The Final Animation: Saturn V Rocket

Storyboards:

Stills (A lot of my custom-made Saturn-V model because it didn't get shown off enough)

(Some models I didn't have time to finish/texture completely)

Description:

This project truly was the culmination of everything I've learned about Maya and 3D animation in general this semester. I wanted to go big with this one, but due to technical limitations and some mishaps I had to abort the moon landing. Despite this, I am still very pleased with my completed animation, and am very glad to have ended on this note. The first notable thing I did different on this project than the others was to create all my own textures, models, and assets. Every single texture/shader, other than the grass at the very beginning of the animation, was created by me. All models are also mine, and a huge part of this project for me was the modelling.

I started by committing a few days just to modeling the main Saturn V Rocket, due to it being the focal point of the animation I wanted to add as much detail as possible.

One issue with this approach that is evident when comparing the above photos with the final animation is that of scope and scale. Focusing on one specific aspect (modeling) and ignoring the other necessary developments for a good animation led me to do a lot more work than I really needed to. The above lunar lander, for example, was not featured in the final animation due to render time limitations. However, I had a ton of fun designing the entire main model, and enjoyed trying to make it as detailed and realistic as possible. Everything was achieved with basic geometry, as I didn't want to have to deal with curves. 

The next step was designing the environment, which I did with two planes. The first was textured with the one externally-sourced shader (grass), and the other I used to make a water surface.

The water is just a flat plane textured with bump and displacement nodes connected to a noise function. The land was formed using the sculpting tools, I ran into some issues with the geometry rendering correctly with the Arnold renderer that limited some of the shots. In addition, ensuring that the edge of the terrain could not be seen in any of the final shots was an interesting task, I often hid it behind terrain or shot scenes at downward or zoomed-in angles. 

The launch itself featured the most annoying part of this whole process for me, which was the particle systems. From what I can tell, Maya has at least three different particle/fluid systems that can be utilized. These are the legacy fluids system, nParticles, and bifrost. Each one had its separate set of issues that caused me to spend multiple days attempting to figure out and develop a good render, and even still fall a bit short in the end.

I started with nParticles, which seemed simple enough to create good visuals, and rendered relatively quickly. However, during the period of developing the initial particle system for the rocket exhaust visuals, I found myself being slowed down by constant computer crashes. When attempting to render long frame sequences and shots, the render would fail and the program would crash consistently, and due to the nature of nParticles, any sequence would need to be rendered from frame one of the particle simulation to look right. Due to this, I found myself repeating extremely lengthy renders, wasting hours of potential time. I also tried caching the simulations via nCache, but doing so would consistently crash Maya with no way of prevention. This led me to explore bifrost, which is the newest and somewhat experimental particle simulation software added to Maya, but at this point I simply did not have enough time to create a visually-appealing exhaust effect, so I had to give it up.

This all led me to the system I used in the final product, which I was very hesitant about due to some serious drawbacks. The Maya legacy fluid/particle system was one I had previous experience with from another project, and because of this I was already exposed to its quirks. The reason I didn't want to use it was due to the inability to render legacy particle system effects on the GPU. Up until this point, I had used GPU rendering to produce my sequences much quicker than I otherwise could. Using legacy particles would mean having to utilize CPU Arnold rendering, which would drastically slow down render times. At this point, the deadline was approaching and I wasn't sure I'd be able to render everything in time. Still, I committed to using the legacy particles, and since I had prior experience with the system I was able to quickly make some passable exhaust effects. Of course, I was limited on time now, so they definitely could have been better if I had more time to develop and re-render sequences, but since each CPU render took around one to two minutes a frame, I had time for basically one take per sequence. 

The much longer render times than anticipated did mean I had to cut down on the scope of the project. I decided to feature just one stage separation and some shots of space in the second scene, and end the animation with that. Definitely a bit of a bummer that I couldn't fully show off the intricacies of my models, but I am still happy with how the final sequences in the second scene came out. I did not include any particle systems in the second scene, so I was able to take full advantage of the speedier GPU rendering. As for engine exhaust, I used basic poly cone geometry with an opacity and color ramp to simulate a fiery engine plume. Everything was filmed far away, so the effect worked and its simplicity was mostly unnoticeable. As for the planet, I sourced the clouds and terrain from satellite imagery on NASA's website, and added a larger sphere with color/opacity ramps to simulate an atmosphere. The final sequence (right after the stage separation) is also one I have problems with but couldn't re-render due to time constraints, in the end it turned into more of a timelapse-esque shot, which kind of works but wasn't what I was originally going for. Either way it wraps up the video somewhat nicely.

I of course do wish I had managed to get more time to render and develop a third scene of the landing itself, as well as some other parts like the docking of the lander and the crew modules. Still, I am very proud with what I produced, and with the limitations of the Maya software, as well as the limited time (I had around four days due to my other finals), I am happy it turned out as well as it did. While sometimes really annoying and tedious, I really enjoyed the process of developing this animation, and having full creative freedom was definitely helpful as well. Thanks for allowing me the opportunity to create something like this for a class, and teaching me about Computer Animation!!