AUDIOSENSE : Advanced Audio Analytics
AUDIOSENSE technology is a set of solutions from the creation of sound signatures to the audio signal analysis in order to detect and classify preselected sound events .
With the concept of pre-defined sound signatures (ACE), our systems become intelligent and triggers alerts only on events that are detected.
AUDIOSENSE technology supports all existing audio digital formats digital (MP3, WAV ...) which may come from any source : TCP connection, RTSP streaming, storage unit, direct sound card.
AUDIOSENSE technology is based on a powerful calculation engine: ORELIA ACE Engine.
ORELIA ACE Engine is a configurable software that allows the implementation of ACE in specific applications with minimal needs to develop specific code.
AUDIOSENSE technology comes in several customer solutions :
► AUDIOSENSE Server : software solution for centralised detection and classification of sounds through TCP or RTP / RTSP audio streams. See an example :
► AUDIOSENSE Package : linux application for embedded signal processing and classification of sounds. Supports low-energy x86 (such as AMD Geode™ LX 800 and Intel® Atom™) and ARM architectures. Compliant with ARMv6 (such as ARM11 from Raspberry Pi) and ARMv7 (such as CORTEX A8 and A9 from smartphones and Beagle Bone) families. Highly optimised using vfp and SIMD instruction sets (NEON, 3DNow...). See the application running on the Raspberry Pi :
► AUDIOSENSE Learn : linux command line software for learning new audio patterns (ACE) and then classify sound files on a PC. Useful for searching existing pattern on sound files given one or two examples. See how it works on this short video :