6 feet from the talker. The remote microphones were
streaming only (no contribution from the hearing aid or environmental microphone).
References
Boothroyd, A., & Iglehart, F. (1998). Experiments with classroom FM amplification. Ear &
Hearing, 19(3), 202-217.
*
GH
z
2.
4
Hz
M
0
90
O
nl
y
O
nl
y
ic
M
ot
e
Re
m
ot
e
Re
m
oo
th
FM
Re
m
ot
e
M
M
ic
ic
O
nl
y
O
nl
y
O
nl
y
HA
GH
z
2.
4
90
0
M
Hz
HA
O
nl
y
O
nl
y
HA
oo
th
/F
M
Un
ai
de
d
-25
5
Bl
ue
t
*
-10
*
HINT SNR-50 (dB)
HINT SNR-50 (dB)
0
conditions. Each condition was assessed one time using two lists (20 sentences for each
condition). The speech was presented at 0° azimuth while continuous, speech shaped noise
expected effect. Future studies using a test protocol that allows for higher noise
nature. Of clinical note, proximity of the remote microphone to a talker's mouth greatly
15
Bl
ue
t
only, remote microphone only and hearing aid microphone plus remote microphone
5
Bl
ue
t
required for correct repetition of 50% of the sentences, was recorded for the hearing aid
comparing remote microphone test conditions with and without the hearing aid microphone
activated. Previous reports of this comparison suggest that remote microphone benefit
+
HA
HA
M
M
ic
ic
+
+
HA
HA
M
M
ic
O
nl
y
Bl
ue
t
Bl
ue
t
FM
oo
th
/F
M
when compared to unaided and any hearing aid only condition (p <0.001).
HA
Un
ai
de
d
-20
prescribed targets in all test conditions. Verification of the remote microphone in-situ
the accuracy of the listener’s response. The HINT SNR-50, or the signal-to-noise ratio
when comparing remote microphone streaming, both with and without the hearing aid
with all four of the tested remote microphone systems.
*
10
presented at a fixed level (55 dB SPL) and the sentence levels were varied depending on
the hearing aid microphone was introduced.
-15
solid black line inside the box. Asterisks indicate significant improvement in HINT SNR-50
transparency was confirmed for all fittings, ensuring that the hearing aid output matched
The Hearing in Noise Test (HINT) was used in its adaptive form during which the noise was
participants were seated both 6 feet and 12 feet from the talker (p >0.05).
As expected, distance between the participant and talker did not affect the benefit that the
percentiles. Mean is shown as the dotted line and the median performance is the
prescribed to age appropriate DSL v5.0 targets (Scollie et al., 2005). Remote microphone
Test Material
No difference was observed between remote microphone only performance when
microphone enabled, to hearing aid only and unaided listening. This benefit was observed
-5
GH
z
and 75
Real-ear measures were completed using an Audioscan Verifit. All hearing aids were
puts.
6 feet and 12 feet from the talker (p <0.001).
Participants in this study demonstrated significantly improved speech recognition in noise
Verification
phones with the remote microphone, the audio mixture was set to equal levels for both in-
condition across all four remote microphone systems when participants were seated both
The results of this study agree with previous reports of remote microphone benefit.
5
O
nl
y
Results
th
pathway to the measurement. When the test condition combined the hearing aid micro-
significantly increased speech recognition in noise when compared to the hearing aid only
Discussion
2.
4
risks or benefits.
output was done at test condition distances to allow for contribution of the direct acoustic
Performance with the combination of remote microphones and hearing aid microphones
15
10
consent process that reviewed study methodology, their required involvement, and any
used a standard size 13 tube.
microphone systems when participants were seated both 6 feet and 12 feet from the
HINT 6 Feet Remote Microphone + Hearing Aid Mic
were financially reimbursed for their participation in the study and completed an informed
system(s). All participants were fit with occluded, full shell custom earmolds; each earmold
-20
-25
Figure 1. Five speaker array used for HINT testing.
In the test set-up, the average Reverberation Time (RT60) was 0.4 seconds across frequency and the Direct to
Reverberant Ratio was 6dB.
sensorineural hearing loss participated in this study. Ten out of sixteen participants were
bilateral set of hearing aids was paired with the respective, compatible remote microphone
Performance with remote microphone streaming significantly increased speech recognition
Performance was similar between three of the four remote microphone conditions when
Sixteen adults aged 52 to 81 years (mean = 68.5 years) with mild to moderately-severe,
and programming software from three hearing aid companies were used in this study. Each
are discussed below.
in noise when compared to the hearing aid only condition across all four remote
Figure 3. HINT SNR-50 is shown as a function of microphone condition. In the figure
above, participants were seated 12 feet from the talker. The remote microphones were
streaming only (no contribution from the hearing aid or environmental microphone).
Participants
Commercially available behind-the-ear (BTE) hearing aids, remote microphone systems,
*
Post-hoc analyses used Tukey’s pairwise multiple comparison test, meaningful observations
-15
to mitigate any learning effects.
Methods
Devices
*
*
*
-10
HA
900 MHz
-5
remote microphone systems; remote microphone/hearing aid mix)
Testing was done over two visits. Three microphone conditions for two remote microphone
significant main effect of microphone condition: F(7,105) = 233.5, p <0.001 for six feet
F(7,105) = 149.4, p <0.001 for 6 feet remote microphone plus hearing aid microphone and
Maybe we could use this section as a lab design
section and show an image of our speaker array?
Hz
Frequency Modulation (FM)
4) Remote Microphone Streaming + Hearing Aid Microphone (6 Feet and 12 Feet; 4
0
M
protocol:
systems)
Four RMANOVAs were completed for the conditions shown in Figures 2-5; each showed a
F(7,105) = 214.8, p <0.001 for 12 feet remote microphone plus hearing aid microphone.
0
microphone systems. Each of the four systems used a different wireless audio transmission
Or
O
nl
y
This study documented speech recognition in noise with four commercially available remote
10
5
3) Remote Microphone Streaming Only (6 Feet and 12 Feet; 4 remote microphone
Results
remote microphone only; F(7, 105) = 255.2, p <0.001 for 12 feet remote microphone only;
90
often used in educational settings in the form of frequency modulated, or FM, systems.
Would probably look nicer, but could be confusing?
HA
Remote microphones provide a direct audio input that results in a consistent input level to
Results
Bl
Un
ue
ai
to
de
ot
h/
d
FM
HA
90
O
nl
0
y
M
Hz
HA
O
2.
nl
4
y
G
Hz
FM
HA
Re
O
Bl
nl
m
ue
y
ot
to
e
ot
M
h
ic
Re
O
nl
m
90
y
ot
0
e
M
M
Hz
ic
Re
O
nl
m
2.
y
ot
4
e
G
M
Hz
ic
Re
O
nl
m
y
ot
e
M
ic
O
nl
y
A remote microphone is a wireless transmitter that routes audio signals to a hearing aid.
Methods
HINT SNR-50 (dB)
Introduction
Wondering if we should put both HINT and ANL
stats here and display both graphs to the right?
Figure 5. HINT SNR-50 is shown as a function of microphone condition. Participants were
seated 12 feet from the talker. The remote microphones were set such that there was equal
audio contribution from the hearing aid microphone and the remote microphone.
Lewis, M.S., Crandell, C.C., Valente, M., & Horn, J.E. (2004). Speech perception in noise:
Directional microphones versus frequency modulation (FM) systems. Journal of the American Academy of Audiology, 15, 426-439.
Nilsson M., Soli S.D., & Sullivan J.A. (1994). Development of the Hearing In Noise Test for
the measurement of speech reception thresholds in quiet and in noise. Journal of the
Acoustical Society of America, 95(2), 1085-99.
Scollie, S., Seewald, R., Cornelisse, L., Moodie, S., Bagatto, M., et al. (2005). The Desired
Sensation Level Multistage Input/Output Algorithm. Trends in Amplification 4(9), 159-197.

Download Report