Wind is a daily phenomenon that can have serious and debilitating effects on hearing aid and cochlear implant (CI) users. People with normal hearing can also experience such negative effects when they use cellular phones. Imagine: You receive a call from a friend who's in a windy environment. He wants to have a conversation with you, but all you can hear is wind noise! In this scenario, your friend is holding a device with its microphone under the influence of wind. You, on the other hand, are holding a device that outputs the signal from that microphone. For hearing aid or CI users, both the microphone and the receiver are located on the same device.
Wind noise is generated in two distinctive ways: (1) when fast-moving air molecules bombard the microphone diaphragm, and (2) when the moving air creates low pressure near the microphone and the hearing aid microphone gets sucked out and bounces back and forth. In theory, the bombardment of the diaphragm could be significantly alleviated by adding physical barriers between the wind turbulence and the microphones (e.g., windscreens or hiding the microphone behind the screened slits). The wind-induced pressure variations around the microphone, however, are much harder to reduce unless the microphone is redesigned to maintain identical barometric pressure in the front and back cavities all the time.
WIND NOISE & HEARING AIDS
Ideally, studies on the impact of wind noise on hearing aids are carried out in acoustically treated wind tunnels designed with multiple measures to reduce the noise level in the testing area. As wind can come from different angles and at different velocities in real-world windy environments, wind noise research often turns the manikin head to examine the wind noise levels when the wind comes from different directions (Fig. 1A). Some wind noise studies report the wind incidence angles and others report the head angles (Fig. 1B). This difference in angle designations is important for the interpretation of the data on wind noise levels, which are often reported using polar plots with specified wind velocities.
When wind noise is recorded with hearing aids of different styles, the wind noise patterns show distinctive patterns (Figs. 2, 3, 4). In general, behind-the-ear hearing aids have a relatively large range of angles with low wind noise levels (i.e., when the hearing aid is facing the wind or when the hearing aid is facing the leeway side of the wind; Fig. 21). In-the-ear or in-the-canal hearing aids, on the other hand, have fewer wind noise variations for a large number of angles, but the wind noise is relatively lower when the hearing aid is facing the direction of the wind (Figs. 32).
The common characteristics among the hearing aid styles are that wind noise levels are the least when the hearing aids are directly facing the wind, which may seem to be counterintuitive. The reason is that when the wind hits an obstacle (i.e., the head), the flow will divert to go around it. The point directly facing the wind is called the point of stagnation, which, in theory and in an ideal world, does not have any wind flow. Situating the hearing aid or the better ear at such a location would therefore result in the least amount of wind noise (watch video online: https://bit.ly/2R5f1lz).
Wind noise is generally loudest when the hearing aids are in the pathway of the wind flow (i.e., when the user is facing the wind and the wind comes at 0° azimuth OR when the user is facing the leeway side of the wind and the wind comes from 180° azimuth). As the wind meets the head, it has to speed up to go around the head. The wind velocity is therefore the highest on the sides of the head, and the wind noise in hearing aids is usually the highest (Figs. 2-31,2).
When the hearing aid is located at the leeway side of the wind, the hearing aid may have a similar wind noise level as when the hearing aid is directly facing the wind at low wind velocities. As the wind velocity increases, turbulence is generated on the leeway side of the head (watch video online: https://bit.ly/2R5f1lz) and the hearing aid will experience higher and higher wind noise levels. These levels, however, are generally lower than when the hearing aid is directly in the pathway of the flow (i.e., when the user is facing the wind or facing the leeway side of the wind).
Although all high-performance hearing aids have implemented some wind noise reduction strategies, user satisfaction in wind noise (57%) still lags behind general hearing aid satisfaction (78.6%).3 Wind noise reduction strategies in commercially available hearing aids can generally be divided into two categories: (1) physical modification or redesign of the form factor and (2) the wind noise reduction algorithms. Some physical modifications are old tricks that are still used in modern hearing aids, e.g., additions of hoods and microphone grids, or hiding the microphone inlets under screened sound slits. Some manufacturers elected to redesign the form factor so that the microphones can be shielded by anatomic structures, e.g., hiding the microphone in the tip of the cymba concha covered by the antihelix.4 The effectiveness of these modifications, however, is largely unknown.
WIND NOISE REDUCTION STRATEGIES
Wind noise reduction algorithms are almost a must-have in modern high-end hearing aids. Effective strategies for wind noise reduction include:
1. Reducing low-frequency gain when speech is not present5;
2. Combining the two omnidirectional microphones in addition mode to take advantage of the 3 dB difference between summing correlated sounds from the far-field and summing uncorrelated sounds from the wind6;
3. Utilizing modulation-based noise reduction algorithms when wind noise is detected because wind noise has different temporal modulation patterns than speech. Some of these algorithms can reduce wind noise levels at the hearing aid output.7 Others may not reduce wind noise, but they did not create any adverse effects;
4. Decreasing gains for low-level inputs and lowering the maximum power output or reducing the high-level gains of hearing aids for high-level inputs. This strategy is especially useful when speech is not present8;
5. Reducing temporal fluctuations by applying gain reduction to the peaks of wind noise to reduce the temporal fluctuations of the wind noise9;
6. Adopting the microphone mode with the lowest wind noise level or the highest speech-to-wind noise ratio at each frequency channel.5 Based on the data reported by Chung5,7 and Zakis,10 this strategy would be the most effective when the wind noise reduction algorithm can choose from
- the front omnidirectional mic,
- the back omnidirectional mic,
- both omnidirectional mics in addition mode (i.e., strategy 2 above), and
- the directional microphone (i.e., omnidirectional mics combined in subtraction mode);
This strategy is most effective for BTE hearing aids and the least effective for ITC hearing aids.5,7 Implemented correctly, it can enhance the signal-to-noise ratios for up to ~30 dB across different frequency regions at some wind incidence angles and wind velocities.8
For binaural hearing aids, the wind noise algorithm can detect and adapt the above listed four microphone modes with the lowest wind noise level or the highest speech-to-wind noise level in each frequency channel between the two hearing aids. The signal of choice at each frequency channel should be transmitted and adopted by both hearing aids5,8;
7. Fitting open-fit hearing aids, if possible, to people who work or enjoy outdoors because BTE hearing aids have more angles with lower wind noise levels than custom hearing aids and wind noise can escape from the ear canal through the vent11; and
8. Extracting correlated signals (sounds from the far-field) from the incoming signal (with wind noise) so that the representation of wind noise is greatly reduced in the hearing aid.12
Some of these strategies are not being implemented in hearing aids yet. Notably, some commonly used strategies have been proven ineffective or detrimental. For example, reducing the low-frequency gain and adopting the omnidirectional microphone mode is the most widely used strategy in hearing devices.6 Yet a recent study showed that CI users obtained the lowest speech recognition scores when using these strategies because:
- reducing low-frequency gain in the presence of speech would also reduce the speech energy; and
- omnidirectional microphones may have higher wind noise levels in the high-frequency regions and yielded lower speech recognition scores (Fig. 4).
Instead, wind noise reduction algorithms should use strategy 6 to take advantage of all possible microphone modes to reduce wind noise interference. The reduction of low-frequency gain should only be used when speech is not present because this strategy can reduce the contribution of wind noise to the overall signal, thus alleviating annoyance, discomfort, and upward spread of masking.
Wind noise reduction algorithms are a work in progress. Before the hearing aid manufacturers incorporate all the above strategies into their wind noise reduction algorithms, clinicians can carry out the following to reduce the interference of wind noise: (1) counsel patients using hearing devices to position themselves in the wind so that their better ear/device is facing the direction of the wind, (2) incorporate strategies 1, 3, 4, and 7 in the hearing aid fitting process (e.g., the wind noise program) if the user is not expected to communicate with others in windy environments; and (2) incorporate strategies 3 and 7 into the hearing aid fitting process if the user is likely to talk with someone in windy environments.
Acknowledgment: The author would like to thank Nicholas McKibben and Steven Taddei for their help in making wind noise recordings, and the staff at Herrick Laboratories at Purdue University for their wind tunnel assistance.
1. Chung, K., Mongeau, L., & McKibben, N. (2009). Wind noise in hearing aids with directional and omnidirectional microphones: polar characteristics of behind-the-ear hearing aids. Journal of Acoustical Society of America
2. Chung, K., McKibben, N., & Mongeau, L. (2010). Wind noise in hearing aids with directional and omnidirectional microphones: polar characteristics of custom hearing aids. Journal of Acoustical Society of America
, 127(4), 2529-2542.
3. Kochkin, S. The Hearing Journal:
2010; 63(1):19. doi: 10.1097/01.HJ.0000366912.40173.76.
4. Dodge, S. (2010). Spatial awareness and localization: ReSound live remote microphone and behind-the-ear hearing instruments. ReSound White Paper. https://www.audiologyonline.com/files/content/00000/00080/66-ao-live-remotemicdoc838v1.pdf
5. Chung, K. (2020). Improving cochlear implant performance in wind through spectral masking reduction: A multi-microphone and multichannel approach. Ear & Hearing
, 41(2), 420-432.
6. Kates, J. M. (2008). Digital Hearing Aids. Wind noise in Digital Hearing Aids. Plural Publishing, Inc., San Diego, 134-61.
7. Chung, K. (2012). Comparisons of spectral characteristics of wind noise between omnidirectional and directional microphones. Journal of Acoustical Society of America
, 131(6), 4508-4517.
8. Chung, K. (2012). Wind noise in hearing aids: I. Effects of wide dynamic range compression and modulation-based noise reduction algorithms. International Journal of Audiology
, 51(1), 16-28.
10. Zakis, J. A. (2011). Wind noise at microphones within and across hearing aids at wind speeds below and above microphone saturation. J. Acoust. Soc. Am.,
11. Chung, K. (2013). Effects of venting on wind noise levels measured at the ear drum. Ear & Hearing
, 34(4), 470-481.
12. Lee, L. (2016). Efficiency of a wind noise attenuation algorithm. Hear Rev,