Periodic Reporting for period 3 - SSID (Soundscape Indices)
Periodo di rendicontazione: 2021-03-01 al 2022-08-31
Eighty million EU citizens are suffering from excessive environmental noise and billions of euros are being spent on noise control, under the EU Directive on Environmental Noise. Unfortunately, the conventional approach, i.e. reduction of ‘sound level’, simply does not deliver the required improvements in quality of life. The growing field of ‘soundscape studies’ is addressing this gap by considering sound environment as perceived, in context, with an interdisciplinary approach. However, soundscapes are hugely complex and measuring them as a basis for environmental design requires a step change to the discipline. This research aims to achieve a ground-breaking development through the establishment of ‘soundscape indices’ (SSID), adequately reflecting levels of human comfort, the impact of which will be reminiscent of that of the Decibel scale created by Bell Systems a century ago. This will provide the underpinning science for soundscape in the field of built environment, with wider intellectual goals of moving from noise control to soundscape creation. Key objectives, as coherent steps for achieving the main aim, are: (1) To characterise soundscapes, by capturing soundscapes and establishing a comprehensive database, which will be a cornerstone for the proposed analysis, and an invaluable resource for scientists for years to come. (2) To determine key factors and their influence on soundscape quality based on the database, by conducting laboratory psychological evaluation, physical/psychoacoustic factors analysis, and more importantly, to research at a physiological/biological level, including the use of functional magnetic resonance imaging. (3) To develop, test and validate the soundscape indices, through analysing the influences by various factors, using a number of inter- & trans-disciplinary approaches. (4) To demonstrate the applicability of the soundscape indices in practice, by establishing frameworks for soundscape prediction, design, and standardisation.
For WP1, we have designed a survey protocol, and this has also been published as a journal paper to encourage other researchers to collect data in the same way. Using the protocol, we have carried out a large- scale field survey and accordingly established a database which includes nearly 5000 responses in English, Spanish, Italian, Chinese and Dutch, across 54 locations worldwide. We also recorded environmental sounds across 49 locations. For WP2, we have also carried out a series of experiments in our SSID Ambisonics lab, such as for determining Galvanic Skin Response (GSR) in response to different loudness levels, and for indoor soundscape assessment. We have also prepared sound samples for the planned behavioural and neuroimaging (fMRI) research. An ordinal logistic regression model has also been developed which predicts the perceptual attributes of urban soundscapes based solely on the spectral slopes of the acoustic time series. For WP3, a new method has been developed to introduce corrections for adjusting the Likert scales, to better predict in both linear and non-linear modelling compared to the ISO/TS 12913-3:2019 To address applicability and reliability of the soundscape perceptual attributes in non-English speaking regions, a collaboration with researchers from 24 universities and 18 countries, covering for 15 different languages, as per ISO 639-3:2007, was initiated. Through a comprehensive feature selection process, multi-level linear regression models were built to predict soundscape pleasantness and eventfulness based on (psycho)acoustic metrics derived from the database binaural recordings. These models were applied to data collected during the COVID-19 lockdown period to predict the change in perceived urban soundscapes due to the lockdown conditions. For WP4, we have initially developed concepts for how the SSID can be implemented in a smart city context and how unmanned soundscape assessments can be performed.
Based on the large-scale survey and recording database, a series of laboratory experiments results, and the analysis of correlations of different factors so far, we are working towards the establishment of the soundscape indices models, where our on-going physiological/biological level experiments including the use of functional magnetic resonance imaging, will be used to ensure the health aspects are integrated into the model.