Knowles, the largest supplier of embedded microphones, said that it had introduced a digital signal processor that could give electronic devices the ability to be controlled with simple voice commands. The processor could also allow smart speakers and other consumer devices to more accurately discern voices in noisy environments and suppress background noise.
The company's latest chip aims to give voice control capabilities to everything from alarm clocks and household appliances to smartphones and other consumer devices that feature voice assistants such as Amazon’s Alexa, Google’s Home and Apple’s Siri. Last year, sales of smart speakers in the fourth quarter were around 4.2 million, up from 600,000 in the previous fourth quarter, according to Strategy Analytics.
The chip is capable of processing process human speech, sifting through background noise, echoes, and reverb while amplifying and sharpening the speaker’s voice. The company added custom silicon to the IA8508 so that it can handle artificial intelligence algorithms that would typically run on the cloud.
Michael Polacek, the company's president of intelligent audio, said that customers are trying to process voice commands “closer to the human, and that means closer to the microphone.” He added to Electronic Design: “If you can move more computing from the cloud to the device and then from deeper in the device to the edge, you can better understand context and how many people are in the room.”
The chip contains three digital signal processing cores. The high-performance core can be programmed to boost sensitivity to faraway voices, while the low-latency core is designed improve active noise cancellation. The low-power core can be programmed to listen for several words simultaneously – enabling users to utter “bake” and “preheat” when controlling an oven – without consuming too much battery life.
On-device processing results in lower latency and greater privacy for users, according to Knowles. Lower latency could be important in smart headphones that have to amplify faint sounds and suppress loud chatter on a crowded city street in real time. The IA8508 will enter production in the first quarter of 2018.
The company said that the chip would lower the bar for manufacturers to add voice control capabilities to devices. Ultimately, that could help Knowles sell more microphones, its core business. The company holds more than 40% of the microphone market. The chip, which is designed to provide electronic devices with more contextual awareness, can connect to eight microphones
The competition is also intensifying. Austin, Texas-based Cirrus Logic sells chips that nullify background noise and deduce the direction of sounds using four-microphone arrays. Last month, Xmos announced that it had raised $15 million in venture capital to fund the development of multicore microcontrollers for voice applications. Another major player is Silicon Valley-based Synaptics.
These companies are focused on improving the current state of far-field voice processing. “Five years ago, there was an assumption that you would be talking to a device within five feet, and many times within one or two feet,” explained Todd Mozer, chief executive officer of Sensory, which sells keyword detection software that can be embedded in light bulbs, kitchen appliances, and other battery-powered devices.
“Over the last few years, people have said that they want to recognize voices from thirty feet away, which means more reverberations, more background noise, and higher signal-to-noise ratio,” he told Electronic Design.
Many companies are also trying to give electronic devices the ability to handle more and more advanced artificial intelligence applications. Last month, Google’s Pete Warden called on the semiconductor industry to develop computer chips that can grasp basic commands like on and off – and simple sounds like a grasshopper’s chirping or knocking in an industrial machine – and are inexpensive enough to be disposable.