Sounding out smart cities: Auralization and soundscape monitoring for environmental sound design

Abstract

Auralization is key in developing a better understanding of how significant changes or infrastructure planning in our urban environment can have an impact on our related environmental soundscape. It allows consultants, planners and other stakeholders to hear the potential acoustic changes that might result, so that designs might be better optimized; it is also a valuable dissemination tool for informing the public as to the nature of such changes. Auralization also facilitates subjective soundscape assessment of proposed developments at the design stage and once construction is complete, smart sensor networks enable soundscape monitoring and objective evaluated on an ongoing basis. Of particular interest is sound emitted from transportation, as it is generally considered as unwanted sound and hence defined as noise. Transportation noise and road traffic noise in particular, is considered a concern for public health by the WHO, and annoyance with some aspect of our daily soundscape is not uncommon. This work presents an overview of how auralization has been used in the context of some recent transportation noise related case studies. The complete auralization chain is presented including source measurement and soundscape monitoring, sound propagation modeling using numerical simulation, soundfield rendering, and the potential for immersive multimodal presentation.

Towards Participatory Design of City Soundscapes

“…There is a rapid expansion of research and the aim of this work is to provide policy-makers and practitioners with operative tools, stand-ards and methods [3] [16] [17]. Wide range of research has been done in the field of noise abatement, noise monitoring, prediction models and auralization [18] [2] [19]. Kang et al have proposed a model, which profiles recorded soundscapes, applies linear regression to soundscape profiles to predict suitable perceptual attributes re-lated to each soundscape, and finally visualize perceptual attributes as layers in geo-graphical maps (soundscape maps) [2].…”

Fine-tuning Audio Spectrogram Transformer with Task-aware Adapters for Sound Event Detection

“…Sound Event Detection (SED) is a task of determining both the categories and timestamps of multiple overlapped events within a given audio clip. It has wide applications for real world systems including smart home devices [1] and automatic surveillance [2]. Access to large corpora with strongly-labeled sound events is expensive and difficult in engineering applications, weakly-supervised SED task has been set up by DCASE challenges 1 to evaluate the progress of SED research.…”

Selective Pseudo-labeling and Class-wise Discriminative Fusion for Sound Event Detection

“…Sound event detection (SED) is a task of detecting both the onset and offset of a sound event. It has widespread applications for real-world intelligent human-interaction systems, including smart home devices, mobile devices, smart headphones, etc [1,2]. However, there are many challenges in real SED applications, including audio signal degradation from acoustic reverberation [3], or the real recordings may contain not only overlapping predefined target events, but also nontarget events and a large number of environmental noise.…”