Winter Wonderland

Over the weekend, I created the first version of the snow flurry element.

Before I started to create the snow, I decided to research the behavior and look so I was aware of how it's supposed to look. After a couple of searches, I found this article from the National Snow and Ice Data center that gave me a good overview. I also searched around for reference and found the video below.

Reference Video

After finding and reading about the general properties of snow, I then looked into what snow might look like in the location we're mimicking. We chose the Donner Pass tunnels near Lake Tahoe, CA as a location to mimic. I found a video of snowfall in Truckee, CA which is very close to the tunnels. 

After collecting all of the relevant reference, I started looked around for tutorials and blog posts related to creating snow flurries in Houdini. I found this great video by Tyler Bay (CG Forge) on how to create the flurries. I used his technique as a starting point for the flurry effect.

I also found this awesome article by Kate Xagoraris (Procedural Kate)about simulating snow in Houdini. She provided a lot of great resources in her post. I would highly recommend taking a look at her whole site.

I then proceeded with creating the effect. I decided to keep my camera static for this test so I could evaluate and dial in the motion of the snow flakes. In version 16 (the test above), I instanced larger flakes and used a higher air resistance to create flakes that fell to the ground slower. I looked back at my primary reference and saw that the flakes were moving faster than this test. 

In this iteration, I lessened the amount of air resistance and instanced smaller flakes. This creates an effect the feels faster and closer to my reference.

After showing these iteration side-by-side, it became apparent that the quantity of particles in version 17 felt lacking. The reduced size of the snowflakes also threw people off a bit.

Some comments from peers included:
"The first one feels more fluffy, but the second one feels more real"
"I feel like there should be more snowflakes in the second one"

Professor Fowler recommended that I use the flake size from version 16 in combination with the air resistance from version 17. I will be following that feedback in combination with any notes provided.

I'm also planning on making adding larger curl noises to the POP sim to add pockets of gusty wind. 

For next review session, I want to integrate velocity from the train's motion. This should cause some nice turbulence that breaks up the vertical motion of my current flakes.