← Back to Blog

Alarm Sound Detection Principles: Understanding Smoke, CO and Car Alarms

Alarms: One of the Most Worthwhile Sounds to Understand

A camera can see smoke, but it cannot hear a smoke alarm ringing while nobody is home. Alarm sound detection turns a passive recorder into an active guardian — but "an alarm" is not one sound. Smoke alarms, CO alarms and car alarms each carry a distinctive acoustic pattern, and recognizing which alarm matters far more than detecting "some alarm".

The Signatures of Three Alarm Types

Alarm signature patterns
Single-tone bursts, patterned pulses, and frequency sweeps

Smoke alarm — — a strong single-frequency pulse lasting several seconds; internationally standardized cadences (T3/T4) are common

CO alarm — — a regular pulse pattern (e.g., four pulses then a pause, the T4 pattern); the rhythm itself is the identity

Car alarm — — a sweeping "wail" or long tone whose frequency glides continuously — clearly different from fixed pulse patterns

These patterns are the key evidence: detection first verifies "this is an alarm-like pattern", then classifies which alarm it is.

From Detection to Graded Notification

Alarm detection and notification flow
Pattern verification, classification, and graded notification rules

The business workflow runs entirely on-device:

1. Alarm event — pattern verification passed

2. Classification — smoke / CO / car alarm / others

3. Graded notification — smoke or CO with nobody home means highest priority; car alarms are routine reminders

4. User handling — review, contact family, or mark as false alarm

Notification channels and grading rules (push / call / SMS combinations) are configurable per product and per household.

Scenes in the Home

Home alarm sensing scene
Detectors, cameras, and push notifications working together

Home security cameras — — alarm detection plus video confirmation

Elderly care devices — — smoke and CO protection, with notifications to children and caregivers

Vehicle and outdoor scenes — — car alarms and emergency signals for fleets and parking areas

Class Composition and Detection Performance

Alarm class and detection data
Class composition and detection rates across environments (illustrative)

In home scenes, smoke alarms dominate; detection stays above 93% even against kitchen noise and street backgrounds. The main confusion sources — music, ringtones and speech — are handled through multi-class negative training.

Item
Spec
Accuracy
96.0%
False alarm rate
<2%
Model size
0.2–1 MB (INT8)
Inference latency
Configurable to platform resources
Platforms
ARM Linux / MIPS / x86_64; pre-adapted for HiSilicon SVP and Ingenic Magik NPU
Sample rate
16 kHz

Conclusion

Understanding alarms means understanding which alarm — and what it should trigger. If you are building a security, home-safety or care product, an online trial with full technical support is available.

Author: softlgh | soundSDK — Offline Sound Recognition SDK for Edge Devices | https://51tage.com

Want to try the detection yourself?