Home | Site Map | Contact | |
| Products | Solutions | News | Technology | Company | Downloads





March 11, 2006

GritTec laboratory announced system of Automatic text independent speaker identification.

Automatic text independent speaker identification system is intended for automatic identification of a speech signal of unknown voice by comparing with "speaker cards", existing in the database of system. "Speaker card" means the structure of data, containing complete information about the given speaker (first and last name, births date, sex and etc) and its individual particularities of voice, including samples of sound files with its voice.

This system can be effectively used:
  • For automatic identification of unknown voice by phonogram of telephone negotiations;
  • In systems with high safety level, for instance, when access to digital information is limited by circle of given persons;
  • Applications where it's necessary to identify a person using peculiarities of his voice.
Designed algorithm of identifications is based on duel comparison spectra features of unknown speaker with the features of speaker card from the database system.

Spectra features are calculated with provision of dynamic determinations of channel distortion level and external hindrances. It allows to compensate channel distortion and influences of external hindrances with comparing spectra features, put into the original speech signal.

Sensitivity to identifications is defined by the level of installing the thresholds of probability of errors 1-th (False Rejection Rate (FRR)) and 2-th (False Acceptance Rate (FAR)) sort. Possibility of regulation of thresholds of FRR and FAR allows to adjust a process of identification flexibly in accordance with system safety requirements.

  • Operated with low SNR;
  • Fast adaptation to changing of channel distortion and external noises;
  • Speaker identification reliability not less than 91% if both of speech signals were recorded in the same channel and duration of input signal was not less than 15 seconds;
  • Speaker identification reliability not less then 85% if both of speech signals were recorded in different channels and duration of input signal was not less than 15 seconds;
  • Database client software;
  • Automatic identification doesn't require special skills;
  • Easy integration with target applications.

Signal requirement
  • Signal format: 16-bits linear;
  • 8 kHz sampling rate;
  • SNR, at least 10 db;
  • Frequency range: 300-3400 Hz or better.

About Grittec

GritTec laboratory specializes on research and development of the unique algorithms and technologies in the field of digital signal processing (DSP). GritTec's research is focused on speech enhancement & concealment, speech recognition, voice biometric, methods of adaptive filtration and other speech & audio technologies.
Url: http://www.grittec.com