In recent years, with the rise of artificial intelligence and the Internet of Things, voice has gradually become a typical way for us to interact with smart devices. Embedding AI voice on connected devices (such as smart home devices, smart office devices, smart wearable devices, etc.) has also become a hot topic. Espressif Information Technology (stock code: 688018) recently released the ESP32-Korvo AI voice microphone array development board.
This is an AI voice development board for IoT embedded devices. It is equipped with a single ESP32 chip and supports 2 wheat / 3 wheat arrays. It is suitable for far-field, low-power speech recognition applications / scenarios, such as smart Smart voice display, smart voice switch. At the same time, in conjunction with front-end acoustic algorithms such as Mic-Array Speech Enhancement (MASE) and Acoustic echo cancellation (AEC), it can effectively implement noise reduction for far-field speech and improve speech wake-up and recognition performance.
ESP32-Korvo consists of two development boards, integrates multiple peripherals, can be plugged in headphones, speakers, SD card, flexible selection of single wheat, double wheat, three wheat, with LED lights and function buttons, forming a variety of AI voice applications .
Motherboard: contains ESP32-WROVER-B module (16 MB flash, 8 MB PSRAM), power module, micro SD card slot, headphone and speaker interface;
daughter board: contains microphone array, function buttons and LED lights.
ESP32-Korvo can be used with Espressif Intelligent Voice Assistant (AI Voice Recognition SDK)-ESP-Skainet to achieve rich application development. Based on the voice wake-up engine WakeNet, users can realize the device's voice wake-up function, and can also customize the wake-up words. Based on the command word recognition network MultiNet, users can easily set rich local command words, add their own voice commands, and realize local voice control of smart devices.
Software Guide:
Go to ESP-Skainet software preparation to view the following steps:
- Obtain ESP-IDF and configure C language PC development environment based on ESP32.
- Get ESP-Skainet and run Espressif intelligent voice assistant. Use ESP-SR under ESP-Skainet to query and call the APIs required by ESP-Skainet, including wake word recognition, voice command word recognition, and front-end acoustic algorithms.
- Compile, program, and run the ESP-Skainet example .
Design reference:
