Conference Proceedings

This paper discusses the acoustic design and performance of a new non-environment control room constructed at JMC Academy in Sydney, Australia. The design goals included a non-environment room at the mixing position, low reverberation times, a flat frequency response, and a half-space acoustic environment for optimal bass response. The design faced constraints related to the existing room shell and limited space for sound absorption. The research details the room's design features, including reflection-free zones, reverberation times, monitoring loudspeakers, and room structure, incorporating specialized Helmholtz absorbers. The study also presents measured acoustic performance results and observations of the listening experience in the room. The resulting room achieves high levels of performance in acoustics.

This paper explores the application of self-supervised learning (SSL) to synthetic aperture sonar (SAS) for target recognition. The study evaluates two SSL algorithms, MoCov2 and BYOL, in comparison to a supervised learning model (ResNet18) for binary image classification. The research aims to address the challenge of limited labeled data in SAS by leveraging SSL to extract features from unlabeled SAS data. The results indicate that SSL models, specifically MoCov2 and BYOL, can outperform supervised models when using a small number of labels, demonstrating the potential of SSL to reduce labeling costs and time while maintaining performance. The study highlights the potential of SSL in improving SAS target recognition, particularly in scenarios with limited labeled data.

This paper proposes an algorithm for generating distributions of sonar depth and the depth of an adversarial target in a realistic environment, based on calculating the Nash equilibrium strategies of the two parties, with the inputted probability of detection modelled by an acoustic ray tracer, Lybin. The method is demonstrated in a sample environment and shows superior performance when compared to simpler distributions. The study explores the application of this method for tactical simulations, emphasizing the importance of accurate modeling of submarine behavior for simulations. The paper concludes by discussing future work related to sampling environmental parameters for more robust strategies.