ESP32-LyraTD-MSC is based on a digital signal processor (DSP) with a three-microphone array (for noise reduction, echo cancelation, voice recognition, near- and far-field voice wakeup, etc.) and ESP32-WROVER, which is an Espressif module combining Wi-Fi and dual-mode (Classic and Low Energy) Bluetooth connectivity. The new audio development board achieves acoustic echo cancelation (AEC), while also being expandable with such peripheral devices as I2C, I2S and UART. ESP32-LyraTD-MSC consists of two development boards: the upper board with function keys and LEDs and the lower board which includes modules and chips.
- Wireless Module: ESP32-WROVER with 802.11 b/g/n, dual-core 240MHz CPU and 4.5 Mbyte SRAM
- Wi-Fi and Bluetooth (Classic & Low Energy) connectivity.
- Digital Signal Processing chip
- Micro-SD card for storing audio files
- Audio driver chip
- Earphone jack
- Dual speaker output ports
- Three-microphone array supporting up to three-meter sensitivity when playing music
- I2C/SPI header
- 6-pin UART header
- I2S header
- Debugging: USB-UART micro-USB interface and JTAG header
- Power Supply: 5V via micro USB port
The ESP32 chip supports wakeup phrases such as “Alexa” and “Espressif”. Say “Alexa” to the board, and a response in Chinese will be played (“您好!有 么吩咐”: nin hao you shen me fen fu, meaning “Hi! How can I help?”). If the board cannot access DuerOS, Baidu’s conversational AI system, the board will support the voice wakeup function only. When ESP32-LyraTD-MSC is connected to DuerOS, users can talk to the speaker and instruct it to play songs, news, etc.