IDOL Speech Server can detect and label segments of audio that contain security-related sounds, including alarms, breaking glass, screams, or gunshots. This functionality is available as the audiosec module, which is implemented in the AudioSecurity standard task.

The audiosec module can also identify the type of alarm that it has detected. To identify an alarm, it must match an alarm template that exists in the database. You must use an audio sample of the alarm to create the template. If a corresponding template does not exist, the task produces a result with the label <UNKNOWN-ALARM>.


To detect breaking glass, HPE strongly recommends that the audio sampling rate is 16 kHz.