AudioClassify.com

Free Server Demonstrations Are Available Now

Custom High Performance C, C++, and NodeJS for Linux Ubuntu Server or 32-bit/64-bit ARM Device
Custom Uncompressed Audio (S16LE PCM) JavaScript for Web Browser (e.g. Chrome or Chromium)
521 Audio Classes in Real-time (TensorFlow YAMnet Model)
Pre-defined Alert Groups (e.g. Danger and People) with Four Selectable Schedules Per Alert Group
User-defined Alert Groups with Easy to Use Browser Based Editor
Pre-defined Schedules, e.g. Calendar ("New Years Day") or Day of Week ("M-F 8:00 a.m. to 5:00 p.m.")
User-defined Schedules for Calendar and Day of Week
Audio Recording of Alerts (up to 120 seconds before and 120 seconds after alert condition)
Alert Notification via Email, SMS, HTTPS Post, and/or Output Pin Activation
Web Browser Microphone Inputs (Uncompressed S16LE, 16,000 samples/sec, 256 Kbps) for Accurate Audio Classification
Internal High Sensitivity MEMS Microphone (Uncompressed S16LE) for Embedded Linux Devices
Optional External High Sensitivity MEMS Microphone
Optional RTSP Audio Input (accepts audio only or video/audio RTSP stream, uncompressed S16LE is recommended)
Custom Audio Classification Models Available
All Custom Application Software has been written in the U.S.A.

Linux Server Demonstration on Ubuntu 20.04 in Major Datacenter
Server Demonstration utilizes Two Web Browser Microphone Inputs (Uncompressed S16LE via Custom Software)
Intel and AMD Octal (and higher) Core 64-Bit Architectures Can Support Eight or more Concurrent Audio Classification Channels
For demonstration, alert recordings are 60 seconds (30 second before and 30 seconds after alert condition)

ARM Demonstration Units on Raspberry Pi 4 Model B 64-Bit OS With Two Simultaneous Audio Inputs
Audio Input 1: High Sensitivity MEMS Microphone Connected to RPi4 via I2S
Audio Input 2: Web Browser Microphone Input (Uncompressed S16LE via Custom Software)
For demonstration, alert recordings are 60 seconds (30 second before and 30 seconds after alert condition)
Please Schedule Server Based Demonstration First
RPi4 Demonstration Units Have Refundable Fee and are Subject to Hardware Availability

Incorporate Audio Classification into Existing Products (e.g. security cameras)
Enhancement of Existing Video Monitoring Systems (e.g. detect activity that cameras cannot see)
Building Security (e.g. detect voice, footstep, or other unauthorized or dangerous activity)
Scientific Monitoring and Studies (e.g. detect animal, bird, machine, and human activity)
Environmental Monitoring and Studies (e.g. detect animal, bird, machine, and human activity)
Patient Monitoring (e.g. detect labored breathing with custom model)
Manufacturing Monitoring (e.g. detect sounds that indicate impending failures with custom model)

Email:

Custom application software source code and a non-exclusive software license will be provided.
Product specifications, terms, conditions, and availability are subject to change.
The user accepts responsibility and liabilty for the ethical and lawful use of audio recording and monitoring.
Any and all Business Agreements must be mutually agreed to in writing.

Audio Classification Custom Software Source Licenses
Server and Embedded Linux Applications