Since Amazon Echo launched in 2014, smart speakers and voice assistants have been enjoying rapid growth. Voice interfaces such as Alexa and Siri have shown faster adoption than almost any other technology in history. According to a research report by Global Market Insights, the smart speaker market is set to exceed 30 billion dollars by 2024. Music streaming has been the dominant play to date, but new technologies are helping to inspire more compelling applications and use-cases that make consumers’ lives easier – bringing the promise of IoT a step closer. Future source’s latest estimates for Home CE with ‘built in’ voice assistants were revised upwards from 2017 based upon increased shipments found within the media streamer category. As a result, shipments increased to 75.6 million units across all device categories.

While the voice market grows exponentially, one of the key challenges facing the long-term viability of personal assistants is the ability to identify and extract a single speech command from within a noisy environment (often called the ‘Cocktail Party Problem’).

Enter XMOS XVF3500: Capturing voice across noisy environments with unrivalled accuracy.
Launched at Mobile World Congress 2018, this is where the unrivalled accuracy of XMOS’ VocalFusion XVF3500 far-field stereo-AEC voice processor comes in to its own. Embedded in 3rd party voice-enabled devices such as smart-TVs, set-top-boxes and soundbars, the powerful UK engineered processor gives you up-close voice capture quality and processing accuracy from across the room – even in noisy environments. A key component in driving the voice-enabled market forward, it remains the first and only Amazon Voice Service (AVS) qualified stereo acoustic echo cancellation development kit available on the market.

The XVF3500 stereo-AEC voice processor provides:

• Stereo acoustic echo cancellation: removes acoustic feedback between the device’s speakers and microphones to capture voice commands accurately (when audio is playing through the stereo device).
• Beamforming: identifies the direction the voice command comes from, even if moving.
• Dereverberation: removes the echoes from your voice as it ‘bounces’ off the hard surfaces in the room.
• Noise suppression: removes exterior noises, e.g. air conditioning, people talking, pets etc.
• Barge-in: enables users to interrupt, when audio content is playing through the speakers, with a voice command spoken at normal volume from across the room.

Use case: Freebox
The Freebox Delta, brainchild of innovative French telecoms operator Free and developed in partnership with Devialet, is a high-performance fusion of technologies. A set-top box, media streamer, soundbar, WiFi server and smart home hub that responds intelligently to both touch and voice. We are proud to have provided the far-field voice capture capability for its two on board personal assistants– Alexa and OK Freebox. The ability to create a smart home product with a choice of voice assistant creates a whole new world of possibilities for manufacturers, operators and consumers alike. The ease of integration into existing products, as well as our development kits help clients such as Freebox move rapidly from prototype to commercial deployment.

For more information click here.