Applications for voice commands to control “things” on the edge of the IoT are infinite, with the mass adoption of services on mobile devices (Siri), smart home systems (Alexa) and increasingly in smart cars (including Tesla’s newest models, which uses their own voice recognition software vs. embedding a third-party technology).
Voice works only as well as the software analyzing and acting on the voice can understand what humans are saying, and this is no small challenge given the amount of languages and dialects, accents and slang that exist across the global market. These challenges are made even harder by the noise inherent in vehicles, and for connected applications (GPS for example) the availability of resilient signals (WiFi in large cities, cellular, and satellite) to hold a session while in motion.
This is a challenge worth solving for drivers, given that very bad things can happen in the 5-10 seconds a driver's eyes are averted from the road to use a touchscreen.
The technology is getting better, with hundreds of millions being poured in by billion-dollar market cap companies like Samsung and others, to close the gap with increasingly advance software that reduces the learning curve for both humans and machines, with the ability to translate and apply unique voice patterns. We are all used to speaking on the phone in our cars today (much more safely with handsfree and voice commands in place for many years) so adding the feature to “set temperature to 68 degrees” is a natural extension that buyers will be willing to pay premium prices for.
DSP Group, an advanced technology, chipset and software company, announced their SmartVoice solution with voice enhancement algorithms built in was chosen by HARMAN (a Samsung company) to power their “Sonique” far field voice input solution, with an initial focus on Smart Home applications.
HARMAN (a massive company known for innovations including in the sound industry) has created an entire portfolio of customizable, high performance microphone arrays and a suite of voice algorithms to accurately capture voice commands even in high ambient noise environments – including smart cars.
They offer full stack integration with popular natural voice engines, and low distortion speakers, and the “voice command ecosystem” in general, and have been collaborating with DSP Group, including embedded their HDClear far-field multi-mic algorithm.
DSP Group, founded in 1986, and trading on Nasdaq since 1994, has its roots in VoIP related software applications and over the last few years has been expanding its capabilities into not just connected humans, but connected machines, leveraging its expertise in voice activated applications, improving those with higher and higher definition technologies including built-in CODECS.
“We’ve solved the hardest voice challenges for many years,” said Tali Chen, CVP Corporate Development & Chief Evangelist for DSP Group. “We’ve built multiple SOC (System on Chip) solutions for enterprise telephony endpoints for nearly every global VoIP company including Avaya and Cisco, supporting the SIP stack beyond traditional voice communications and extending to voice assistants, like Amazon’s echo.”
For HARMAN, DSP Group’s underlying technology keeps the device on and open to voice command with low power consumption, enabling voice activation in wearables, mobile devices and more.
“HARMAN, like others of our customers and partners, also choose us for our pre-processing algorithms, acoustic echo cancelling, and far-field communications to ensure their products deliver the ultimate experience,” Chen continued.
“With our DMBD5, product manufacturers can offer multiple endpoint solutions. Our chips come with the CODEC that translates for each and every application. Given that our software filters out noise – winds and other sounds – assistants, smart TVs, HVAC systems and other smart home solutions work better when the voice commands are more accurately heard and acted upon by the control systems.”
The specific innovation is a unique three-mic configuration intended primarily for consumer electronics applications, beyond cars, including home hubs connecting music devices, TV set-top boxes, Internet routers and appliances.
HDClear suppresses background noise and promotes high speech intelligibility for voice communications and user experience. HDClear also enhances the accuracy and performance of automatic speech recognition (ASR) applications, voice commands and voice activation, all with ultra-low power consumption.
When asked why DSP Group was chosen by HARMAN, Chen said,” It’s straightforward. We’ve been developing our technologies for years and never stop innovating and improving – while economics were important, performance was most important and we were able to balance both to ensure HARMAN’s end products would be commercially successful on a grand scale.”
“By leveraging DSP Group’s SmartVoice technology, HARMAN will be able to quickly add a best-in-class voice-user interface to the Sonique portfolio of mic arrays for enhanced accuracy in high noise environments,” commented Dave Rogers, President of HARMAN’S Lifestyle Audio SBU. “Our customers expect the highest level of technology from our products. Working with DSP Group allows us to deliver our promise of an excellent user experience.”
Parks Associates research shows that 41% of consumers in U.S. broadband households use a voice assistant application or device. “Companies are creating a world of devices that can be commanded naturally with the simplicity of a vocal command already,” said Dina Abdelrazik, Research Analyst at Parks Associates, “Improvements in the underlying technology, such as enhanced far-field capabilities, will lead to a better experience for voice-first activities. For instance, Parks Associates data reveal that 43% of people already use voice control to stream music while 14% use it to control a smart home device.
DSP Group provides wireless chipset solutions for converged communications. Delivering semiconductor system solutions with software and hardware reference designs, DSP Group enables OEMs/ODMs, consumer electronics (CE) manufacturers and service providers to cost-effectively develop new revenue-generating products with fast time to market.
DSP Group provides a broad portfolio of wireless chipsets integrating DECT/CAT-iq, ULE, WiFi, PSTN, HDClear™, video and VoIP technologies, enabling converged voice, audio, video and data connectivity across diverse mobile, consumer and enterprise products – from mobile devices, connected multimedia screens, and home automation & security to cordless phones, VoIP systems, and home gateways.
HARMAN designs and engineers connected products and solutions for automakers, consumers, and enterprises worldwide, including connected car systems, audio and visual products, enterprise automation solutions; and services supporting the Internet of Things.
Edited by
Ken Briodagh
