Abstract

Depending on the acoustic scenario, people with hearing loss are challenged on a different scale than independent of normal hearing people to comprehend sound, especially speech. That happens especially during social interactions within a group, which often occurs in environments with low signal-to-noise ratios. This communication disruption can create a barrier for people to acquire and develop communication skills as a child or to interact with society as an adult. Hearing loss compensation aims to provide an opportunity to restore the auditory part of socialization.

Technology and academic efforts progressed to a better understanding of the human hearing system. Through constant efforts to present new algorithms, miniaturization, and new materials, constantly-improving hardware with high-end software is being developed with new features and solutions to broad and specific auditory challenges. The effort to deliver innovative solutions to the complex phenomena of hearing loss encompasses tests, verifications, and validation in various forms. As the newer devices achieve their purpose, the tests need to increase the sensitivity, requiring conditions that effectively assess their improvements.

Regarding realism, many levels are required in hearing research, from pure tone assessment in small soundproof booths to hundreds of loudspeakers combined with visual stimuli through projectors or head-mounted displays, light, and movement control. Hearing aids research commonly relies on loudspeaker setups to reproduce sound sources. In addition, auditory research can use well-known auralization techniques to generate sound signals. These signals can be encoded to carry more than sound pressure level information, adding spatial information about the environment where that sound event happened or was simulated.

This work reviews physical acoustics, virtualization, and auralization concepts and their uses in listening effort research. This knowledge, combined with the experiments executed during the studies, aimed to provide a hybrid auralization method to be virtualized in four-loudspeaker setups. Auralization methods are techniques used to encode spatial information into sounds. The main methods were discussed and derived, observing their spatial sound characteristics and trade-offs to be used in auditory tests with one or two participants. Two well-known auralization techniques (Ambisonics and Vector-Based Amplitude Panning) were selected and compared through a calibrated virtualization setup regarding spatial distortions in the binaural cues. The choice of techniques was based on the need for loudspeakers, although a small number of them. Furthermore, the spatial cues were examined by adding a second listener to the virtualized sound field. The outcome reinforced the literature around spatial localization and these techniques driving Ambisonics to be less spatially accurate but with greater immersion than Vector-Based Amplitude Panning.

A combination study to observe changes in listening effort due to different signal-to-noise ratios and reverberation in a virtualized setup was defined. This experiment aimed to produce the correct sound field via a virtualized setup and assess listening effort via subjective impression with a questionnaire, an objective physiological outcome from EEG, and behavioral performance on word recognition. Nine levels of degradation were imposed on speech signals over speech maskers separated in the virtualized space through Ambisonics’ first-order technique in a setup with 24 loudspeakers. A high correlation between participants’ performance and their responses on the questionnaire was observed. The results showed that the increased virtualized reverberation time negatively impacts speech intelligibility and listening effort.

A new hybrid auralization method was proposed merging the investigated techniques that presented complementary spatial sound features. The method was derived through room acoustics concepts and a specific objective parameter derived from the room impulse response called Center Time. The verification around the binaural cues was driven with three different rooms (simulated). As the validation with test subjects was not possible due to the COVID-19 pandemic situation, a psychoacoustic model was implemented to estimate the spatial accuracy of the method within a four-loudspeaker setup. Also, an investigation ran the same verification, and the model estimation was performed with the introduction of hearing aids. The results showed that it is possible to consider the hybrid method with four loudspeakers for audiological tests while considering some limitations. The setup can provide binaural cues to a maximum ambiguity angle of 30 degrees in the horizontal plane for a centered listener.

Introduction

Individuals with normal hearing often can effortlessly comprehend complex listening scenarios involving multiple sound sources, background noise, and echoes. However, those with hearing loss may find these situations particularly challenging. These environments are commonly encountered in daily life, particularly during social events. They can negatively impact the communication abilities of individuals with hearing loss. The difficulties associated with understanding complex listening scenarios can be a significant barrier for individuals with hearing loss, leading to reduced participation in social activities.

Several hearing research laboratories worldwide are developing systems to realistically simulate challenging scenarios through virtualization to better understand and help with these everyday challenges. The virtualization of sound sources is a powerful tool for auditory research capable of achieving a high level of detail, but current methods use expensive, expansive technology. In this work, a new auralization method has been developed to achieve sound spatialization with a reduction in the technological hardware requirement, making virtualization at the clinic level possible.

Key Chapters

Conclusion

Throughout the course of this study, a new auralization method called Iceberg was conceptualized and compared to well-known methods, including VBAP and first-order Ambisonics, using objective parameters. The Iceberg method is innovative in that it uses "Center Time" (TS) to find the transition point between early and late reflections in order to split the Ambisonics impulse responses and adequately distribute them. VBAP is responsible for localization cues in this proposed method, while Ambisonics contributes to the sense of immersion.

In the center position, the Iceberg method was found to be in line with the localization accuracy of other methods while also adding to the sense of immersion. Also, a second listener added to the side did not present undesired effects to the auralization. Additionally, it was found that virtualization of sound sources with Ambisonics can implicate limitations on a participant’s behavior due to its sweet spot in a listening-in-noise test. However, these limitations can be circumvented and extended to Iceberg, resulting in subjective responses that align with behavioral performance in speech intelligibility tests and increasing the localization accuracy.