The rumble generated by humanity took a big dive during the Covid lockdowns.
Everything we do - from driving our cars to operating our factories - produces ground motions that can be detected by seismometers.
An international team of researchers says this noise fell by up to half when coronavirus restrictions were enforced.
The period March-May represents "the longest and most prominent global anthropogenic seismic noise reduction on record", they tell Science journal.
The group obtained their motion data from a global network of 268 seismic stations in 117 countries. Many of the stations were citizen science efforts incorporating Raspberry-Pi mini-computers.
These instruments were sensitive to all types of vibrations but also that band of frequencies, in the region of 4-14 Hertz, where human activities show up.
Their information reveals how the quieting started in January in China, the origin of the Covid crisis, and then spread like a wave to the rest of the globe.
As people were ordered home, travel restrictions were imposed, and places of work came to a halt - the usual vibrations put into the ground were abruptly dialled down.
The biggest reductions were recorded in the most densely populated areas, like Singapore and New York City, but drops were also observed in remote areas like Germany's Black Forest and Rundu in Namibia. And the phenomenon wasn't confined just to the surface; the quieting was evident even at stations placed in boreholes hundreds of metres underground.
Seismometers have long recognised a drop in this shaking at nights, at weekends and during holiday periods - but this lull was far more pronounced and prolonged.
"I think one of the most interesting things for me is that this is really our first look at what actually contributes to the human-caused field of noise," observed co-author Dr Steve Hicks from Imperial College London, UK.
"And as populations get bigger, and as cities get bigger - particularly those in geologically hazardous areas - we need to work out how we're going to monitor those hazards, such as earthquakes, volcanoes and landslides. Because as time goes on, more and more important signals that tell us about these kinds of events are going to get concealed."
Dr Hicks highlighted a shallow, Magnitude 5 earthquake that occurred in Mexico during its lockdown.
The location meant that its signal would not normally be detected in urban areas without special processing of the data.
On this occasion, however, the "filtering" wasn't necessary because the human vibrations that would otherwise smother the natural signal were reduced by 40%.
Dr Hicks says a better understanding of human seismic noise could therefore improve the detection and interpretation of delicate signals that might give warning of potentially harmful events, such as a volcanic eruption.
The team says it saw very strong correlations with the mobility trends compiled by the likes of Google and Apple. These trends are derived from the locations and movements of cell phones.
Governments used this information to gauge how well populations were following lockdown rules. Dr Hicks says if we go back into wide-scale restrictions then the seismic data offers an alternative form of monitoring that may carry fewer privacy concerns.
"I guess people worry that their mobile phone is tracking them, even though this information should be anonymised. But the very strong correlations with the mobile phone data mean the seismic noise can also be used to monitor mobility, from a more general point of view on, for example, the city scale; or in those places where the phone data simply isn't available. We weren't able to access any data like that for China, for example."
The scientific collaboration that examined the seismic data involved 76 authors from 66 institutions in 27 countries. The group was led by Dr Thomas Lecocq from the Royal Observatory of Belgium.