DSPG Unveiled Ultra-Low-Power Always-On Voice Processor

22 February, 2016

The new DBMD4 chip is Optimized for Google Hotword and Sensory TrulyHandsfree, and designed for Mobile, Wearables and IoT

The new DBMD4 chip is Optimized for Google Hotword and Sensory TrulyHandsfree, and designed for Mobile, Wearables and IoT


DSP Group from Herzliya, Israel, unveiled DBMD4 – its next-generation, ultra-low-power, always-on voice and audio processor. The DBMD4 platform features a notably small footprint, offers the lowest power consumption and best-in-class enhancements for always-on voice operation, elevating performance and functionality of battery-operated devices.

Adoption of voice-enabled applications such as smart assistants, voice commands and voice search is proliferating in mobile, smartwatches and IoT devices. DBMD4 allows battery-operated devices to actively listen and sense voice activity and commands while in ultra-low power mode – alleviating battery strain, improving device usability and extending battery life.

DBMD4’s power-optimized implementation for always-on voice is based on Sensory TrulyHandsfree and Google Hotword. In addition, DBMD4 supports six seconds of voice buffer that enables seamless operation. DBMD4 is a completely autonomous, audio-centric processing solution enabling power-hungry application processors in mobile and wearable devices to actually sleep during voice sensing.

“The consumer and enterprise use of mobile smart assistants is on the rise. This creates a real need for a low-power solution that enables users to enjoy these functionalities via a simple voice-enabled interface,” said Harry Wang, Research Director at Parks Associates. “Integrating with both mobile and wearable devices, DSP Group’s solution will further increase the usage of such applications – creating value for application developers, service providers, and end users,” he concluded.

“We are excited to be working closely with DSP Group to enable the first DSP with Sensory’s low power always listening hardware block,” said Todd Mozer, CEO of Sensory. “DBMB4 features built-in handshaking with Sensory’s TrulyHandsfree, which gives it a winning combination of power consumption and accuracy in trigger word listening environments,” he continued.

High-Performance Voice Enhancement Suite

DBMD4 incorporates a suite of voice enhancement algorithms including noise suppression that significantly improve user experience and accuracy of speech-driven applications, particularly in high noise environments. DBMD4 also supports advanced audio and voice features including Acoustic Echo Cancellation (AEC), Automatic Gain Control (AGC), Beam Forming and more.

“We foresee a bigger role for voice and audio in consumer devices for smart processing and sensing – including mobile, wearables and IoT,” said Ofer Elyakim, CEO of DSP Group. “With advanced features like low-power always on and best in class voice enhancement solutions, we believe that DBMD4 is uniquely positioned to play an instrumental role in this evolution,” he added.

Share via Whatsapp

Posted in: News , Semiconductors