How Sound Detection Works

The sounds the app classifies

Rather than treating every noise as a snore, the app sorts what it hears into six main categories:

• Snoring, the main event. Each detection lands on the timeline with its volume level.
• Gasps, sudden intakes of breath that often follow a breathing pause.
• Coughs, tracked as their own category rather than folded into snoring.
• Sleep talking: speech, whispers, and mumbling.
• Laughter, because sleepers do laugh.
• Loud sounds, sudden noises above a threshold you set (covered below).

Snoring, gasps, and coughs form the respiratory group. The app also monitors breathing patterns throughout the night; a silent stretch of about 10 or more seconds followed by a recovery sound that is clearly louder than the silence before it gets flagged as a breathing disruption. Breathing Disruptions covers those in depth.

How does it tell a snore from a cough? Each sound type has a recognizable acoustic signature. Snoring carries most of its energy in the low and mid frequency range, roughly 50 Hz to 3 kHz, which separates it from speech, coughs, and ambient noise. The classifier weighs both the sound pattern and the frequency characteristics of each audio segment before assigning a label. Most ambient noise gets ignored.

No detection system is perfect. When two sounds overlap or bedding muffles one, an event can land in the wrong category. If a label looks off, play back the audio for that moment and hear what happened. Everything here is audio analysis for personal insight; the app does not diagnose sleep apnea or any other condition.

Real-time analysis, no sampling

Some apps save battery by sampling: they wake up periodically, listen for a moment, and sleep again, which means they can miss whatever happens in between. Snore Timeline analyzes your audio continuously, using Apple's Sound Analysis framework running on your phone. Every sound is processed as it happens. Nothing is skipped, sampled, or uploaded.

Continuous analysis has two consequences you'll notice:

• Detection starts immediately. There is no calibration or setup step; the first snore of the night counts as much as the hundredth.
• The timeline is complete. Because the app never stops listening, a quiet gap on your timeline means the room was quiet, and the gaps themselves become data for breathing and sleep stage analysis.

All of this runs on-device. Your audio never leaves your phone, and the Privacy Policy covers what that means in practice.

Sensitivity and false positives

Not every sound makes the cut. The app logs a sound as an event only when it's a strong enough match for one of its categories, which is how it keeps a creaking radiator from filling your night with phantom snores.

You control how strict that bar is with the sensitivity setting, which has five tiers: Minimal, Low, Balanced, High, and Maximum. Balanced is the default.

• Minimal captures only clear, loud snoring and ignores almost everything else. Choose it for a noisy room or when a partner's snoring keeps tripping the detector.
• Maximum catches very faint snores but also picks up more background noise and logs more false detections.

Decibels and frequency colors

Loudness appears throughout the app in dB SPL, on a scale from about 28 dB, the reference for near-silence, up to 105 dB for an extremely loud sound. Read it like a volume meter: higher numbers mean louder sounds. As a rough guide to your snoring:

• Below about 40 dB: faint snoring
• 40 to 48 dB: light snoring
• 48 to 56 dB: audible snoring
• 56 dB and above: heavy snoring

The app tracks both peak and average decibel levels for each episode. These numbers are for personal reference, not a clinical measurement.

The app detects which microphone is in use, and it adjusts readings while your phone plays audio, such as music or a podcast.

Every sound is a blend of frequencies, and a frequency is simply how fast the air vibrates, measured in hertz (Hz). Low sounds vibrate slowly; high sounds vibrate fast. Press play and drag to hear it and watch the wave tighten:

160 Hz

Play Drag the slider — higher pitch vibrates faster

Zoom all the way in on the waveform and the bars split into stacked shades of orange showing where each sound's energy sits across frequencies:

• Darker orange: low frequencies, roughly 50 to 250 Hz, the deep rumble of snoring.
• Medium orange: mid frequencies, about 250 to 1500 Hz, harmonics and vowel sounds.
• Bright orange: high frequencies, about 1500 to 8000 Hz, sibilants like the "s" in "sss".

Breathing Snoring

Play 0:00 High Mid Low

The bright, high-frequency band is the one that matters most for breathing. Every exhale makes a faint hiss, like a soft “sss”, and that hiss sits in the high band. Snore Timeline listens for it to follow your breathing through the night, which is what powers breathing-disruption detection and sleep stage estimates. It is also why a phone placed too far away or a noisy room weakens those features: the hiss is quiet, and it is the first thing lost.

The frequency detail appears only on the most zoomed-in view; at wider zoom levels the bars show as solid color. Timeline & Playback walks through reading the waveform as a whole.

Zoomed in

Loud Sound Detection

The classifier handles snoring and sleep talking on its own. Loud Sound Detection exists for everything else: it creates an episode whenever a sound rises above a volume threshold you choose, regardless of what the sound is. That catches noises the classifier can't name, such as whispered sleep talking too quiet to register as speech, teeth grinding, sounds you make while moving, or other unidentified nighttime noises.

The default threshold is 55 dB. To pick a threshold suited to your room:

1. Start a recording and stay quiet.
2. Watch the dB level the timeline shows for your room's baseline noise.
3. Set the threshold just above that level.

A quieter room lets you use a lower threshold and catch more.

You may also see sounds land in the Loud Sound category that you expected to be snores. That happens when background noise masks the breathing patterns the snoring classifier listens for; the classifier needs a clear signal to identify snoring, and when a room sits above about 45 dB at baseline, more sounds tend to register as Loud Sound signals rather than snoring episodes. The next section explains what to do about it.

How background noise affects detection

Constant background noise such as an air conditioner, fan, traffic, music, rain, or ocean sounds is tracked separately and does not create snore events on your timeline. The app recognizes these as continuous ambient noise rather than discrete snoring, so a humming AC by itself will not fill your night with false snores.

The real cost of steady noise is masking. A loud noise floor drowns out quiet breathing and faint snores, leaving the classifier less signal to work with. Two things follow:

• Faint sounds become harder to detect at all.
• In rooms above roughly 45 dB at baseline, sounds that would otherwise classify as snoring shift into the Loud Sound category.

To get cleaner classification, quiet the room where you can. The usual suspects are fans and white noise machines, HVAC and air purifiers, open windows facing traffic, and TVs or audio left playing. A quieter room gives more accurate detection overall.

When you can't control the noise, in a hotel room for example, work with what you have:

• Move the phone closer to your head so breathing sounds stay audible over the hum. Getting Started covers placement in detail.
• Raise the sensitivity so the app still picks up softer sounds over the background.

Zoomed in