How loud should focus noise be?

What volume should steady noise be at while you work? If the research includes decibel figures, why can’t anyone just tell you which one to use?

Because those figures describe something else. They come from short lab tasks and from sound measured in the air around people, not from someone spending an afternoon on spreadsheets with headphones on. The useful answer is a method you can repeat. Find the quietest level that takes the edge off whatever is actually breaking into your work, and stop there.

Why can’t a paper choose your level?

Mehta, Zhu and Cheema ran five experiments on ambient noise and creative thinking. They compared three levels: a low 50 dB, a moderate 70 dB and a loud 85 dB. The moderate level improved performance on creative tasks, and the loud level made it worse. Their explanation was that moderate noise made processing a little harder, which pushed people toward more abstract thinking. Loud noise simply reduced how much people could take in.

That’s a finding about brief tasks and the sound around people. It doesn’t tell you where to put a slider for a track playing in your ears, so read 70 dB as a description of a lab setup, not as advice.

Two broader reviews add context, but neither gives you a number either. Zhou and colleagues reviewed 43 studies of moderate broadband noise, generally under about 85 dBA. They found only minimal evidence that it hurts cognitive performance, and the harms that did show up were tied to specific circumstances. Minimal evidence of harm is not evidence of a benefit. It only means moderate levels aren’t the obvious threat people sometimes assume.

Szalma and Hancock pooled a large body of noise-and-performance research in a meta-analysis. Speech disrupted performance more than non-speech noise did. They also found no simple rule that louder is always worse. Compared with quiet, lower-intensity noise was linked to bigger performance drops than higher-intensity noise. But when two noise levels were compared with each other, the louder one did worse.

Taken together, the kind of sound and the size of each volume change matter at least as much as hitting a target figure. That’s why the steps below rely on listening rather than measuring.

Seven steps to a level you can stand behind

1. Name the interruption. Before you touch anything, decide what you’re trying to cover: an air vent, a printer, nearby typing, a colleague on a call. Test one target at a time. If the target is talk, write that down, because speech is the harder thing to work against and volume alone may not beat it.

2. Pick a task you repeat. An inbox pass, a reading block or an hour of code all work. It has to come around often enough that Tuesday’s session can be compared with Thursday’s.

3. Start close to nothing. Choose one steady noise, white, pink or brown, from moonMind or anything else with a simple volume control. Set it low enough that you’re half unsure it’s playing.

4. Climb in small steps. Raise the volume a little, wait twenty or thirty seconds, and listen for the interruption you named. Then raise it again. Small steps matter because the jump itself is part of what you hear.

5. Stop at the first level where the interruption softens. Don’t wait until it disappears completely, and don’t add one more notch for luck. That first softening point is your candidate level.

6. Leave it there for the whole session. Wanting to turn it up after twenty minutes is common, and the urge usually comes from habit rather than a new noise.

7. Log one line when you finish: did you still notice the noise, and did the interruption still get through? Over a week, change only the volume between sessions and compare the lines. The level that keeps producing the fewest interruptions and the least awareness of the noise is your level. It applies to that task, at that desk, with those headphones.

While you’re climbing, a rough gauge helps:

• Too low: the interruption still comes through clearly. The noise is background texture, not cover.
• About right: the interruption softens, and within a minute or so you stop registering the noise at all.
• Too high: the noise has become a presence of its own and may be distracting you. Turning it down here costs you nothing.

What to skip

• Copying a figure from a paper. The 70 dB and 85 dB levels were lab conditions measured in the air. They aren’t headphone settings.
• Fighting conversation with volume. Turning up the noise against nearby talk isn’t the reliable fix it seems to be.
• Big jumps. Going from quiet to loud in one move tells you almost nothing about where the softening point is.
• Changing two things at once. If you switch noise color and volume in the same week, your log can’t tell you which change mattered.
• Reading “minimal harm” as “proven help.” The 43-study review found that moderate levels don’t look clearly harmful. That’s all it found.
• Pushing through strain. A level that leaves you tense is too high, whatever your log says.

Troubleshooting: when the level won’t stay put

You keep creeping it up. If each session ends louder than it started, you’ve most likely gotten used to the noise, and the climb is habit. Reset to your logged level at the start of each session. Before you touch the slider, ask whether the interruption is actually getting through.

Nothing covers the voices. If talk still cuts through at a level that already feels loud, stop climbing. This is the pattern Szalma and Hancock’s analysis points to: speech is harder to work against than steady sound. Moving desks, closing a door or saving word-heavy work for a quieter hour may help more than another notch.

The right level shifts from day to day. That’s expected. Desks near shared spaces change by the hour, and headphones change how much of the outside reaches you. Keep separate log lines for loud days and calm days instead of averaging them into one number that fits neither.

The noise gets more noticeable as the day goes on. Turn it down a step. You’re looking for a level you can forget about, and a sound that keeps demanding attention is doing less for you, not more.

Limits: what a volume test can’t tell you

This is a personal experiment, and it isn’t care for anything. The studies above looked at short tasks or at sound in a shared space, so none of their figures maps onto your headphone slider. Whatever level you settle on, it should be comfortable enough to sit under for an afternoon. If a level feels tiring, take that as a reason to turn it down, not proof that it isn’t loud enough yet. If listening at any level leaves your ears feeling off, or poor concentration keeps eating into your weeks, put the slider aside and ask a doctor, or a friend whose judgment you rely on, to weigh in.

Sources

1. Mehta R, Zhu R, Cheema A. Is Noise Always Bad? Exploring the Effects of Ambient Noise on Creative Cognition. Journal of Consumer Research, 2012, 39(4), 784–799. https://doi.org/10.1086/665048
2. Zhou H, Molesworth BRC, Burgess M, Hatfield J. The effect of moderate broadband noise on cognitive performance: a systematic review. Cognition, Technology & Work, 2023, 26(1), 1–36. https://doi.org/10.1007/s10111-023-00746-2
3. Szalma JL, Hancock PA. Noise effects on human performance: A meta-analytic synthesis. Psychological Bulletin, 2011, 137(4), 682–707. https://doi.org/10.1037/a0023987