“Acoustic Scene Analysis” and Digital Hearing Aids

Digital hearing aids have gained rapidly in popularity in recent years, now comprising over 93% of the US market (figure 1). Interestingly, despite this increase, customer satisfaction, benefit, use, and residual disability have not changed appreciably during the past decade (figure 2). One hearing aid feature, however, that has produced considerable increases in patient satisfaction is the directional microphone (figure 3). Directional microphones, of course, do not require the use of digital technology, as they have been available commercially in hearing aids for over thirty years. That said, a recent trend has been to incorporate directional microphones into digital hearing aid systems that monitor the listening environment and automatically activate fixedor adaptive-directional microphone arrays only when appropriate. These systems represent the first approximation of “acoustic scene analysis”, which involves a classification and decision-making process that can recognize a wide variety of sound environments and adapt the hearing aid characteristics accordingly. Numerous studies have been published that suggest that automatic activation of hearing aid features may provide increased use by patients than when they are manually activated (Surr, Walden, Cord and Olson 2002; Cord, Surr, Walden and Dyrlund 2004). In addition to directional microphones, acoustic scene analysis may be used to adapt features including feedback reduction, noise/reverberation cancellation, and multiple-channel compression. The classification accuracy of this system has proven to be highly sensitive and specific in laboratory settings under simulated conditions (Buchler 2002). As these systems move from laboratory to clinic, however, it raises the issue of how to best evaluate classifier performance for “real-world” conditions. Further complicating this issue is the prevalence of “datalogging” features with modern digital hearing aids, which may be used to further optimize hearing aid performance to meet individual patient needs, moving from “acoustic scene analysis” to “auditory scene analysis”. The appropriate balance between controlled laboratory