Support instruction set and FPU floating point operation unit accelerator: supports up to 1024 points of complex /IFFT operations or 2048 points of real /IFFT operations Customized voice algorithm operator Storage: Support 1 analog Mic input Support dual-channel output Support 12S input/output Power supply and clock: Built-in 5V to 3.3V, 3.3V to 1.2V LDO power supply for the chip RC 12MHz clock source and PLL phase-locked loop clock source Set POR (Power on Reset), low voltage detection and watchdog Peripheral interface: Support up to 10 GPIO All GPIO can be configured as external interrupt input and call source 1 standard SPI Master interface, the highest speed car is 30MHz 1 SPI Slave interface with a maximum rate of 30MHz 1 full-duplex UART maximum rate 3Mbps I I2C master/slave controller with a maximum speed of 400kHz 2 PWM outputs 1 12-bit SAR-ADC maximum 450Khz sampling rate Custom voice (provided by customers, customized by factory) Customized start-up announcement Custom voice words Custom voice words Custom TTS voice broadcast Support up to 10 -free words (no need to wake up, direct recognition of words) Colour: blue Material: plastic Size: 18 x 16mm Package Contents: 4 x Voice Control Module Only the above package content, other products are not included. Note: Light reflection and different displays may cause the color of the item in the picture a little different from the real thing. The measurement allowed error is +/- 1-3cm. The solution supports offline recognition of 150 local commands, freely customizable words, words and response broadcast words, and has a wealth of peripheral interfaces. HLK-V20 is the latest high-performance pure offline speech recognition module for a large number of pure offline control scenarios and products. It can be widely and quickly applied to smart homes smart small appliances, 86 boxes, toys, lamps, industrial, medical, Products that require voice control, such as the Internet of Things, automobiles, security and lighting. Processing :32bit RISC core, running frequency 240M The module uses a 32bit RSIC architecture core, and adds a instruction set specifically for signal processing and speech recognition, an FPU arithmetic unit that supports floating-point operations, and an accelerator. The audio signal is trained and learned through the neural network to improve signal Recognition ability.