Participative media and volumetric rendering using transient_prbvolpath

Overview

In this tutorial, you will modify the scene from 0-render_cbox_diffuse.ipynb so that the cubes in the Cornell Box have participating media inside (i.e. cubes filled with smoke).

This tutorial assumes that you’ve read 0-render_cbox_diffuse.ipynb. Most of the setup/visualization code is the same. Here you get to see some cool specular reflections in the rendered video though :)

🚀 You will learn how to:

• Edit the scene to add participating media, and render using the transient_prbvolpath integrator

• Visualize the edited scene

# If you have compiled Mitsuba 3 yourself, you will need to specify the path
# to the compilation folder
# import sys
# sys.path.insert(0, '<mitsuba-path>/mitsuba3/build/python')
import mitsuba as mi
# To set a variant, you need to have set it in the mitsuba.conf file
# https://mitsuba.readthedocs.io/en/latest/src/key_topics/variants.html
mi.set_variant('llvm_ad_rgb')

import mitransient as mitr

print('Using mitsuba version:', mi.__version__)
print('Using mitransient version:', mitr.__version__)

Using mitsuba version: 3.6.4
Using mitransient version: 1.2.0

We use the short alias mitr for mitransient for improved code readibility.

Editing the Cornell Box scene

In this case we provide a cbox_volumetric.xml that we will load with mi.load_file. The necessary changes are:

<shape type="obj" id="smallbox">
  <string name="filename" value="meshes/cbox_smallbox.obj"/>
-   <ref id="gray"/>
+   <bsdf type="null" />
+   <medium type="homogeneous" name="interior">
+     <rgb name="albedo" value="0.99, 0.4, 0.46"/>
+     <float name="sigma_t" value="0.05"/>
+     <phase type="hg">
+       <float name="g" value="0.1"/>
+     </phase>
+   </medium>
</shape>

In summary:

1. We change the “gray” BSDF with a “null” BSDF. This means that, when a ray hits the box it will go inside and not reflect as it would do with a gray diffuse surface.

2. We add a homogeneous medium. For the details of all the variables and configuration you can check Mitsuba’s documentation. The text above shows the changes for smallbox which has red albedo. Note that the largebox also has a different medium with blue albedo.

# Load XML file
# You can also use mi.load_dict and pass a Python dict object
# but it is probably much easier for your work to use XML files
import os
scene = mi.load_file(os.path.abspath('cornell-box/cbox_volumetric.xml'))

Render and visualization

This uses the same methods as in the previous notebook.

data_steady, data_transient = mi.render(scene, spp=1024)
print(data_steady.shape, data_transient.shape)

(256, 256, 3) (256, 256, 400, 3)

Note that this function will take longer to execute, as it’s doing the rendering job (lazy evaluation)

# Plot the computed steady image
mi.util.convert_to_bitmap(data_steady)

# Plot the computed transient image as a video
data_transient_tonemapped = mitr.vis.tonemap_transient(data_transient)

mitr.vis.show_video(
    data_transient_tonemapped,
    axis_video=2,
)

# Plot some frames of the computed transient image
import matplotlib.pyplot as plt

data_transient_tonemapped[data_transient_tonemapped > 1] = 1

plt.subplot(1, 2, 1)
plt.axis("off")
plt.imshow(data_transient_tonemapped[:, :, 72])
plt.subplot(1, 2, 2)
plt.axis("off")
plt.imshow(data_transient_tonemapped[:, :, 120])
plt.show()