Malaspina Labs extends standalone biometric speaker verification to NXP LPC5410x microcontrollers
September 22, 2015 — Malaspina Labs announces the release of its completely standalone VoiceBoost™ Biometric Speaker Verification solution for the LPC5410x family of microcontrollers from NXP Semiconductors.
In contrast to “speech recognition” solutions which interpret what commands are spoken to control a device, “speaker verification” solutions use voice recognition to identify who spoke those commands. Speaker verification is useful for user authentication (device unlocking) as well as for parental controls or access controls (limiting access to certain functions) and for device personalization (multi-user profiles). All speaker verification solutions require that users first “enroll” their voice prior to a device being able to “verify” or “authenticate” the user by recognizing his or her individual voice. This voice enrollment step extracts the unique characteristics of each user’s voice and distills those characteristics into a “voiceprint” that is used to subsequently recognize the user. Traditionally this voice enrollment step has required large amounts of computing power and memory, requiring that the voiceprint be generated on a very capable computer or as a cloud-based solution, and then downloaded to the target device for use during recognition. This has limited the use of biometric speaker verification on low cost or low power consumer devices.
Malaspina Labs has achieved a breakthrough 100x reduction in the size of the software required to create the voiceprints used for subsequent recognition. This breakthrough allows Malaspina Labs to provide the industry’s only biometric speaker verification solution in which the voice enrollment step executes on the same low power device as the recognition. Malaspina’s breakthrough allows manufacturers to now create low power consumer devices such as wearable, Internet of Things and home automation devices which incorporate device locking, parental controls, and multi-user capabilities without the additional cost or complexity of pairing the device with a mobile phone, computer or Internet for voice enrollment.
Malaspina has released an on-device voice enrollment and voice recognition solution optimized for the LPC5410x family of microcontrollers from NXP Semiconductors. The voice enrollment step executes entirely on the LPC5410x device requiring less than 90KB of memory and less than 50 Mcycles for real time operation. Once one or more users have enrolled their voices, Malaspina’s Biometric Speaker Verification solution executes entirely on the same LPC5410x device in an ultra-low power “always listening” mode, recognizing each individual user’s voice even in noisy environments or with users speaking at arm’s length and beyond, while consuming just 0.08mA current on average. This completely standalone on-device voice enrollment and recognition solution enables manufacturers to create smarter wearable and IoT devices without the additional costs and complexities of connecting the device with a mobile phone or computer in order to enroll users’ voices.
“On-device enrollment is a user-experience game changer for low power devices, such as wearables and portable IoT devices,” said Brendon Slade, Director of Tools & Embedded Ecosystem, NXP Semiconductors. "The NXP LPC5410x microcontroller's ultra-low power consumption enables both on-device enrollment of voice biometrics and efficient always-on listening for voice commands, a dramatic improvement in ease of use and battery life in a wide range of end products.”
Dean Neumann, CEO of Malaspina Labs said “Malaspina is pleased to be able to offer our standalone biometric speaker verification solution for NXP’s LPC5410x microcontrollers. The capabilities and ultra-low power consumption of the LPC5410x family makes it a great choice for manufacturers looking to add voice features and user-specific functions to any consumer device. Users can now enroll and use voice biometrics on any LPC5410x based device from anywhere. No more USB tethering or Bluetooth pairing or WiFi connection to a hotspot needed.”
About Malaspina Labs
Malaspina Labs develops speech interface and enhancement software for ultra-low power devices which improves person-to-person and person-to-device voice communication in high noise environments. Malaspina Labs' VoiceBoost™ suite of solutions provides speech isolation and voice interface technology that is effective at handset, speakerphone and arm's length distances and beyond. VoiceBoost solutions operate deterministically at bone-conduction, telecom, and voice interface latencies within the constraints of milliwatt-class processors designed for mobile, embedded, medical and consumer devices.
For more information about Malaspina Labs, visit