Seeed Studio reSpeaker 2-Mics Pi HAT V2.0 for Raspberry Pi

Seeed Studio ReSpeaker 2Mics Pi HAT (V2.0) is a dualmicrophone HAT for Raspberry Pi , purposebuilt for AI and voice applications . It enables responsive, farfield voice interaction and integrates smoothly with mainstream platforms such as Amazon Alexa Voice Service and Google Assistant .

Note: ReSpeaker 2Mics Pi HAT has been upgraded to V2.0. The codec changed from WM8960 to TLV320AIC3104, adding Raspberry Pi 5 support and expanding the sampling rate range from 896 kHz.

Key features

Voice user interface for Raspberry Pi

Designed for seamless voice interaction on Raspberry Pi. The HAT integrates 2 analog microphones and supports VAD (Voice Activity Detection), DOA (Direction of Arrival), and KWS (Keyword Spotting) via software algorithms running on the Pi.

TLV320AIC3104 audio codec

A lowpower, highfidelity stereo codec optimized for portable digital audio applications, delivering clean, accurate voice capture and robust input gain control.

Controllable peripherals

Includes 3 programmable APA102 RGB LEDs and 1 user button for visual feedback and interaction flows in voice applications.

Easy assembly

Snap it onto your Raspberry Pi with the 40pin headerno soldering required. Ideal for developers, makers, and classroom use.

Smart audio expansion

The compact HAT leverages the TLV320AIC3104 codec and dual analog mics for highdefinition voice capture. With NLU and noise reduction algorithms on the Pi, it supports:

• Voice Activity Detection (VAD)
• Direction of Arrival (DOA)
• Keyword Spotting (KWS)

Widely used in smart speakers, voice assistants, and AI voice interaction systems.

How ReSpeaker 2Mics Pi HAT works

Two cornermounted analog microphones capture sound and feed the onboard audio codec, which features:

• Synchronized ADCs (AnalogtoDigital Converters)
• Mic boost amplifier for enhanced clarity

The codec outputs multichannel audio to Raspberry Pi via:

• Standard I2S or PCM over I2S ports
• TDM for synchronized multichannel transmission

Intelligent voice processing on Raspberry Pi

After capture, NLU algorithms on the Pi process the audio to enable:

• VAD detect speech segments
• DOA localize the speaker
• KWS recognize wake words and triggers

Ideal for voicecontrolled applications including remote control cars, voice UIs, wakeword systems, and smart home devices.

To improve input quality, ReSpeaker supports Acoustic Echo Cancellation (AEC), Beamforming, WebRTC Noise Suppression (NS), and Automatic Gain Control (AGC), delivering cleaner audio streams to Raspberry Pi.

Specification

Hardware Overview

Application

• Voice interaction learning
• Secondary software development
• Voicebased remote control cars

Get started with Google Assistant

How to use button to trigger Google Assistant

Documents

Respeaker_2_Mics_Pi_HAT_SCHRespeaker_2_Mics_Pi_HAT_PCBRespeaker_2_Mics_Pi_HAT_SCHRespeaker_2_Mics_Pi_HAT_PCBReSpeaker 2 Mics Pi HAT 3DSeeed-Voice DriverAlgorithms includes DOA, VAD, NSVoice Engine project, provides building blocks to create voice enabled objectsAECReSpeaker2-MicsPiHAT_V2.0_SCH&PCB

Technical details

Part list