VoiceGrid™ LN

A single server network solution for 1:N searches, supporting a database of up to 10,000 records and up to 10 workstations. Quickly identifies or eliminates speakers by matched ranking scored by percentage of probability.

VoiceGrid LN interfaceVoiceGrid LN interface
Product Library:

Voice Grid DB

The voice matching in VoiceGridLocal system is accomplished through the utilization of the STC exclusive methods of processing, segmentation, feature extraction and identification of voice samples. This ensures the highest levels of accuracy and provides an additional and valuable tool in assisting in investigations where voice evidence may be present such as:

  • Kidnapping
  • Extortion
  • Domestic Violence
  • Terroristic Threats
  • Prank or False Report Calls
  • Corruption
  • Drug Trafficking
  • Gang & Organized Crime
  •  Abuse of Emergency or 911 Communication Channels

VoiceGrid Local is ideal for agencies with a larger voice database requirement but with only the need for a stand-alone system or multiple workstations within the same facility or campus. Additionally VoiceGridLocal can be delivered as either a software-only solution to be installed on customer provided hardware or as a full turn-key solution including both hardware and software.

The turn-key solution delivers additional benefits by providing out-of-the-box plug and play capabilities. This will allow the user to quickly utilize the system and begin immediate operations without having to tax their existing IT infrastructure or resources. Additionally it will alleviate any compatibility issues with existing hardware and be pre-configured to the exact specifications required.

Voice Grid LN

In addition to managing your own voice database SpeechPro offers an essential voice VoiceGrid identification module that allows you to make a biometric search of probable speakers through the whole voice database. Whenever you have unidentified voice recording you can always match it with the samples you have in your database.
Voice recognition draws similar parallels to fingerprint and face recognition biometrics in that a voiceprint also has available parameters that provide unique information about a person. These parameters are individual to each person, are created through the specific anatomical makeup of a person and therefore can provide distinctive information about the person’s identity.
Voice recognition technology can be utilized to provide unique and valuable enhancements to law enforcement investigations and identification operations. It is applicable in any situation were audio may be the only lead or evidence. Such cases include domestic abuse or calls in violation of protection orders, pranks or false emergency calls, agency radio communication abuse, inmate call monitoring, kidnapping, extortion, corruption, gang and organized crime communications and audio from video recordings where the speakers are not visually identifiable.


  • Quickly identify or eliminate speakers under investigation by matched ranking scored by percentage of probability
  • Create, monitor and manage the voice database
  • Full control over the administration and management of the voice data including access rights and security levels
  • Automatic speech signal resampling and segmentation
  • Ability to store and search by additional data (demographic, biographic) associated with the voice sample
  • Identification results can be viewed, analyzed and printed
  • Adjustable filtering to fine tune results
  • Compatible with IKAR Lab for expert analysis and verification

Technical specification

Maximum Voice Database Size 10,000 records
Maximum Number of Client Workstations 10
Estimated Search Time at Maximum DB Size 1 minute
Biometric feature extraction time 9.5 sec.
Standard Server Hardware Intel Core i7, 2.66GHz/8Gb DDR3/SSD 180Gb
Standard Server System Software CentOS Linux 5 (x64) / Oracle XE10 / Tomcat 6

Recommended Workstations

Workstation Hardware Intel® Core™ 2 Duo, 2.66 GHz
Workstation O/S Microsoft® Windows® XP, Microsoft® Windows® 7