Implementing and validating headphone-based virtual acoustic methods to study auditory distraction

Master Thesis of Guo, Shuyao
Forschungsgebiet: Acoustic Virtual Reality
Betreuer: Yadav, Manuj / Fels, Janina

Implementing headphone-based virtual acoustic methods to study auditory distraction 1. Introduction Background noise, including noise irrelevant to the task, brings impairment in cognitive performance, for both adults and children; the so-called irrelevant sound effect. According to current state of research, mainly two different kinds of mechanisms lead to auditory distraction, namely sound-related attention distraction and specific interference. However, there are certain controversies on ISE, given the facts that the similarity of background noise and task-related language proved to be irrelevant, and ISE can also be evoked by non-linguistic sound. Besides, the relationship between attention distraction and specific interference worth a further study. Series of experiments have been done to clarify these mechanisms, in which participants were asked to finish certain cognitive tasks in an environment with or without background noise. In earlier studies of this topic, the background noise used has mostly comprised of unrealistic sounds that were presented diotically, which differs a lot from real scenarios. Stimuli closer to real background noise should be used in such experiments for a more reasonable conclusion. Besides, the effect of room acoustics and sound reproduction methods on cognitive performances is also not clear. Different types of noise stimuli are thus necessary for this purpose, including stimuli closer to reality than simple diotic or monaural sounds used in early studies. According to current research, acoustic technologies can establish a realistic virtual acoustic environment via headphones and HRTFs. Head and eye-tracking equipment and technique are introduced to bring interaction between participants and environment. In actual experiments, these implementations can be reproduced via headphone while participants perform certain cognitive tasks. This is currently being explored in a joint DFG project of acoustics (at IHTA) and psychology researchers at RWTH and TU Kaiserslautern. In this master thesis, which will benefit from the expertise within the DFG project, different classes of background noises based on headphones will be implemented in order of increasing realness, using room acoustics, generic and individual HRTFs, head and eye-tracking technology. Finally, at least one short case study per implementation will be accomplished to test the validity of these implementations. 2. Methods 2.1 Implementations 2 different room models, namely classroom for children and open office for adults have been designed using Sketchup. Room acoustic simulation and auralization are done with MATLAB, RAVEN, VA and certain sound engines. Parameters of room acoustics and psychoacoustics are also considered in the simulation, based on existing literature about these parameters in classrooms and offices. Starting from here, new elements are to be introduced to the workflow. a. Generic HRTF A generic artificial HRTF for adults and children is used to produce better acoustic environments compared to the diotic stimuli used in previous studies, which brings spatial distribution of sound sources into consideration. With binaural technology, this stimulus can be presented via headphones during experiments. b. Individualized HRTF Next stage of noise stimuli is to take individualized HRTFs into account instead of using the same generic HRTF for all participants. Based on data of head dimensions, HRTFs are slightly modified with specific algorithm to adapt different participants. Thus, more precise spatial information will be provided for the experiments. c. Headtracking All the simulations mentioned above are static, meaning no interaction between participants and environment. To introduce certain interaction to the scenery, head and eye-tracking equipment will be used to collect the spatial orientation of participants in time, such as position and angle of the head. With real-time processing, the virtual acoustic environments used in the experiments will be adapted to react to the participant’s movement, making it one step closer to reality. d. Individual HRTF Finally, individual HRTFs are considered in this workflow. Measurement of HRTFs will be executed in an anechoic chamber, taking approximately 10 minutes. Individual HRTFs will be combined with all the implementations planned before. Thus, the virtual environment is highly personalized, to provide a background noise as authentic as possible. 2.2 Validation Section 2.1 lists the types of simulation that will be implemented, and the corresponding background noise stimuli that will be created, which fit well to the need of this project. To verify the validity of these stimuli, trial tests will be run using all type of background noises and cognitive tasks designed by psychologists. Given the logistical limitations in terms of running experiments with human participants at present, at least one individual HRTF will be completely measured and used in implementation of the virtual acoustic environment closest to reality. 2.3 Scheduling A total period of 26 weeks is planned for this dissertation. First 2 weeks will be used for familiarizing with the current state of this project and tools required. Next, 4 weeks of time are occupied for implementing generic and individualized HRTFs. Implementations of headtracking will be the most difficult part of the project, which should take 6 weeks. To finalize the whole virtual environment will use another 6 weeks. Finally, 4 weeks for validation and experiments and last 4 weeks for writing and revising the paper.