Big cities are loud, New York is louder. At least that’s my personal impression. Finding an apartment where sirens, garbage trucks or subways don’t wake you up every night is almost like winning the lottery. Claudio Silva, engineer and project manager at New York University, confirms this. Noise is a major disruptive factor in city life. But so far, statements about the noise level in the city are based on anecdotes, like the personal experiences of Claudio Silva and me.
So far, the most important source for noise researchers has been a database in which complaint calls to a municipal hotline are stored. This data is spotty and unreliable. NYU’s Brooklyn campus is relatively quiet. High-rise office buildings stand around an open space. A few students are playing boules. A sign hangs from a first-story window that reads, “Recording in progress! Researchers are investigating street noise.” Next to it, a microphone protrudes in the direction of the campus. You can only really see it if you know it.
How New York sounds
“In some places we record everything. We use these recordings to train machine learning algorithms so that they can detect the source of noise,” says Yitzchak Lockerman.
This would allow them to continuously measure how noise pollution is developing, instead of relying on one-off measurements. The minicomputers should evaluate the noises themselves and only save the result or forward it to a server. Instead of an audio recording, only the information “fire engine” is transmitted, or “sledgehammer” or “street musician”. This sound information is then given to another member of the research team, Harish Doraiswamy. His job is to combine the data in a 3D model of the city with other datasets.
“We have data on crime, car traffic and subways. We also know where building permits have been issued. And with all this data we can then see how different aspects influence noise pollution. In which neighborhoods it is particularly bad and how in different places that will be dealt with.”
Collect more data
This would give the researchers a reliable source of data. But the data is still spotty. Covering the whole city with microphones is impossible. That’s why Claudio Silva and his colleagues also want to use data sources that make the 3D model more precise. Data on the condition of individual buildings. For example, which building materials were used.
“If we know the geometry of buildings and streets and the building materials they are made of, it is theoretically possible to simulate the physical conditions,” says Claudio Silva.
Based on the selective data of the noise recording, an always up-to-date noise map of the entire city could be created.