Malaspina Labs Launches VoiceBoost Voice Activation and Biometric Speaker Verification Solutions
Malaspina Labs has launched VoiceBoost Voice Activation solution and VoiceBoost Biometric Speaker Verification solution for NXP's LPC54100 Series of microcontrollers. Both solutions provide voice wakeup and voice-command capabilities in an always-listening mode using very low power and are effective in high background noise with users speaking at distances up to several meters away.
Malaspina's VoiceBoost Speaker Independent Voice Activation solution allows LPC54100-based devices to perform always-listening voice activation and voice commands entirely on the device, with an incremental current consumption of just 0.1mA on average in noisy environments. Devices using the VoiceBoost Speaker Independent Voice Activation solution allow users to interact with devices without the need for gestures, typing, or button presses by continuously listening for a wakeup phrase, even while in an ultra-low power sleep state. The complete system, including microcontroller core, all integrated peripherals, voice activation solution and microphone, consumes as little as 2.2 mA on average in noisy environments. In addition to waking devices with a spoken phrase, the solution can be used to implement factory-defined voice commands.
Used separately or in conjunction with the speaker independent solution, the VoiceBoost Biometric Speaker Verification solution enables voice authentication of one or more users. This allows devices to add access control functionality, such as device locking and parental controls, by responding only when an enrolled user's private passphrase is spoken in the user's own voice. In addition to its authentication uses, the speaker verification solution allows device manufacturers to add multi-user capabilities to devices. The speaker verification solution requires incremental current consumption of just 0.08 mA on average in noisy conditions. The complete system, including microcontroller core, all integrated peripherals, speaker verification solution and microphone, consumes less than 1.0 mA on average in noisy environments.
"Using voice commands to interact with an increasing number of smart wearable and IoT devices is a hot new trend in user interface design. But continuously listening for, recognizing, and processing voice commands is challenging from power, memory, and performance standpoints. Malaspina's VoiceBoost solutions and our LPC54100 Series microcontrollers are a great combination: both were designed to consume minimal power in always-on listening and awake processing states. And Malaspina's voice solutions remain accurate even in less than ideal environments, such as noisy rooms or when users are more than an arm's length away," said Brendon Slade, director of tools and embedded ecosystem at NXP Semiconductors, in a statement.
"NXP's LPC54100 Series of microcontrollers offer great performance at low power, which our VoiceBoost speech software can take full advantage of," said Dean Neumann, CEO of Malaspina Labs, in a statement. "We are delighted to be able to offer optimized voice activation and speaker verification solutions that showcase the performance and versatility of the LPC54100 Series."