Automatic Augmentation of Soundscapes

Opportunity

The World Health Organisation (WHO) has identified noise as a significant environmental pollutant, second only to air pollution. Noise exposure has been compared to second‐hand smoke, with increasing evidence linking it to health issues such as ischaemic heart disease incidence, annoyance and sleep disturbance. Current sound masking interventions involve static playback of sounds through loudspeakers, which require the manual selection of sound levels and tracks. This invention introduces a novel AI model that automatically selects the most appropriate soundtrack and sound level to optimise perceptual responses. This AI model is trained on a comprehensive dataset of human subjective responses collected by NTU’s research team, making it highly effective in enhancing the acoustic environment. 

Technology

This system configures an audio track (“masker”) to be introduced to an existing soundscape, either through loudspeakers or digital addition via headphones, based on optimising predicted perceptual responses. The term “masker configurations” refers to any masker that can be added to the acoustic environment via any form of playback, alongside any necessary parameters or effects for intended playback. These parameters include digital gain levels, spatialisation, dynamic envelope and filtering. The masker configuration strategy is autonomous and adapts to the dynamics of the present soundscape through monitoring or digital input. This dynamic adaptability ensures the soundscape is continuously optimised to improved perceived qualities such as pleasantness, annoyance, calmness, vibrancy and eventfulness.  

Applications & Advantages

Application areas include:

• Outdoor urban areas with poor sound quality, such as parks next to highways, to mitigate the adverse impact of surrounding noise.
  
• Indoor areas to improve acoustic comfort and overall quality.
  
• Indoor or outdoor environments where the acoustic environment needs to be maintained in a fixed or desired condition.
  
• Crafting soundscapes in virtual, mixed or augmented reality based on listener perception. 
  

Advantages: 

• Autonomous and adaptive soundscape augmentation. 

• Real-time optimisation based on AI model predictions. 

• Improved acoustic environments leading to enhanced comfort and reduced health risks. 

• Flexibility in application across various settings, both indoor and outdoor.