We're almost there! Time is moving very quickly and it's time to speed things along. Admittedly, this week was a bit rushed for me, as I've been scrambling to get things done quickly, so forgive the shortness of this post 🥹
Main Pyro
First thing to address is the main pyro sim. My velocity is far too high right now and is causing the fire and smoke to blow up dramatically. It's also way to fast. So, I turned down the velocity on my custom velocity field, as well as reduced the radius of the source torus.
Much better. Now though, there is a distinct donut shape that can be seen above the turbine as the velocity field kicks in. So, to fix this, I did two things:
Add curl noise to the velocity field. I already had a bit of this in there, so I just increased the amplitude and decreased the frequency to get more. Unfortunately, the curl noise in attribute VOP does not let you animate the noise, so I added an additional point velocity node set to "keep input" and added a tiny amount of animated curl noise through that. The reason I didn't use just the point velocity curl noise is because I didn't like it's resulting shape as much. When I turned up the amplitude to the level I needed the circular shape was getting destroyed, and I couldn't quite save it. So, I used point velocity curl noise at a low amplitude to simply add a little animation to the VOP curl noise. This is probably the worst way to go about things though, so I do plan to play more with the settings to find a more efficient solution.
I keyed the velocity field so that it's not so intense in the beginning of the simulation. I also masked out the velocity from my initial source around the velocity field area so that the two don't interfere with one another. I keyed the intensity of the mask to match the strength of the velocity field accordingly.
Fireballs
The biggest issue on my two fireballs was stepping, as well as their odd shaping (a teardrop like shape, tiny tip with smoke billowing out behind it). A suggestion from my professor to solve this was to face my velocity backwards more. Previously, I had my velocity pointing more radially, so as to get a noisier and more ball-shaped trail (which clearly wasn't too successful). I also increased the magnitude of the velocity. This change helped a lot, although, I had lost quite a bit of the fireball shape, now my trails looked more like a rocket trail. I had to split the fireball element into a smoke and a fire pass as well, to get more control of shading in Solaris.
Pyro Trail
I applied similar knowledge from my fireballs to the smaller pyro trails as well, which helped a lot to fix their shaping from week 11. I also increased the offset on when the trails begin inside the pyro trail path node, so that they don't all burst out at the same time (Although, in hindsight, this didn't make as much a difference as I would have liked).
Unfortunately though, since these trails are moving much faster than my fireballs, I once again encountered the stepping problem. I was having a lot of difficulty troubleshooting this though. I genuinely couldn't figure out why stepping persisted despite going through each node to try and figure out the issue. Unfortunately, due to time, I'll have to save this for the next (and final!) week.
Initial Pyro
The only change here was that I went back to the previous simulation before last week's. When I tried resimming this element by sourcing from the new model, I was once again encountering the old blobby source issue. However, since the shape of the helicopter on this particular area barely changes between the proxy and high-res models, going back to the old simulation was no harm at all. (Unlike with the main pyro, for which the shape of the cockpit has changed quite a bit).
Fixing the Rock Column
Recall from week 11 that this particular island/column was not ...responding well to the volume VOP noises. There was just too much mass to cover and it was resulting in my PC freezing up. So, to resolve this, after the voronoi fractures, I deleted all the pieces that could not be seen on camera. Then, I was able to go in and properly texture the rocks using volume VOPs again. Same process as previously, though now things were much faster with the reduced size. I still couldn't get as much detail as I would have liked for an element so relatively close to camera, but I couldn't increase my voxel size much more due to hardware limitations. I suppose I may have to rely a bit on textures to complete the look.
Texturing the Samson
To end the week, I brought the Samson model into Adobe Substance Painter to texture. Since there was no time to manually unwrap the model, I made use of Substance's auto-UVing tool. I also made use of UDIMs, which allows you to split a model's UVs across multiple texture maps, allowing for much higher texture resolution. I used a grayish metal as my base then used Substance's smart textures and masks to add levels of wear and tear. I also painted on the blue and white patterns that Trudy has on her particular Samson Helicopter in the movie.
The Final Render
Unfortunately, due to time, I wasn't able to render out the updated island/columns, nor the textures on the Samson. But anyways, here the final for this week:
Two glaring things I've noticed: 1) I've added way to much curl noise to my velocity field, and its shape can barely be seen now, and 2) the shaders on the small pyro trails put far to much density on the smoke. I think I had increased the density in the pyro bake node to better see how the smoke was looking, but forgot to turn it back down afterwards.
For Next Week:
Next week will be my final write up for this project, where I will explain my setup for each element in the shot, including any fixes that will be made between now and then. See you there!