Final Outcome
Credits:
epic game launcher fab(envir), alorema.gumroad(Mego V2), Turbo Squid, Sketchfab
Special Thanks:
George, Serra, Friendred, Shihong
The Making of Video
Final Fix
Compositing
Rendering
Lighting & Camera Change
Adjustments
Animation: Spline
Animation: Blocking
I began working on the animation and quickly realized that the process required significantly more time than I had initially anticipated. Nonetheless, here is Shot 2.
Special thanks to George for his valuable feedback on this shot. One of the major revisions involved the awkward animation of the clown making the balloon in front of the audience. Initially, I had planned to cover the clown’s hands with smoke during the post-production compositing phase. However, George suggested making the clown face away from the camera, effectively concealing the issue and reducing the need for additional adjustments later on.
This revision proved to be quite effective, and I also took the opportunity to enhance the little girl’s facial expressions with additional detail. That concludes the work on this shot, there is still much to catch up on.
For the ballon dog shot, apart from recording myself for reference, I have also did some research online about how cartoon characters fall in order to make the shot entertaining and interesting.
While creating the blocking for the animation, I realized that animating the balloon character was more challenging than expected, despite its lack of complex anatomical features. This difficulty stemmed from the need to convey emotions and narrative solely through body movements, without relying on facial expressions.
To ensure the emotional intent was effectively communicated, I continuously sought feedback from peers, asking how the animation felt to them and adjusting accordingly.
Rigging
As the model I selected for the main character, Loopy the dog balloon, did not include a pre-existing rig, I was required to rig it manually. This process presented several challenges. Firstly, the initial rig was overly simplistic and lacked the necessary functionality for more nuanced movements such as head tilting. Furthermore, the rig was missing key control systems for individual body parts. For instance, there was no master control to synchronize the movement of the ears with the head, which made the animation process significantly more time-consuming, as each ear had to be animated separately.
In addition, I encountered difficulties with weight painting, particularly related to the rotational control of one of the ears. These issues underscored the importance of beginning the animation process as early as possible. Despite performing initial checks on the rig, many problems only became apparent during actual animation. This experience highlighted that certain technical limitations are difficult to identify prior to practical use.
Test Render
I conducted preliminary tests of several camera shots from my previsualization using the models and character assets. While the intended atmosphere was successfully conveyed, the lighting proved suboptimal, and certain texturing issues became apparent, particularly on the little girl’s face and in the rendering of the grass, which appeared unnaturally white. Due to current time constraints, these elements will be addressed and refined at a later stage of production.
Setting up
In order to allocate the majority of available time to the animation process, I elected to forego the modeling stage and instead utilize pre-existing rigs and environments sourced from online repositories. My initial attempt involved the use of Fab (Unreal Engine) and Truong’s resources. While I was able to locate several professional-quality rigs that met my technical and aesthetic requirements, their acquisition proved cost-prohibitive, as they were available only through subscription-based licensing models that required recurring monthly payments.
I identified a suitable asset from Fab and subsequently integrated it into a new Unreal Engine project file. To evaluate its potential within the production pipeline, I conducted a series of test renders, experimenting with multiple camera angles to determine those most consistent with the visual composition and narrative intent established in my previsualization (previs) stage.
However, several technical challenges were encountered during the process. One notable issue involved the inability to successfully import FBX rigs containing both textures and animations from Autodesk Maya into Unreal Engine (UE). To address this, I experimented with multiple troubleshooting approaches, including exporting the assets via Alembic cache, upgrading to a more recent version of UE, and separately importing an FBX file containing only the texture data into UE. Despite these attempts, the problem persisted, necessitating further investigation into compatibility and asset pipeline optimization.
Ultimately, the only viable solution was to re-rig the model entirely within Maya prior to export. It is plausible that the original rig contained underlying technical inconsistencies such as non-standard naming conventions, residual construction history, or other rigging irregularities that contributed to the incompatibility issues experienced during the import process into Unreal Engine.
I also conducted a series of tests to evaluate the textures, reflection quality, and other rendering attributes. These issues required resolution prior to commencing animation, as addressing them in post-production could result in an excessive and potentially overwhelming cleanup process at the final stage of the pipeline.
Previs
This previs was originally developed during Term 2. No subsequent modifications were made, as it effectively conveyed the key narrative elements I intended to present. Moving forward, I will refine this previs by replacing the characters with updated references and proceeding with blocking and spline stages. The updated version will be stored in the same project folder.
Storyboard
The narrative recounts an incident in which a clown, while attempting to hand a balloon to a child, inadvertently released it. The balloon drifted into a hazardous location, namely, a busy street filled with moving vehicles. Remarkably, it neither burst beneath the weight of the cars nor succumbed to the sharp spines of a nearby cactus. Instead, it was eventually retrieved and “adopted” by an individual.
The underlying moral emphasizes that even under adverse or perilous circumstances, outcomes can take an unexpectedly positive turn. This lesson also serves as a broader metaphor for the importance of animal adoption. Many animals face challenging and distressing conditions beyond their control, and adopting them offers a chance for safety, care, and a renewed quality of life.
Ideation
This narrative originated from an experience I had one evening while returning home. As I walked, I observed a solitary balloon adrift in the night air, carried intermittently by the wind. It drifted between the roadway, where vehicles passed continuously, and the pedestrian pathway, moving back and forth in an almost rhythmic motion. The account that follows is a creative interpretation of the scene I witnessed that evening.
I remained stationary for some time, observing the scene. The striking contrast in color and texture, combined with the sense of anxiety I experienced while witnessing the event, served as the primary source of inspiration for this story.
Subsequently, I began sourcing reference images to create mood boards, enabling a more comprehensive understanding of the intended color palette. In addition, I developed character biographies for the principal figures in the narrative.