Wind noise in hearing aids
Harvey Dillon, Richard Katsch, Inge Roe,
National Acoustic Laboratories,
Australian Hearing,
With the support of GN
Resound, Oticon, Phonak, &
Widex
The problem
Wind + hearing aid = noise
But why, and how, and how
bad is the problem?
Dillon, Roe & Katsch, NAL.
Turbulence
U
L
f = US/L (Hz)
S = Strouhal number
Dillon, Roe & Katsch, NAL.
1/3 octave level (dB SPL)
Spectrum of noise behind wire
110
100
90
80
1 cm
2 cm
4 cm
mic alone
70
60
50
25
50
100
200
400
800
1600
3150 6300
Frequency (Hz)
Dillon, Roe & Katsch, NAL.
Laser Doppler Velocimeter
Dillon, Roe & Katsch, NAL.
Velocity down the wind tunnel (CIC aid)
Dillon, Roe & Katsch, NAL.
Velocity out of head (CIC aid)
Dillon, Roe & Katsch, NAL.
Velocity upwards (CIC aid)
Dillon, Roe & Katsch, NAL.
Turbulent velocity (CIC aid)
Dillon, Roe & Katsch, NAL.
Turbulent velocity (ITE aid)
Dillon, Roe & Katsch, NAL.
Turbulent velocity (BTE aid)
Dillon, Roe & Katsch, NAL.
Effect on turbulence of distance from head
Dillon, Roe & Katsch, NAL.
Solution 1: Extend the microphones
Dillon, Roe & Katsch, NAL.
Wind velocity inside and outside the concha
Dillon, Roe & Katsch, NAL.
Turbulence in the concha
Tragus
Dillon, Roe & Katsch, NAL.
Sensitivity of ITC
Tragus
Dillon, Roe & Katsch, NAL.
Smoothness of ITE
Tragus
Dillon, Roe & Katsch, NAL.
Noise measurements: Specially
designed wind-tunnel.
Noise measurements: Outlet of Wind
Tunnel
Noise measurements: Outlet of Wind
Tunnel
z
x
y
Wind velocity
5
m/sec
 18 km/hr
 11 m.p.h.
 Level 3 on 13 point Beaufort Scale
 Flags unfurl but droop
 Scattered whitecaps
 Gentle Breeze
 Exceeded 6% of time
Dillon, Roe & Katsch, NAL.
1/3 octave level (dB SPL)
Noise at BTE position
100
90
-90
-60
-30
0
30
60
90
80
70
60
50
40
25
50
100
200
400
800 1600 3150 6300
Frequency (Hz)
Dillon, Roe & Katsch, NAL.
1/3 octave level (dB SPL)
Aid comparison at 0 degrees
100
90
80
BTE
ITE
ITC
CIC
70
60
50
40
25
50
100
200
400
800 1600 3150 6300
Frequency (Hz)
Dillon, Roe & Katsch, NAL.
KEMAR at 0 degrees to wind
Dillon, Roe & Katsch, NAL.
1/3 octave level (dB SPL)
Aid comparison at 30 degrees
100
90
80
BTE
ITE
ITC
CIC
70
60
50
40
25
50
100 200 400 800 1600 3150 6300
Frequency (Hz)
Dillon, Roe & Katsch, NAL.
100
90
BTE
ITE
ITC
CIC
BTE no ear
80
70
60
50
16
00
31
50
63
00
80
0
40
0
20
0
10
0
50
40
25
1/3 octave level (dB SPL)
Aid comparison at 30 degrees
Frequency (Hz)
Dillon, Roe & Katsch, NAL.
Solution 2:
Dillon, Roe & Katsch, NAL.
Solution 2: Remove the pinnae
Dillon, Roe & Katsch, NAL.
Solution 2: Remove the pinnae
Dillon, Roe & Katsch, NAL.
1/3 octave level (dB SPL)
ITE noise versus azimuth
100
90
80
100 Hz
200 Hz
2 kHz
70
60
50
-90
-60
-30
0
30
60
90
Aziimuth (degs)
Dillon, Roe & Katsch, NAL.
KEMAR at -50 degrees to wind
Dillon, Roe & Katsch, NAL.
1/3 octave level (dB SPL)
Aid comparison at -50 degrees
100
90
80
BTE
ITE
ITC
CIC
70
60
50
40
25
50
100
200
400
800 1600 3150 6300
Frequency (Hz)
Dillon, Roe & Katsch, NAL.
1/3 octave level (dB SPL)
Aid comparison at -50 degrees
100
90
BTE
ITE
ITC
CIC
BTE no ear
80
70
60
50
40
25
50
100 200 400 800 1600 3150 6300
Frequency (Hz)
Dillon, Roe & Katsch, NAL.
Solution 3: Shed the vortices gracefully
Dillon, Roe & Katsch, NAL.
1/3 octave level (dB SPL)
ITE noise versus azimuth
100
90
80
100 Hz
200 Hz
2 kHz
70
60
50
-90
-60
-30
0
30
60
90
Frequency (Hz)
Dillon, Roe & Katsch, NAL.
KEMAR at -90 degrees to wind
Dillon, Roe & Katsch, NAL.
Solution 4: Keep the aid towards the wind
Dillon, Roe & Katsch, NAL.
ITE noise re CIC noise
10000
15
10
1000
5
dB
0
100
-5
-80
-60
-40
-20
0
20
40
60
80
-10
Dillon, Roe & Katsch, NAL.
Factors affecting wind noise
 Levels
are very intense
 Obstacles (head, pinna, tragus) act as:
– Wind guards
– Turbulence source
– Turbulence shredder
 Large
obstacles create low-freq turbulence
– head
 Medium
obstacles create mid-freq turbulence
– pinna
 Small
obstacles create high-freq turbulence
– tragus, inlet port
Dillon, Roe & Katsch, NAL.
Other observations
 As
wind speed increases:
– noise levels increase
– frequency spectrum extends upward
 Two
microphone ports produce:
– correlated noise if a common source
(e.g. head or pinna)
– uncorrelated noise if separate
sources (e.g. inlet port)
Dillon, Roe & Katsch, NAL.
Potential solutions
Wear one aid and orient the head
 Wear a scarf
 Don’t fit a BTE
 Don’t fit a fixed directional microphone
 Low distortion input circuitry

– up to at least 110 dB SPL

Low-cut filtering
– especially over the vent-transmitted range
 Smooth
design
 Electronic signal processing from multiple
microphones
Dillon, Roe & Katsch, NAL.
For a copy of this talk, send an
Email to:
Research@NAL.GOV.AU