Breathing Disruption Detection

What is a breathing disruption?

A breathing disruption is a specific pattern in the audio. Audible breathing, then a long silence, then a recovery sound. That pattern can mean you stopped breathing for a bit during sleep, so the app gives it its own event type instead of lumping it in with snoring.

Each disruption is logged with its timestamp, how long the silence lasted, and how strong the recovery breath was. The app also grades each one by the length of the silence. Under 12 seconds is mild, 12 to 20 seconds is moderate, and over 20 seconds is severe. Your nightly summary calls out the single longest pause of the night with its own rating, so one long event still stands out on a night with few disruptions.

The three detection conditions

The app only flags a disruption when three things happen in order:

1. Baseline breathing. The app has to hear breathing for at least a minute before the silence. It checks twelve 5-second windows and needs breathing in at least 11 of them. This proves your breathing was loud enough to hear in the first place, so the silence that follows means something. Without that check, a quiet stretch from someone whose breathing never registered would look exactly like a real pause.
2. Extended silence. Complete silence for 10 seconds or more. Anything shorter never qualifies, which rules out the normal pauses between breaths.
3. Recovery sound. The silence has to end with a recovery sound, a gasp, a snort, a sudden loud breath, that measures at least 15 dB louder than the average during the silence. That confirms breathing came back with some force, as opposed to you having rolled away from the phone.

The strictness is deliberate. Each condition rules out a different false alarm. The baseline rules out people whose breathing can't be heard, the 10-second floor rules out ordinary gaps, and the recovery requirement rules out silence from rolling away from the phone. The trade-off only goes one way. When the app gets it wrong, it's by missing events, never by inventing them. The section on quiet breathers covers what that means for your counts, and How Detection Works explains the decibel measurements behind the third condition.

Markers on your timeline

Breathing disruptions show up as muted gray-blue markers on your timeline. The color keeps them visible without fighting the snoring episodes for attention, so you can scan a whole night and spot clusters. Disruptions piled up in one stretch of the night mean something different from the same count spread out evenly.

You can also tap a breathing disruption in your nightly summary to jump to that moment on the timeline and listen to it. Hearing the silence end in a gasp lands differently than reading a number, and the audio is what makes this worth bringing to a doctor.

Zoomed in

The metrics on your summary

Disruptions per hour is the main metric. The app divides the disruptions it found by your hours of sleep and rates the result. Fewer than 1 per hour is Minimal, 1 to 3 is Mild, 3 to 5 is Notable, and more than 5 is Significant. Each level comes with a short summary and a color. Since the app can only count what it hears, even a low number can be worth paying attention to, and a higher rate is worth raising with a doctor. The rate isn't a medical score.

Recovery breath strength shows as a decibel value like +16 dB, meaning the recovery sound was 16 decibels louder than the silence right before it. A bigger number means a more forceful recovery, which can suggest the body worked harder to get air moving again. Your summary shows the strongest recovery of the night.

The nightly assessment needs at least 2 hours of sleep. If your recording was shorter than that, or the app didn't hear enough breathing to work from, the assessment stays hidden for that night. A full night with the phone close by gives it what it needs.

Your summary can also show:

• Average breathing rate in breaths per minute, with the normal adult range of 12 to 20 shown for reference
• Average snoring intensity in decibels
• Longest pause of the night with its severity rating
• Gasps as a total count and per-hour rate
• Blood oxygen, if you've paired an Apple Watch that records it: how much of the night you spent below 95% saturation, and your lowest reading

Breathing disruptions also feed into your sleep score. Snoring and disruptions are among the factors that can pull it down, along with sleep duration, deep and REM sleep, efficiency, and awakenings. The breathing section is the detailed view of disruption activity, and the score is the overall picture. Disruption counts show up in your Weekly Summary and the factor comparison cards too. The "How does this work?" button in the app's breathing section opens a full explanation of detection and its limits.

Gasps vs. breathing disruptions

The app counts gasps separately from breathing disruptions, and the difference matters. A gasp is one sound, a sudden forceful breath in, which may follow a pause and can be a sign of the body struggling to get air. A breathing disruption is the whole three-part pattern of baseline breathing, a silent gap of 10 seconds or more, and a recovery sound. A gasp can happen without that pattern around it, which is why the two counts drift apart. Gasps are reported as a total and a per-hour rate. Disruptions are reported per hour. Both appear as their own items on your nightly summary, and Episodes & Events covers gasps with the other event types.

CPAP nights and quiet breathers

Detection works by listening, so it depends on hearing both your breathing and the recovery sound. Quiet breathing, a soft CPAP machine, or a phone that's too far away can all make it less reliable, and quiet disruptions can go unrecorded. Remember the first condition. The app needs a full minute of steady audible breathing before it starts watching for a pause. If your breathing stays barely audible all night, the baseline never forms and the app undercounts or misses events.

Expect the count to run low. Listening alone misses:

• Quiet disruptions the microphone never picks up
• Events with no breathing effort, since there's no sound to hear
• Events that end without an audible recovery sound
• Anything buried under background noise like a fan, a white noise machine, or HVAC

Treat the count as a conservative estimate, and use it to track trends night to night. A rate that rises or falls across weeks tells you more than any single night's total.

Medical disclaimer

The results make a good conversation starter. The Doctor-Ready Data guide shows how to turn your recordings into something a doctor can use, and Export & Sharing covers the report and audio files you can bring to an appointment. Everything stays on your device until you decide to share it. See the Privacy Policy.