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





April 4, 2008

GritTec laboratory updates Speech Enhancement technology on the base of dual microphone array solution.

    In the updated version of dual microphone array (DMA) module Multiple Canceller (MC) updating has been made. Also integration DMA with technology on the basis of Noise Cancellation (NC) has been made (see fig.1).

Fig.1. Scheme of modules GritTec's Combo-SpeechEnhancer

   GritTec's Combo-SpeechEnhancer is complex using of technologies DMA and NC allows at first by means of DMA to suppress all external hindrances and background noises in a direction of the chosen speech source, and then by means of NC to suppress all remained external noises and absent-minded hindrances what DMA could not suppress completely. The sources of external hindrances can be an external noise of car, speech of collocutor, music and any other sounds, coming from directions, different from direction of the chosen source of the speech.
   The made updating in MC allowed to receive on an output of the dual microphone array the clear speech with lower speech distortions and higher suppression of external hindrances and noises. This result was possible to reach at the expense of work optimization of the adaptive filters which used in module MC.
   Also this updating in MC automatically allows expanding quantity of adaptive filters for suppression of other external hindrances and suppression of an acoustic echo. At the moment experiments on suppression of other external hindrances and an acoustic echo are not finished yet.

General features of GritTec’s Combo-SpeechEnhancer (ver.1,0):

  • Support of 1 or 2 microphones configurations;
  • Configurations: DMA&NC, DMA or NC;
  • Delay of the algorithm: 8 ms;
  • Noise suppression not less 25 db (for DMA&NC);
  • High quality of speech enhancement.

General features for new version DMA (ver.1,3):
  • Support of 2 microphones configurations;
  • Delay of the algorithm: 8 ms;
  • Directional microphones: unidirectional configuration (omni-uni directional);
  • Distance between microphones not less 0.04 meters;
  • Distance to the speaker not less 0.5 meters (up to 5 meters (?));
  • Choice of a direction of a speech source: hand control;
  • Noise suppression not less 15 db;
  • Low level of speech distortion.

Examples of work GritTec’s Combo-SpeechEnhancer (ver.1,0) with use of various configurations:
  • initial signal from 1st microphone ;
  • previous version DMA(ver.1,21) ;
  • new version DMA(ver.1.3) ;
  • DMA(ver.1.3) and NC ;
  • suppresion with NC .

At the moment demonstration version of GritTec’s Combo-SpeechEnhancer is accessible only on request. Some parameters at the moment are still specified.

About GritTec
GritTec Laboratory (GritTec Ltd.) specializes on research and development of algorithms and technologies in the field of speech and audio processing. GritTec's research is focused on speech enhancement, speech concealment, voice biometric, speech recognition, speech synthesis and other speech and audio technologies.
url: http://www.grittec.com