🎙️ Seeed Studio ReSpeaker 2‑Mics Pi HAT (V2.0) is a dual‑microphone HAT for Raspberry Pi 🍓, purpose‑built for AI 🤖 and voice applications 🗣️. It enables responsive, far‑field voice interaction 📡 and integrates smoothly with mainstream platforms such as Amazon Alexa Voice Service 🔊 and Google Assistant 💡.
⚠️ Note: ReSpeaker 2‑Mics 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 🎶 8–96 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 low‑power, high‑fidelity 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 40‑pin header—no soldering required. Ideal for developers, makers, and classroom use.
🔊 Smart audio expansion
The compact HAT leverages the TLV320AIC3104 codec and dual analog mics for high‑definition 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 2‑Mics Pi HAT works
Two corner‑mounted analog microphones capture sound and feed the onboard audio codec, which features:
- 🔄 Synchronized ADCs (Analog‑to‑Digital Converters)
- 🔊 Mic boost amplifier for enhanced clarity
The codec outputs multi‑channel audio to Raspberry Pi via:
- 📡 Standard I2S or PCM over I2S ports
- 🔗 TDM for synchronized multi‑channel 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 voice‑controlled applications including remote control cars, voice UIs, wake‑word 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
| Specification | Details |
| MIC | 2 analog microphones |
| LED | 3 APA102 RGB LEDs |
| Raspberry Pi headers | Supports Raspberry Pi 5, Raspberry Pi Zero, Raspberry Pi 1B+/2B/3B/3B+/4B |
| TLV320AIC3104 codec | Low power, high quality stereo codec for portable digital audio |
| Software algorithms | VAD (Voice Activity Detection), DOA (Direction of Arrival), KWS (Keyword Search) |
| User button | 1 programmable button |
| Voice capture distance | Up to ~3 m radius |
| Battery | Exclude |
| Dimensions | 65 mm × 30 mm × 15 mm |
| Weight | G.W 33 g |
🧩Hardware Overview
🔧 Application
- Voice interaction learning
- Secondary software development
- Voice‑based remote control cars
📚 Documents
🧾 Technical details
| Dimensions | 65 mm × 30 mm × 15 mm |
| Weight | G.W 33 g |
| Battery | Exclude |
📦 Part list
| ReSpeaker 2‑Mics Pi HAT V2.0 | x1 |