Leading mixed-signal IC company Awinic has launched the high-performance AWA89601 NPU-powered intelligent voice processing chip. Integrating a dedicated audio NPU, the chip combines sound model training with deep hardware integration to achieve breakthroughs in AI noise reduction, voice enhancement, far-field pickup, and voiceprint recognition, delivering precise “face-to-face clarity” interaction for smart devices.
As a highly integrated semiconductor solution, the AWA89601’s AI deep noise reduction engine excels as an “environmental noise crusher.” It supports deep learning-based noise cancellation and complex environmental noise processing, effectively filtering wind, keyboard, traffic, and other common interferences. For core voice processing, the chip features far-field pickup, precise voice separation, and dual-MIC beamforming, significantly improving long-distance speech recognition accuracy. Built-in Acoustic Echo Cancellation (AEC) eliminates speaker echo for clear call quality, while Voiceprint Recognition (VPR) provides biometric authentication for secure interactions.
The chip’s key innovation lies in its edge-side NPU architecture. The integrated dedicated audio NPU delivers powerful on-device inference capabilities, enabling real-time processing of four uplink voice channels. Compared to traditional cloud-based solutions, local edge-side processing inherently avoids privacy risks associated with voice data transmission, ensures low-latency response, and maintains stable performance even without network connectivity—effectively addressing the cloud model's three main pain points: network dependency, high latency, and privacy concerns.
In terms of specifications, the AWA89601 comprehensively supports single/dual-MIC noise reduction, echo cancellation, dual-MIC enhancement, far-field pickup, on-device voiceprint recognition, and special sound detection. It is tailored for diverse applications including AIoT devices, security systems, smart home products, AI robots, and AI toys, providing customized voice processing for smart terminals across industries. Benchmark tests show significant improvements in key metrics such as noise reduction in complex environments, far-field pickup accuracy, and voiceprint recognition precision, setting a new standard for edge-side AI voice processing.
The launch of the AWA89601 marks a significant breakthrough for Awinic in edge-side AI audio chips, advancing technological upgrades and application innovation in AI voice processing within the semiconductor industry and accelerating the transformation of smart device interaction experiences.
ICgoodFind : Awinic’s edge-side NPU voice chip addresses key interaction challenges, enabling innovation in smart devices. We are committed to helping the industry access high-quality audio chip resources and seize opportunities in interaction upgrades.
