VoiceGrid™ SDK

VoiceGrid™ SDK provides the same voice biometrics tools and algorithms available within the VoiceGrid™ line, but gives business partners with development resources the opportunity to integrate voice recognition technology into their own platforms.

VoiceGrid™ SDK
Product Library:

Application

The VoiceGrid™ SDK provides voice matching biometric tools and technologies which developers can integrate into their own platforms or systems. The VoiceGrid™ SDK offers reliable, accurate and rapid identification on all sizes of databases including those in excess of millions of records. The VoiceGrid™ SDK is ideal for both uni-modal and multimodal biometric identification within systems designed specifically for fast and accurate results.

Highlights

  • 3 exclusive SpeechPro algorithms
  • Decision-level fusion
  • Audio preprocessing
  • Analysis of recordings as short as 3 seconds
  • High speed of matching – 1,000,000 voice samples in less than a minute
  • Noise filtering (up to 10dB sound-to-noise ratio)
  • Automatic dialog segmentation (2 speakers)
  • VoiceGrid™ SDK can work under Windows 32-bit and 64-bit as well as Linux

Features

  • Voice as a new biometric modality in your system
  • Higher accuracy with additional modality
  • Opportunity to add new capabilities to existing systems
  • Suitable to large-scale identification and verification
  • Development and integration time significantly reduced
  • Voice biometric algorithms are top performers at NIST evaluation tests
  • Technical support 24/7

Technical specification

Platform
Windows 32/64
Linux 32/64
Hardware requirements
100,000-sample database (not including disk space for audio):
- Core 2 Duo CPU @ 2 GHz
- 2 GB RAM
- 2 GB free disk space
Each additional 100,000 samples:
- 1 GB RAM
- 1 GB free disk space
All the requirements are given for simultaneous application of all three methods.
OS (Operating System)
Microsoft Windows 7 (32-bit or 64-bit)
Microsoft Windows XP (32-bit)
Microsoft Windows Server 2008 (32-bit or 64-bit)
Cent OS Linux 5
Red Hat Linux 5
Development tools
Microsoft Visual Studio 2005, 2008, 2010 for C++ API
Any development tool for HTTP API
g++ 4.1 for C++ API
Any development tool for HTTP API
Programming APIs
C++ API
HTTP API
System functions
Preprocessing and signal quality evaluation
2-speaker dialog separation (diarization)
Enrollment, verification, identification
Computing performance
From 6 to 12 biometric models per minute on single core processor
From 3,000 to 40,000 comparisons per second on single core processor
Recommended database size
100,000/2,000,000 samples
100,000/2,000,000 samples