Sound for deep work

The useful framing is not "what audio makes me focus" but "what is currently stopping me". The answer changes what you should play.

Diagnose first

A loud environment. Your problem is masking. Reach for a session with a heavier noise bed and a shallow, slow modulation — nothing that asks for attention. The color matters more than the beat rate here.

A wandering mind in a quiet room. Masking is irrelevant. What helps is a stable, low-information field and a session with a defined end.

Anxiety dressed up as distraction. This is common and usually misdiagnosed. If you cannot settle because you feel wired, faster audio makes it worse. Go to alpha, or something explicitly calming, and come back to work afterwards.

Genuine tiredness. No audio fixes this. A 20-minute session will not replace sleep, and using stimulating audio to push through is how you get a bad afternoon and a worse evening.

Pick a band deliberately

Counter to the intuition that focus means fast:

• SMR, 12–15 Hz — the default for sustained work. Calm alertness. Sensorimotor Calm Focus, 14 Hz, 19 minutes.
• Beta, 15–20 Hz — for work needing momentum: drafting, arguing, building. Hyper-Focus Drive, 18 Hz, 31 minutes.
• High beta, 20 Hz+ — short bursts of dense analysis only. Deep Analytical Processing is 22 Hz and deliberately only 16 minutes.
• Alpha, 8–12 Hz — for reading, editing, admin. Work with rhythm rather than hard reasoning.

More on the distinctions in beta and SMR.

Match the session to the block, not the day

This is the recommendation with the most practical value, and it has nothing to do with frequency.

A session has a length. Use it as the work block. Start the session, work until it ends, stop. The end of the audio is a natural boundary that costs no willpower to observe, which is most of what makes timed work methods effective in the first place.

Running focus audio continuously for six hours does the opposite. It stops being a signal and becomes the room.

Practical settings

• One session per kind of work. Reuse deliberately. The consistency becomes a cue, and cues do more for starting than any frequency does.
• Low volume. Under your thinking, not over it.
• Lyrics are the real enemy. Nothing here has any, which is most of why instrumental beats vocal music for verbal work — no frequency claim required.
• Try the carrier drift on a long session. A Shepard-Risset carrier sweeps the timbre continuously so it never settles into something you can predict. See carrier drift.

What the evidence supports

Modestly. A 2019 meta-analysis found a moderate pooled effect across memory, attention, anxiety and pain, varying with frequency, duration and timing. A 2025 parametric study of 80 participants found gamma-frequency beats on a low carrier with noise masking best for attention — and called the effects modest.

Modest is a fair word for it. The reliable part of this is the behavior: sitting down, starting a timer, and not doing anything else until it ends.