SUBSCRIBE TO OUR BLOG
It’s a big problem for urban areas and it won’t be solved overnight. But there is hope in the form of ingenious mapping technologies that pinpoint noisy areas and help us mitigate problems before they arise.
Mapping noise isn’t new. In 2002, the EU’s Environmental Noise Directive required Member States not only to create maps but also to make them available to the public. The DEFRA website for England is typical of the results. It’s a site where you enter your postcode to see local sources of noise from road, rail, air and industry, both during the day and at night.
The DEFRA map, and similar ones for other countries, depict noise levels in bands of decibels, such as 65.0-69.9 and 70.0-74.9. They’re similar to weather maps that use isobars to depict areas of equal pressure.
But what you also need to know is that the decibel scale isn’t linear; it’s logarithmic. For every 3dB increase, the sound intensity doubles. On the maps, the areas with the highest decibels are often shown in red, with quieter areas in different colours.
This approach, however, has its critics. “It’s too abstract for most people as it’s difficult to envisage what the noise will really be like,” says Erlend Viggen, a researcher at the Norwegian Research Institute SINTEF. “It’s much better to generate an artificial noise that people can listen to,” he explains.
Viggen and his colleagues at SINTEF made recordings of cars travelling at 30, 50 and 80km/h and combined them to create a typical traffic sound. The software then takes into account your distance to the sound, as well as how it’s refracted and reflected by the environment.
“It’s currently at the proof-of-concept stage, not yet a finished product. But it would definitely be possible to show the user a 3D map of the listener’s position, the surrounding terrain including buildings and roads, and the paths that sound takes from vehicle to listener,” says Viggen.
The goal is to give authorities and companies information on the impact of proposed changes, such as an increase in traffic from the expansion of a nearby airport, or alterations to road layouts. It would also be handy for assessing mitigation measures, such as lower speed limits or the effectiveness of adding an earth bank to screen a busy road.
On the other side of the Atlantic, a company called HowLoud has developed an online tool anyone can use today. Unlike DEFRA’s map, which does not “represent or imply noise values at individual locations”, HowLoud promises to tell you how loud your own future house will be.
HowLoud has so far created 3D models of Los Angeles County and Orange County, and the company has successfully raised funds through Kickstarter to extend this to the whole of North America.
Americans using HowLoud don’t see maps with a decibel rating, however, but a number called the Soundscore. It incorporates local sources such as car repair workshops, restaurants and schools. These are the noises you want to know about when you buy a house, which is why the real estate business could be a huge growth area.
Crucially, Soundscore is based on more than loudness alone. As founder Brendan Farrell, a former computing and maths researcher at Caltech, told radio station KPCC, “Noise is subjective – some types of noise are more unpleasant than others.”
The fact that noise is so subjective – the same sound is far more likely to bother us at night than during the day – combined with the sheer impossibility of putting a microphone on every street corner, led to a different approach by Microsoft Research.
Yilun Wang, Yu Zheng and Tong Liu developed a noise map of New York that uses data collected by the city’s 311 service. Residents use the service to report non-emergency complaints made by phone calls, text or mobile apps. From 2010 to 2014, noise was the third biggest category.
The complaints contain the time, location and more detail on the type of noise. The different categories include banging and pounding, loud talking, parties and construction going on outside permitted hours.
To confirm that the 311 data faithfully represented what was actually happening in the real world, the researchers visited 36 locations in Manhattan at different times of day and made recordings on a mobile phone.
The 311 data did indeed correspond to the real world, except for places with few or no complaints. After all, no complaints doesn’t necessarily mean no noise. So the researchers supplemented the 311 data with information collected from location-based social networks Foursquare and Gowalla (now part of Facebook).
On Foursquare, check-ins are associated with different categories, including Art & Entertainment, Food, and Nightlife Spot. The research showed a strong correlation between vehicle noise complaints and Art & Entertainment check-ins over the course of a day, and loud music complaints were concentrated in similar geographical areas to Nightlife Spot check-ins.
Microsoft Research, HowLoud and SINTEF, among others, have shown that you can make pretty accurate city-wide noise maps without having to record sounds absolutely everywhere. And with noise pollution on the rise, the clamour for reliable maps is only likely to get louder.