Voice Logger with Real-Time Playback and Compression

This tool continuously records audio input, displays real-time volume levels, and plays back speech segments after pauses. It saves all audio to .wav files and logs each playback chunk in .jsonl. Optionally, finished .wav files are compressed to .mp3 using ffmpeg if it is available.

---

🔧 Features

• 🔊 Continuous audio recording
• 🎛 Real-time volume meter and adjustable threshold
• ⏸ Automatic detection of pauses of adjustable duration
• ↺ Playback after each detected pause
• 📏 Saves all audio (including silence)
• 🗒 One-line JSON log per playback (JSONL format)
• 📁 Hourly file splitting with timestamped filenames
• 🎼 Optional compression to high-quality MP3 with ffmpeg

---

▶ How to Use

1. Install Dependencies

1. (Optional) Install FFmpeg and add it to your system PATH for .mp3 compression.
2. Run the installation script:

install.bat

• If a virtual environment (venv/) already exists, the script will ask whether to reuse it or delete and recreate it.
• All required Python packages will be installed into the virtual environment.

2. Run the Logger

run.bat

Speak naturally. When you pause for a bit, the last spoken phrase will be played back automatically.

---

🎹 Controls

Key	Action
↑ / ↓	Increase / Decrease volume threshold
← / →	Increase / Decrease pause duration
n	Start a new recording file immediately
space/Enter	Stop playback early
q	Quit the program

---

📁 Output

All session data is saved to the recordings/ folder:

• .wav file (raw audio)
• .jsonl file (playback chunk metadata)
• .mp3 file (optional, if ffmpeg is available)

Example log entry in .jsonl:

{
  "fs": 16000,
  "chunk_duration": 0.1,
  "start_chunk": 103,
  "end_chunk": 156,
  "start_time": 1718546724.13,
  "end_time": 1718546729.72,
  "threshold": 0.2,
  "pause_duration": 1.5
}

---

✅ Requirements

• Windows OS
• Python 3.8+
• Microphone input
• (Optional) ffmpeg in PATH for .mp3 compression

---

📌 Notes

• The tool saves everything, including silence.
• Playback happens only between detected speech-pause regions.
• If ffmpeg is missing, the .wav file will still be saved — compression will simply be skipped.

---

📄 License

MIT License. Use it, tweak it, share it.

---

Enjoy your real-time voice playback! 🎙️