Sound wave simulation

This project provides a ready-to-use simulator of acoustic wave propagation in a field including sound obstacles, based on cellular automata. Lightweight and fast, it complements Alexander3's implementation and relies on the work of T. Komatsuzaki et al..

Installation

The current project is implemented in python 3.7 and requires the following dependencies:

• click: for creating a command line interface
• joblib: for CPU parallel computing
• scikit-image: for generating random obstacles in the field

pip install click joblib scikit-image

Getting started

First, let's start by generating a map with random obstacles. We prepared a random shape generator in the module utils.

Note that the simulator works with any other binary matrix you provide.

import matplotlib.pyplot as plt
from simulator import generate_random_map

# Generate random map of obstacle
# Note that you can use any other binary map you want
map_size = (100, 100)
random_seed = 1
obstacle_map = generate_random_map(map_size, random_seed=random_seed)

# Plot generated map
plt.pcolormesh(obstacle_map, cmap="gray")

Now, we run the simulator for 200 time steps and compute the resulting sound pressure level map.

from simulator import SoundSimulator

# Simulation
duration = 200
simulation = SoundSimulator(map_size=map_size, 
                            obstacle_map=obstacle_map, 
                            duration= duration)
simulation.run()
spl = simulation.spl(integration_interval=100)

Finally, we display the obtained pressure field:

# Display the sound pressure level field
fig, ax= plt.subplots(nrows=1, ncols=1, figsize=(8,4), sharey=True)
spl_plt = ax.pcolormesh(spl, cmap="coolwarm")
plt.colorbar(spl_plt, ax=ax)
ax.set_title("Sound Pressure Level (dB)")
fig.tight_layout()

How to use the CLI

• To run 10 random simulations lasting each 200 iterations and saving the results in the current directory, run the following line:

python -m src.simulate -n 10 -d 200 -p .

• The result of each simulation is stored separately in a file named example_x.pickle. One can load and visualize the results by running src.load_n_viz.

python -m src.load_n_viz

• For additional details about the arguments, a help page can be displayed:

python -m src.simulate --help