AON1010 is designed as a highly optimized neural network core specifically for voice and audio recognition. It processes microphone data efficiently, suitable for applications ranging from voice recognition to sound recognition. Delivered in Verilog RTL, the AON1010 is suitable for incorporation into ASIC and FPGA designs, making it adaptable to varied technologies.
This IP supports always-on environments with features like multi-wake-word detection and adaptive voice activity detection, which are key for continuous listening applications. Additionally, it provides high accuracy even in noisy environments, and is speaker-independent, meaning it can adapt to user and environmental variabilities without loss of performance. This adaptability allows it to handle multiple commands simultaneously and with precision.
With notable capabilities in speaker identification and context awareness, the AON1010 is optimized for integration into high-precision voice control systems. Its proprietary processing engine ensures substantial power savings, critical for battery-operated devices and systems requiring persistent voice interaction functionality.
