Speech Technology Magazine

Conexant and Harman Collaborate to Improve Speech

Together the two companies are set to deliver premium end-to-end audio and voice solutions for manufacturers of voice-enabled devices.
Posted Jun 20, 2017
Page1 of 1
Bookmark and Share

Conexant Systems, a provider of audio and voice technology, has begun collaborating with HARMAN International, a wholly-owned subsidiary of Samsung Electronics, to deliver end-to-end audio and voice solutions for manufacturers of voice-enabled devices, smart home, automotive, and enterprise applications, even when used in high ambient noise environments.

"Voice is everywhere. We're seeing it flourish in consumer electronics and appliances of all types, automobiles, robotics, and even in work-place solutions," said Saleel Awsare, president of Conexant, in a statement. "By expanding our relationship with HARMAN to offer co-developed embedded far-field voice solutions, we're able to offer an even greater selection of comprehensive premium audio and voice solutions to leading brands around the world, allowing them to produce innovative voice-enabled applications that offer accuracy and performance that is second-to-none."

Enabling speech recognition and voice control from a distance requires overcoming substantial acoustic challenges related to echo cancellation, background noise, microphone positioning, speaker placement, and more. Conexant's 2 and 4-microphone voice processing solutions are designed to work with any wake word and natural voice engine to deliver verbal commands from anywhere in the room, even in noisy, real-world environments. The solution also enables voice barge-in, allowing users to interrupt their voice-enabled devices when they're playing sound. Through this collaboration, Conexant and HARMAN will help customers reduce engineering time and development costs for creating high-performance, noise-robust voice-enabled devices.

Offered in microphone configurations for a variety of device applications and design identities, the core component of HARMAN's 2- and 4-microphone high-performance far-field voice solutions is Conexant's AudioSmart voice input processors running Conexant's far-field voice pre-processing technology. Conexant's AudioSmart voice input processor system captures the user's voice from anywhere in the room, separates the voice commands from music and voice prompts being played out of the device, even with background noise present, and provides a clean audio signal to the AI system for speech recognition processing.

Key benefits of the Conexant AudioSmart voice input processors include the following:

  • Robust far-field voice interaction compatible with all major automated speech recognition (ASR) services;
  • Proprietary Smart Source Locator technology that identifies the direction from which the user's voice is coming;
  • Enhanced Smart Source Pickup noise suppression technology that filters out ambient noise, including non-stationary TV noise and voices coming from sources other than the user; and
  • Voice barge-in enabled by full duplex acoustic echo cancellation (AEC) that detects the wake word even when the device is playing music or voice prompts loudly.

"We expect the market for voice-controlled devices in the consumer electronics, enterprise, and robotics space to grow dramatically. Partnering with Conexant will enhance our foothold in this valuable market. We're very excited about this partnership and the opportunities it will present to both parties," said Dave Rogers, senior vice president of consumer audio at HARMAN, in a statement.


Page1 of 1