Alert sounds should not place critical information solely in the high frequencies
Aliases: alert sound design · notification audio · pitch selection
What it is
A common trap when designing alert or notification sounds is putting the main pitch variation, or the bulk of the acoustic energy, in a high frequency band — high frequencies sound crisper and more attention-grabbing, and that makes them tempting as a way to make an alert stand out. But the upper bound of the audible range narrows preferentially from the high-frequency end with age, which means that if an alert sound's only distinguishing feature sits in a band that is especially vulnerable to age-related loss, older users or those with occupational high-frequency hearing loss may fail to perceive the critical information the sound was supposed to convey at all — not just find it a bit quiet.
Why it happens
This isn't about the alert being completely inaudible — as long as its overall loudness is high enough, the presence of a sound is usually still noticed. The problem is that the specific information the alert is meant to convey (which kind of notification, which status it represents), if it is distinguished only by a pitch feature sitting in a high-frequency band that happens to be the one most affected by age or noise exposure, leaves the user aware that "a sound occurred" without being able to tell what it was saying — or simply missing the band that was supposed to draw their attention. The variation in where different users' audible range tops out directly determines whether they can reliably receive the information a designer placed in that high band.
Where it stops holding
This advice targets the specific case where critical information is distinguished by a single high-frequency pitch feature alone. If the alert already comes paired with a synchronized visual cue or haptic feedback, users have another channel to pick up the same information even if the high-frequency part is missed, and the risk from high-frequency encoding drops considerably. Conversely, moving critical information to the mid frequencies is not an automatic guarantee of safety either — low-frequency ambient noise (traffic, HVAC, machinery) can mask a mid-frequency alert too. Shifting to the mid band only addresses the specific risk of age-related loss, not every risk that can make a sound hard to hear; the background noise profile of the actual use context still needs to be checked on its own.
Applying it
- Put the pitch variation that carries critical distinguishing information (different notification types, different urgency levels) in a mid-frequency band — roughly a few hundred hertz up to three or four kilohertz — where most people's hearing is least affected by age, and which also happens to be where speech information itself concentrates, making it reliably perceivable.
- High-frequency content can be used as decorative polish (to make a sound feel crisper), but should not be designed as the sole cue a user must rely on to judge what kind of alert it is.
- To verify: play the designed alert sounds to a test group spanning different age ranges, including older users, and ask each to describe the pitch features they hear and what information they infer from them. If older and younger users' descriptions diverge noticeably for the same sound — a pitch change younger users can distinguish that older users miss entirely — the critical information sits too high in frequency and needs to be shifted toward the mid range.