Description

Reefwing Noise Suppression Carrier Board

This carrier board hosts the real-time audio noise suppression project from Chapter 10 of Embedded AI: Run at the Edge (No Starch Press, 2026). The project runs a recurrent neural network on the Raspberry Pi Pico 2 (RP2350) to suppress background noise in a live audio stream, with a real-time spectrogram rendered to an attached LCD.

Running neural network inference on a live audio pipeline is one of the more demanding embedded AI workloads: the RNN must keep up with the incoming PDM microphone stream while the second core drives the display. This board removes the breadboard wiring from that problem so you can focus on the signal chain and the model.

What the board does:

Provides a socketed footprint for the Raspberry Pi Pico or Pico 2
On-board TDK InvenSense MMICT3902 MEMS microphone for audio capture
Header and mounting for the DFRobot DFR0664 2.0" TFT LCD (ST7789 driver) used for the live spectrogram
User pushbutton, power slide switch, and status LED
4-pin expansion header

Code

Embedded AI Book

https://github.com/Reefwing-Software/Embedded-AI

Schematic and Layout

Schematic

Schematic_Noise-Suppression.pdf

Download Gerbers Download BOM(Bill of materials)

Aug 27,2026

37 views

Reefwing Noise Suppression Carrier Board

Carrier board for real-time neural network audio noise suppression on the Raspberry Pi Pico 2, from the book Embedded AI (No Starch Press).

Published: Aug 27,2026

Standard PCB

Download Gerber file 0

BOM(Bill of materials)

PCBWay Donate 10% cost To Author

Only PCB

PCB+Assembly

Add to cart

Project images are for reference only. Actual production is based on the manufacturing files on the project page. Please review the designer's notes (e.g., PCB thickness) and select the appropriate options. PCBWay is not responsible for issues caused by unsuitable parameter selections.
For more important ordering information, please refer to Read More

Copy this HTML into your page to embed a link to order this shared project

Copy