The industry leader in mixed-signal, Awinic Electronics, proudly unveils its new high-performance NPU neural network intelligent voice processing chip: the AWA89601. Integrating a dedicated audio NPU (Neural Processing Unit) and combining sound model training with NPU hardware, this chip achieves groundbreaking advancements in AI noise reduction, AI voice enhancement, AI far-field pickup, and AI voiceprint recognition. It empowers smart terminal devices to deliver an interactive experience where "the voice feels as clear as being face-to-face."
Product Advantages
AI Deep Noise Reduction Engine
Supports Deep Learning Noise Reduction (Deep Learning NR) and complex environmental noise reduction (e.g., wind noise, keyboard sounds, traffic noise).
Intelligent Voice Processing
Supports far-field pickup, precise voice separation, and dual-MIC beamforming to enhance far-field pickup accuracy.
Supports Acoustic Echo Cancellation (AEC), effectively suppressing device speaker echo to ensure interference-free calls.
Supports Voiceprint Recognition (VPR), providing biometric-level identity verification for secure voice interaction.
Innovative Integration of Edge-Side NPU
Integrates a dedicated audio NPU, providing edge-side inference computing power and supporting real-time processing of 4-channel upstream audio.
Compared to cloud-based solutions, local data processing avoids privacy leakage, achieves low-latency response, and operates without network dependency.
Product Specifications & Features
The AWA89601 supports single/dual-MIC noise reduction, echo cancellation, dual-MIC enhancement, far-field pickup, local voiceprint recognition, and local special sound recognition.
Applicable Scenarios
Suitable for AIoT devices, security equipment, smart home devices, AI robots, AI toys, and other scenarios.
Product Block Diagram
Image 2: Product Block Diagram
Technical Challenges/Pain Points vs. Solutions
Table 1: Technical Pain Points and Awinic's Solutions
沪ICP备10011954号 |
沪公网安备 31011202011955号
EN