Ultrasonic Visualizer

Real-time ultrasonic spectrum display using your microphone. Captures audio via PortAudio (sounddevice), computes an FFT, and plots only the ultrasonic band (>20 kHz). Includes a control panel to adjust device, sample rate, block size, min frequency, and FFT size while running.

Features

• Live FFT with Hann window and dBFS scaling
• Adjustable min frequency, FFT size (2^11–2^18), and block size
• Device selection by index or name; change sample rate on the fly
• Pause/resume; info banner with Δf, Nyquist, latency

Requirements

• Python 3.11+
• numpy, matplotlib, sounddevice
• PortAudio runtime (Linux): sudo apt-get install -y libportaudio2
• GUI backend (recommended): sudo apt-get install -y python3-tk

Quick Start

1. Create a venv and install deps:

   python3 -m venv .venv && source .venv/bin/activate pip install -U pip pip install -e .

2. Run it:

   list audio input devices

   python ultrasonic_visualizer.py --list-devices

   launch with specific device and samplerate

   python ultrasonic_visualizer.py --device "hw:0,0" --samplerate 96000

   or after installation: ultrasonic-visualizer --device "hw:0,0" --samplerate 96000

GUI Controls

• Min Freq (kHz): lower bound of displayed spectrum
• FFT Size (2^n): frequency resolution (Δf = SR / N)
• Blocksize: audio callback buffer size (latency vs. update rate)
• Device, SR (Hz): change input and sample rate; click Apply Device
• List Devs: print available device indices/names to the terminal
• Pause: freeze/unfreeze the plot

Troubleshooting

• PortAudio not found: sudo apt-get install -y libportaudio2
• No GUI window: install python3-tk or use X/Wayland forwarding
• Roll-off around 20–24 kHz: likely your input path is 48 kHz (Nyquist ~24 kHz) or hardware/mic low-pass filtering

License

MIT License — see LICENSE.