Abstract:

In this study conducted in Singapore, we aimed to enhance the urban soundscape by creating a soundscape intervention device. With the complex nature of the urban soundscape, it was crucial for the device to have the ability to sense the surrounding sound, select the most appropriate natural sound masker to mask negative noise (such as traffic noise), and adjust dynamically to changes in the acoustic environment.

To achieve this, we utilized the power of cloud computing and the Internet of Things (IoT) to provide real-time listening and playback of the augmented sound via an array of ambisonics loudspeakers. The device features an innovative AI-based model to choose the optimal sound masker and determine its gain level, taking the burden offhuman supervision.

We also created the Affective Responses to Urban Soundscapes (ARAUS) dataset to serve as a benchmark for evaluating models that predict and analyze soundscape perception. The working prototype of the system was first tested in a gazebo located in a university garden near a busy road, and several more units are being installed in residential areas to study the human response to the improved soundscape. By combining technology and nature, we hope to create a more pleasant and harmonious urban soundscape for all.