Reviews
Part of Inside desk noise testing: what happens at each stage
Is measuring your desk noise worth the trouble?
Measurement helps in a few specific situations and wastes an evening in all the rest. How to tell which one you are in before you start recording anything.
Measurement is a tool, not a virtue. It earns its place when it changes a decision. This is how to work out in advance whether yours will.
What to take away
- Measure when there is a dispute, a repeated change to evaluate, or a pattern nobody can describe.
- Do not measure to prove a complaint wrong. It rarely works and it costs goodwill.
- Office desks sit far below the levels that occupational limits address, so those numbers will not help you.
- The cheapest useful test is a comparison between two states of your own room.
When it is worth it
- Someone disputes what is happening, and a dated record would settle it.
- You are about to spend real money and want a before reading.
- The complaint is about certain times of day, and nobody can say which.
- You are testing several changes and need to tell them apart.
When it is not
- Everyone already agrees the noise exists. Skip to changing something.
- You want a number to argue with. A reading below a limit does not make a person wrong.
- The change you are considering costs nothing. Just make it.
- You cannot control the conditions, so no two readings would be comparable.
Why occupational numbers will not settle a desk argument
Hearing-damage thresholds are much higher than anything a desk produces. NIOSH sets a recommended exposure limit of 85 A-weighted decibels averaged over an eight-hour workday and explains that for each 3 dBA increase above that, the allowable daily duration should be halved. Its practical field test is that sound at about 85 dBA usually forces you to raise your voice to be heard by someone three feet away, and 95 dBA means shouting. An office with a mechanical keyboard is nowhere near that. Distraction and annoyance are real problems, and they are not the problems those numbers were written for.
What a good result looks like
A useful measurement produces a sentence, not a number: typing at the doorway measured about the same before and after, or the level dropped by about four decibels and the listener says the sound is duller. Both are results. A single number with no comparison is not.
Accuracy you can actually reach at home
You do not need laboratory precision, and the research on getting closer to it is clear about where the limit sits. An experiment testing whether iOS phones and similar devices could serve as reliable instruments for noise exposure compared three different types of microphone on the same devices against reference equipment, in the context of the very large number of workers with potentially hazardous noise exposure. The lesson for a desk is proportion: your phone plus a fixed position plus repetition gets you far enough for a household decision, and the last few decibels of accuracy cost more than the decision is worth.
Deciding in two minutes
Write the decision you are about to make. If a number would change it, measure. If it would not, do the cheap change instead and ask the listener. The instrument questions, if you decide to go ahead, are in questions to ask before buying a sound meter, and the full method is in what happens at each stage.
Common questions
Can I use a measurement to push back on a complaint? Rarely well. Annoyance does not track level closely, and a low reading tends to read as dismissal rather than evidence.
Is it worth measuring my keyboard against a colleague's? Only if you can put both on the same desk in the same room with the same microphone position. Otherwise you are measuring the two desks.
How often should I re-measure? When something changes, and once a year otherwise if you are keeping a record at all.
What if the numbers contradict what people hear? Believe the people. Level is one part of annoyance, and frequency, timing and predictability are the rest.
