Fusion Voice Engine™

Fully-Integrated, Modular Suite of Codecs, Algorithms, and Critical Voice Services

The Fusion Voice Engine is a comprehensive solution consisting of an extensive set of algorithms and codecs designed for Digital Voice applications. The Fusion Voice Engine is designed for a wide range of applications, like PC VoIP, Smart Phones, RoIP devices, and Embedded Voice applications. It allows application developers, service providers and hardware manufacturers the ability to easily build VoIP or Radio over IP (RoIP) technology into their solutions.


Integrated Codecs, Algorithms, and Voice Services

The Fusion Voice Engine has a high level API to easily manage Analog Channels, Digital Channels and the Mixer. Fusion Voice Engine is extensible, with a plug-in architecture, allowing the user to add in custom codecs and/or proprietary echo cancellation components. It is designed to support sampling rates of 8kHz (Narrow Band Audio), 16kHz (HD Audio), 32kHz (Ultra-HD Audio) and 48kHz (CD Audio).


Fusion Voice Engine Features

The Fusion Voice Engine can be run on virtually any OS/hardware configuration. Along with all the various codecs and algorithms, Voice Engine customers also receive some of Unicoi's Fusion Embedded Protocols, including:



  • G.711U
  • G.711A
  • G.722
  • G.726-16
  • G.726-24
  • G.726-32
  • G.726-40
  • G.729
  • DVI4
  • DVI4 HD
  • DVI4 Ultra HD
  • iLBC
  • Speex
  • Speex HD
  • Speex Ultra HD
  • SILK
  • L16 PCM
  • L16 PCM HD
  • L16 PCM Ultra HD
  • L16 PCM CD


  • Gain Control
  • Automatic Gain Control (AGC)
  • DC Blocker
  • High-Pass Filter
  • Voice Activity Detector (VAD)
  • Audio Resampler (8K, 16K, 32K, 48K)
  • DTMF Generator
  • DTMF Detector
  • Call Progress Tone Generator
  • Custom Ring Tone Generator
  • Comfort Noise Generator (CNG)
  • Conference Bridge (Mixer)
  • Packet Loss Compensation
  • Acoustic Echo Canceller (AEC)
  • Noise Reduction
  • Frequency Equalizer

Voice Engine Architecture

Fusion Voice Engine’s architecture centers on conversations, channels and streams. A conversation contains one or more channels which in turn contain a single or a mated pair of streams which are, in general, a sequence of audio processing algorithms. More specifically, an analog channel always contains a mated speaker and microphone stream, and a digital channel contains an encoder and/or a decoder stream. A typical 3-way conference call scenario is depicted here:

Voice Engine Architecture Diagram
InstaVoIP Architecture Diagram
*Click here to download .png version of diagram

Phone: +1-678-208-2250 E-mail: