🎙️ ESP8266 Wi-Fi Microphone

Stream audio from an INMP441 I2S microphone over Wi-Fi using an ESP8266, then receive and transcribe it on your PC in real-time using Python.

---

happy to help at any time (i do not respond fast i have school)

Key Features

• UDP Audio Streaming — Low-latency audio packets sent from ESP8266 to PC over your local network
• Arduino OTA (Over-The-Air) Updates — Flash new firmware to the ESP8266 wirelessly; no USB cable needed after first flash
• 160 MHz CPU Overclock — The ESP8266 is set to run at 160 MHz (system_update_cpu_freq(160)) for smooth, uninterrupted audio capture and transmission
• Software Gain Boosting — Python-side gain amplifier (GAIN_FACTOR) compensates for the INMP441's quiet output; clipping is prevented automatically
• Speech-to-Text via Google — Received audio is automatically transcribed using Google's Speech Recognition API
• Mono 16-bit PCM, 16 kHz — Clean audio format compatible with most speech recognition engines

---

🗂️ Project Structure

esp/
├── esp8266_mic/
│   └── esp8266_mic.ino      # ESP8266 firmware (I2S capture + UDP send + OTA)
├── python_receiver/
│   ├── main.py              # PC-side receiver (UDP listen + gain boost + transcribe)
│   └── requirements.txt     # Python dependencies
└── README.md

---

🧰 Hardware Requirements

Component	Details
ESP8266	NodeMCU / Wemos D1 Mini
Microphone	INMP441 (I2S digital mic)

INMP441 → ESP8266 Wiring

INMP441 Pin	ESP8266 Pin
VDD	3.3V
GND	GND
WS (LRCK)	GPIO14 (D5)
SCK (BCLK)	GPIO13 (D7)
SD (Data)	GPIO12 (D6)
L/R	GND (for left channel)

---

⚙️ Step-by-Step Setup

Part 1 — Arduino IDE (ESP8266 Firmware)

1. Install the ESP8266 Board Package

1. Open Arduino IDE → File → Preferences
2. Paste this URL into Additional Boards Manager URLs:

   http://arduino.esp8266.com/stable/package_esp8266com_index.json
   

3. Go to Tools → Board → Boards Manager, search esp8266, and install it

2. Install Required Libraries

In Arduino IDE go to Sketch → Include Library → Manage Libraries and install:

• ArduinoOTA (usually bundled with the ESP8266 core)
• I2S (bundled with ESP8266 core — do not use the Arduino built-in I2S)

3. Configure Arduino IDE Board Settings

⚠️ CRITICAL — the 160 MHz setting is required for stable audio streaming.

Go to Tools and set the following:

Setting	Value
Board	NodeMCU 1.0 (ESP-12E Module) (or your board)
CPU Frequency	160 MHz ← must be 160 MHz
Upload Speed	921600
Flash Size	4MB (FS:2MB OTA:~1019KB)

4. Edit the Firmware Configuration

Open esp8266_mic/esp8266_mic.ino and update these lines:

const char* STASSID = "your-ap";         // ← Your Wi-Fi network name (SSID)
const char* STAPSK  = "your_password";   // ← Your Wi-Fi password
const IPAddress listener(192, 168, 1, XX); // ← Your PC's local IP address
const int port = 3333;                   // ← UDP port (keep default or change both sides)

💡 Find your PC's local IP by running ipconfig (Windows) or ip a (Linux) in a terminal.

5. Upload via USB (First Time)

• Connect your ESP8266 via USB
• Select the correct Port under Tools → Port
• Click Upload (→)
• Open Serial Monitor at 115200 baud — you should see Streaming Started...

6. Upload via OTA (After First Flash)

Once the ESP8266 is connected to Wi-Fi:

1. Under Tools → Port, a network port called esp8266-mic will appear
2. Select it and click Upload as usual — no USB needed!

---

Part 2 — Python Receiver (PC Side)

1. Find Your Python Version

Make sure Python 3.8+ is installed:

python --version

2. Create a Virtual Environment

✅ Always use a virtual environment to avoid dependency conflicts.

# Navigate to the python_receiver folder
cd esp/python_receiver

# Create the virtual environment
python -m venv venv

# Activate it
# On Windows:
venv\Scripts\activate

# On macOS/Linux:
source venv/bin/activate

3. Install Dependencies

pip install -r requirements.txt

4. Edit the Python Configuration

Open python_receiver/main.py and check/update these values:

UDP_IP = "0.0.0.0"   # Listen on all interfaces (keep this)
UDP_PORT = 3333       # Must match the port set in the .ino file
SAMPLE_RATE = 16000   # Must match i2s_set_rate() in the .ino (16000)
GAIN_FACTOR = 5.0     # Increase if audio is too quiet (try 3.0–10.0)

5. Allow Firewall Access

When you first run the script, Windows Firewall may prompt you — click "Allow" to permit UDP traffic on port 3333.

6. Run the Receiver

python main.py

You should see:

Listening with 5.0x gain on port 3333...

Speak near the INMP441 — transcribed text will print to the console every ~5 seconds.

---

###part2.2 ffplay

ffplay -fflags nobuffer -flags low_delay -probesize 32 -analyzeduration 0 -sync ext -nodisp -af "volume=5.0" -f s16le -sample_rate 16000 -ch_layout mono -i "udp://0.0.0.0:3333?fifo_size=0&overrun_nonfatal=1"

🔧 Tuning Tips

Problem	Fix
Audio is too quiet	Increase GAIN_FACTOR in main.py (e.g. 8.0)
Audio is distorted / clipping	Decrease GAIN_FACTOR (e.g. 3.0)
Packets dropping / stream stutters	Ensure ESP8266 CPU is set to 160 MHz in Arduino IDE
OTA port not showing	Make sure ESP8266 is on the same Wi-Fi network as your PC; restart Arduino IDE
Transcription fails	Check your internet connection (Google Speech API requires internet)

---

📦 Dependencies

See python_receiver/requirements.txt.

Package	Purpose
sounddevice	Audio device interface
numpy	Gain boost math (fast array operations)
SpeechRecognition	Google Speech-to-Text interface
PyAudio	Audio I/O backend required by SpeechRecognition
cffi	C Foreign Function Interface (sounddevice dependency)

---

📝 Notes

• The 160 MHz CPU frequency is set in firmware via system_update_cpu_freq(160) in setup(). You must also set it in Arduino IDE under Tools → CPU Frequency before uploading, or the setting may be overridden by the bootloader.
• Audio is captured as 16-bit signed integers, mono, 16 kHz — the left I2S channel only (the INMP441's L/R pin is tied to GND).
• The Python script buffers ~5 seconds of audio before transcribing to give the Google API enough context for accurate results.
• OTA hostname is esp8266-mic — you can ping it at esp8266-mic.local if mDNS is working on your network.

---

📄 License

MIT — free to use, modify, and distribute.