Voice Enhancement

Voice Enhancement Package

About

Voice Enhancement Package (VEP) is a suite of real-time software DSP technologies designed for improving speech recognition performance in voice-controlled multimedia devices. VEP is a front-end solution that performs audio signal pre-processing prior to the Automatic Speech Recognition (ASR) and Key Word Recognition (KWR) engines. 

VEP is an ASR-engine agnostic signal processing package that can be used in conjunction with any ASR engine. The technologies included in VEP are universal and do not interfere with signal processing performed as part of modern ASR’s.

VEP Components

VEP includes a number of signal processing blocks necessary to improve speech recognition, including:

• DC-removal
• Acoustic echo-canceller (Mono or Stereo)
• Multi-microphone beam-former
• Gain control

VEP DSP blocks are highly configurable allowing full flexibility in hardware, acoustic and software design of the final product.

Alango Voice Enhancement Package (VEP) confined in grey background:

• N-microphone (N microphone) input, up to 16 microphones
• M-microphone output (dependent on number of microphones)
• Stereo Acoustic Echo Canceller with configurable echo tail (receiving 2 echo-reference signals)

Technical Information

PERFORMANCE

• Echo canceller
  • Fast convergence (< 300ms)
  • Robustness to speaker signal distortions
  • Long echo tail (up to 1000ms)
• Beamforming array
  • Configurable geometry with up to 16 microphones
  • Wide dynamic range (close and distant sounds)
  • Attenuation of different types of noises: direction acoustic noises, diffused noises, reverberation and self-noises of microphones

AVAILABILITY

VEP is available* on the following platforms:

• ARM cores
• Cadence (Tensilica) HiFi3
• CEVA TeakLite IV
• Synopsys ARC core

* Implementation of VEP may differ between platforms. Please, contact Alango technical support for specific information.

Porting on other platforms can be performed quickly.

Audio Examples

Additional Information

VEP Integration Guide is available upon request.

Please contact us for more information, current MIPS/memory consumption, audio examples and real time demo availabilities.