Embedded Files
To the left, you can see the final image generated to demonstrate the volumetric rendering that I implemented.
Some main features:
- The smoke is homogeneous, but bounded. The cloud only obscures the turret and part of the ship, not the shells in front of it.
- The cloud casts a faint shadow (more obvious in the test images than here).
- The Henyey-Greenstein scattering phase function was used here, but it would be relatively simple to add in Rayleigh or Mie scattering phase functions.
This was one of the first images I generated with volumetric scattering. Things of note:
- There's a dark border around the volume. This is due to the entry and exit points being too close together, causing only the entry to be detected. This was latter fixed by having the Evaluate method check if the passed point was inside the bounding box of the volume.
- There's a number of black dots on the ground beneath the volume. These are due to a NaN value generated by a typo in one of the equations in the scattering phase function.
You can clearly see in this image that the dark border and black dots have been eliminated. However, the green box doesn't shadow the volume, and the shadow cast by the volume is far too sharp. Unfortunately, neither of these issues have been fixed.
This is a test of the emission property, using a color temperature of around 2250 K with an intensity of 30. It doesn't seem to be illuminating anything, which I originally though was an error...
...So I cranked up the intensity to 3000. As you can see, the volume does illuminate the surrounding area (albeit noisily) for sufficiently high intensities. The issue at lower intensities is that objects do not specifically check to see if they are illuminated by the volume. As a result, illumination is only spread by rays which happen to intersect the volume.
Report abuse