Recent times have seen significant advances in the voice technology landscape, with devices like smart speakers and home assistants being welcomed into consumers’ homes across the world.
However, despite their popularity, these gadgets are far from perfect. There are still many challenges that stand in the way of a hassle-free experience for the user. Perhaps the most common issues are when users are trying to communicate with their device in a noisy environment or when using words that sound alike.
Many will have experienced a device like Alexa misunderstanding their request. And so, it’s safe to say that we’re still a long way away from an optimal voice solution that can be seamlessly integrated into our everyday lives.
To address this issue, XMOS has partnered with software company, Sensory. The company develops and licenses technologies for speech recognition, natural language understanding, face and voice biometrics, wake words, computer vision, sound identification and more.
This partnership is set to usher in a new era of user-friendly, effective, and innovative voice control technology – where interaction with digital devices will be as natural and seamless as speaking with a friend.
How the partnership works
xcore-voice is an open-source, software-defined, low-power voice solution built on the unique xcore.ai platform. Along with voice interface algorithms, It is an innovative technology that unlocks a new level of integration and customisation, including AI models, for developers, allowing them to design within software. It comes with immediate out of box example application that can be run on a L71 voice development board
Sensory’s TrulyHandsfree technology, which is integrated in one of the example applications, offers a unique solution in voice recognition, which perfectly complements the capabilities of the xcore-voice offering.
The many benefits of this technology integration include:
- High accuracy in challenging environments – the ‘wake word engine’ boasts high accuracy, even when faced with loud background noise, accented speech or far-field conditions.
- Unparalleled offline performance – Local commands can be run offline using xcore-voice without the need for a connection to the cloud. This benefits users and helps manufacturers create unique products in a crowded market.
- Total versatility and high compatibility – Designed to integrate seamlessly into solutions such as xcore-voice.
- Low power performance – Once given a command, it activates and actions it, before reverting back to low-power mode once commands are no longer being received or detected.
- Online support for developers – Sensory’s VoiceHub portal allows developers to create up to ten custom wake words in multiple languages, plus various phrase spotted commands, and natural language grammars at no additional cost.
- Unmatched flexibility – Being compatible with both mobile phones and desktop PCs, VoiceHub can also be used across many applications in-situ or while on the go – providing further flexibility for users.
- User freedom – Devices powered by XMOS and Sensory can be used across multiple platforms such as Siri, Amazon, Google, and more, giving users one central device to conveniently access multiple services.
All of these features and benefits result in listening devices designed for your everyday needs that can respond and listen to you as a friend, rather than acting like a robotic and rigid servant.
Looking to the future
The partnership between XMOS and Sensory is set to pave the way for a truly intelligent voice solution. This will provide superior voice-activated control and experiences for more diverse applications than ever seen before, whether it’s personal assistants or industrial automation or in-cabin automotive use cases.
Together we are shaping a future where interaction with digital devices is as effortless, efficient, and flexible as possible. It’s an exciting time for developers, companies, and consumers alike as we navigate this transformative era of voice technology.
To find out more about our partnership and our joint technological offerings, click here to watch our recent webinar.