A leading wireless audio SoC provider adds Audio and Voice processing without the battery penalty

Customer case study | Ceva-NeuPro Nano

A leading wireless audio SoC provider selected Ceva NeuPro technology to add embedded intelligence to high-volume audio and IoT devices. The objective is to enable always-on AI audio and smarter sound processing while meeting strict battery and cost constraints. 

The NeuPro Nano implementation builds on Ceva IP already present in the customer platform, supporting a faster integration path for neural-network, DSP and control processing.

Customer’s target markets & applications

  • Home audio 
  • Wireless audio 
  • IoT devices 
  • Sound processing 

 

The challenge

The customer needed audio and voice processing capabilities in Bluetooth SoCs and a fast integration path on existing platforms, without creating an unacceptable battery or cost penalty.

Why Ceva

  • Power-efficient Audio: Ceva enables embedded audio processing while addressing the battery and cost constraints of high-volume connected devices. 
  • Simplified SoC integration: NeuPro Nano combines neural-network, control, and NN processing and control processing on one core, reducing integration complexity. 
  • A complete platform approach: NeuPro can be paired with Ceva Bluetooth IP already associated with the customer’s platform. 
  • Support for voice-enabled applications: Access to a wake-word model ecosystem provides a practical path for adding always-on voice intelligence. 
  • Faster deployment: Building on Ceva IP already integrated into the platform helps shorten the path to adding new AI capabilities. 

Implementation Highlights

  • Ceva-NeuPro Nano (32) 
  • Neural-network, DSP and control processing on one core 
  • Pairs with Ceva Bluetooth IP