Noise Floors in Shared Housing: What a Meter Shows Over Three Weeks

The quietest hour in a shared flat measured 34 decibels and the loudest measured 71, and the loudest hour was the same hour every weekday. Three weeks of readings produced a schedule that no amount of good intentions had produced.

The flat held three people, two of whom worked from home. The third worked nights and slept during the day, which is the arrangement that makes noise a scheduling problem rather than an insulation problem.

How the readings were taken

A phone meter logged the average and peak every ten minutes, in the room used as an office, for three weeks. The phone sat on the desk at head height, and the door was closed as it would be during work.

The absolute values are approximate. The pattern across the day is reliable, because the same instrument measured everything.

The weekday pattern

Between 08:00 and 09:30 the room averaged 52 decibels, driven by the bathroom and the kitchen. Between 10:00 and 12:00 it fell to 36. Between 12:00 and 13:30 it rose to 61, which is lunch. The afternoon averaged 44 until 17:00.

The noisiest hour of the working day was lunch, and it was the hour when the person had been scheduling calls.

The night shift complication

The third flatmate slept from 08:00 to 16:00, which meant the quiet hours of the flat were the exact hours the office room was unused, and the loud hours were the ones that had to be kept quiet for someone else’s benefit.

Two people were tiptoeing during their own working day and working loudly when someone was trying to sleep. Nobody had noticed the inversion because nobody had written the hours down.

What the meter changed

The three flatmates put the readings on the kitchen wall. That was the entire intervention. With a shared record, the arrangement changed within a week: calls moved to 10:00 to 12:00, and noisy kitchen activity moved after 16:00.

Nobody had to be persuaded. The disagreement had been about facts, and the facts had not existed.

The physical fixes that helped

A door seal reduced peak noise from the corridor by about 6 decibels, which is audible. A rug under the desk reduced the sound of the chair and, more usefully, reduced the echo on calls.

Both cost under 40 euros. Neither was as effective as moving a call two hours later.

What did not help

Foam panels were bought and placed on one wall. They reduced reverberation in the room, which improved recording quality slightly, and did nothing about noise arriving through the door and the window.

Absorbing sound inside a room does not stop sound entering it, which is a distinction the packaging does not make.

What remained unsolvable

Two hours a week of building work outside could not be scheduled around, because it had no pattern. Both flatmates learned to treat those hours as calls-free and used them for reading.

Accepting an unmeasurable block is a reasonable outcome of measurement, and it is easier to accept when everything else has a number.

Three weeks on

The quiet hour still measures 34 decibels and lunch still reaches 61. Calls no longer happen at lunch, and the night-shift flatmate reported being woken twice in a month rather than most days.

The reading that changed most

Peak readings were more useful than averages. A 61-decibel lunch hour and a 71-decibel drilling hour average out to something unremarkable, and both make a call impossible. The chart that mattered plotted peaks, not means.

The flatmates kept the peak chart on the wall and stopped looking at the averages after the second week.

Seventy-one decibels at the worst hour was never the problem. The problem was that it was the hour everyone had chosen for the tasks that need quiet, and a wall chart of readings fixed that faster than any purchase.