Eric R. Bernstein - eric.r.bernstein@gmail.com
Anthony J. Brammer – brammer@uchc.edu
University of Connecticut Health Center
263 Farmington Ave.
Farmington, Ct 06030-2017
Popular version of paper 3pNS6
Presented Wednesday afternoon, November 2, 2011
162nd ASA Meeting, San Diego, Calif.
In the 1980s, the National Institute of Occupational Safety and Health found that over 9 million U.S. workers were exposed to noise levels that exceeded the 85 dBA limit known to cause hearing damage with prolonged exposure. These environments typically require workers to use hearing protection devices that may have an integrated electronic communication channel, such as aviation headsets that allow pilots to hearing instructions from air traffic control towers. An obvious solution to improve speech intelligibility in high noise environments with these types of hearing protectors is to simply increase the volume of the communication system, thus improving the ratio of signal power to noise power, or signal-to-noise ratio (SNR). Unfortunately, a standard volume control increases the communication channel gain across the entire frequency range, and depending on the background noise of the environment, can increase the sound pressures underneath the ear cup to levels associated with hearing loss. In many industrial environments, noise power is rarely equally distributed across the entire frequency range and so excess signal power can be added beyond what is necessary to maintain speech intelligibility.
To address this problem, a communication hearing protector has been developed that adaptively alters the characteristics of the communication channel based on the environmental noise. The system uses a modified active noise reduction algorithm that breaks the frequency range into a series of bands, called subbands, in order to better identify the power spectrum of the environmental noise underneath the ear cup. A similar filter breaks down the original communication signal to determine its power spectrum. This information is then used to select an appropriate gain for each frequency band to either reach an optimal SNR or limit the total power of the combined noise and speech signals. These gains are used to filter the original speech to provide a new communication signal that is presented to the user that improves speech intelligibility without providing excessive sound pressures that can lead to hearing loss.
Sounds pressure waveforms from a simulation of several different hearing protector designs are given in Figure 1 and can be heard in the accompanying audio files. The first thirty seconds of each simulation exposes the user to a Leopard tank noise recorded from the drivers position without any speech communication. A passive hearing protector does not provide any further reduction in sound pressures detected under the ear cup over time and thus the amplitude of the signal remains relatively constant during the noise only period. The traditional active noise reduction (ANR) and subband ANR hearing protectors both slowly reduce the sound levels at the ear over the first fifteen seconds until a quieter steady-state level is reached. The key difference between the two ANR systems is demonstrated when a speech signal is added during the last thirty seconds of the simulation. The volume of the communication signal is set so that the total sound pressure underneath the ear cup does not exceed a level that could cause hearing damage over time using the passive system. Due to the high level of environmental noise the resulting SNR is too low and the noise masks the speech signal. A traditional ANR hearing protector headset provides a user controlled volume level and thus user interaction is necessary to increase the volume levels after the environmental noise is reduced by the active control to regain speech intelligibility. The subband system automatically recognizes that additional signal power is required for speech intelligibility and that active control has reduced the environmental noise levels such that additional communication power can be added without exceeding hearing damage thresholds. Over fifteen seconds, the speech signal gain is increased until a comfortable listening level is established and the sound of the speech can be easily distinguished from the remaining environmental noise. The subband system has improved the word intelligibility to nearly 95% which is suitable for most critical communication, from about 50% for the passive and traditional ANR systems.
The technology developed through this research will help provide smart hearing protectors that can adapt to the environmental noise to optimize speech volume and improve intelligibility without risking hearing damage. It is expected that these devices can help mitigate some of the communication problems of modern hearing protectors and encourage workers to better limit occupational noise exposures.
| sample 1 | sample2 | sample3 |