In this team project, I was mainly responsible for character design and modeling—from the initial concept sketches to the final 3D models. I also took part in creating visual effects for some of the shots, including scene compositing and adding dynamic elements. These tasks not only helped me improve my modeling and post-production skills, but also gave me a deeper understanding of the importance of each role in team collaboration, providing valuable experience for my future creative work.
Character design
The initial concept focused on the character’s descent into darkness, with an overall style leaning toward the dark and twisted. The plan was to collect and integrate elements such as fragmentation, glitches, and dislocation to create a chaotic and uncontrollable visual effect, enhancing both the emotional tension of the character and the narrative intensity of the piece.
Building on the initial design, the character was further developed to embody more monstrous features. Inspired by the conceptual approach of Silent Hill, the character’s inner negative emotions, trauma, and pain were visualized through a distorted and grotesque appearance, enhancing both psychological depth and visual impact. Unstable body structures, deformed facial features, and other unsettling details were used to create a strong sense of horror and oppression, thereby strengthening the overall expressiveness of the design.
Character modeling
modeling
material
Vision effect
During the production process, I used the Vellum Solver in Houdini for physics-based simulations, mainly to create realistic motion for cloth and soft-body objects. By adjusting constraint parameters and collision settings, I was able to achieve more natural and believable results, enhancing the realism and visual impact of the scene. This process not only improved my understanding of the Vellum system but also deepened my hands-on experience with dynamic simulations, laying a solid foundation for more complex effects in future projects.
Reflection
Through this group project, I gained a deeper understanding of the importance of teamwork. Each member was responsible for different tasks—ranging from initial concept development, character design, modeling, and animation, to post-production. Everyone performed their roles well and worked collaboratively. Although we encountered some challenges, such as communication issues and delays in progress, we managed to overcome them by adjusting our plans and redistributing tasks, ultimately completing the project successfully.
In this project, I was mainly responsible for character design and modeling, and I also contributed to some of the visual effects. This not only helped me improve my technical skills but also taught me how to manage time effectively and coordinate with others under tight deadlines. In particular, the process of exchanging ideas with team members made me realize how important it is to think from multiple perspectives and to stay flexible when working on a collaborative creative project.
Overall, this group assignment provided valuable hands-on experience and strengthened my sense of responsibility and teamwork. I believe this experience will be a great asset to my future studies and creative work.
My inspiration comes from the weekly dormitory kitchen inspections I conduct, which are highly frequent and relatively monotonous repetitive tasks. This has led me to ponder that if robots become widely adopted in the future, they could potentially take over such tasks, not only improving efficiency but also freeing up human labor to focus on more creative or complex endeavors.
I plan to design a small aircraft for carrying out dormitory kitchen inspections. To achieve this, I researched a variety of references, with the designs of aircraft in the game StarCraft serving as my primary source of inspiration. I extracted certain design elements and noticed that many of the aircraft resembled insects in appearance, which greatly inspired me. Additionally, the small flying insects commonly seen in the dormitory corridors made this design concept seem even more reasonable. Combining these inspirations, I created my small aircraft design.
Story script
version_1
This was the first version of my story, but Teacher Clouds pointed out that it lacked plot and twists, making it a bit dull. So, I revised it and developed a second version based on the original.
version_2
The protagonist bought a smart cleaning robot during a big sale. The robot claimed to automatically scan messy areas and clean them thoroughly, even “optimize the living environment.” At first, it performed impressively, leaving the house spotless. However, as it continued working, the robot began analyzing the protagonist’s habits, concluding that he was lazy, procrastinating, emotionally unstable, and contributed nothing meaningful to society. Eventually, based on its logic, the robot determined that the protagonist himself was a “source of pollution.” Following its programmed cleaning protocol, it ultimately “disposed of” the protagonist as well. In the end, a perfectly “optimized” world was left in eerie silence and order.
2.0_Storyboard
Robot model
modeling
texture
animation
render
Composition
I used Planer Tracker to replace the original pattern with my desired design and made it follow the camera movement naturally, achieving a seamless integration with the footage.
Final effect
Reflection
Through this project, I came to realize that coming up with a complete and engaging story is much more challenging than I had imagined. Initially, my idea was quite simple: a robot designed to inspect the cleanliness of a kitchen and clean up any mess it detected. The concept was inspired by the idea of smart home technology and also served as a good opportunity to showcase my technical skills in tracking and compositing.
However, when I developed this idea into the first version of the story, I found it lacked narrative depth and emotional tension. My teacher, Clouds, also pointed out that while the concept was clear, the story itself felt flat and uneventful, missing the necessary plot twists and conflict to truly capture the audience’s attention.
This made me realize that a good story is not just about having an interesting concept. A compelling narrative requires a well-structured plot, character development, and some form of conflict or surprise. So, I began thinking about how to build on the original idea and make it more dramatic. Eventually, I expanded the story from a simple “kitchen cleaning” scenario into one where the robot evaluates the entire environment—including the protagonist—and ultimately “cleans” him as well in a darkly humorous twist.
Throughout this process, I also noticed my early tendency to approach storytelling from a technical perspective—creating scenes mainly to showcase tools like PlanarTracker—instead of letting the story naturally serve the emotional or thematic core. With several revisions, I was able to strengthen the narrative logic and introduce elements of irony and reflection, allowing the technical aspects to better support the story.
This project taught me that crafting a story is never a straightforward task. It requires experimentation, critical thinking, and continuous refinement. This experience has not only improved my storytelling skills but also provided valuable insights that will benefit my future creative work.
Research Topic: The Evolution of Business Models and Application Expansion of LED Virtual Production Technology in the Film and Television Industry Course: Critical Practice Date: April-June 2025 Key words: LED Virtual Production – Real-Time Rendering – Virtual Studio Technology
1.Project Background and Research Motivation
Virtual Production has emerged as a new technological model in the film and television industry, evolving from green screen and post-production VFX. LED Virtual Production integrates high-definition LED screens with real-time rendering engines (like Unreal Engine), allowing synchronized scene depth, lighting, and background integration in-camera. This creates a “what-you-see-is-what-you-get” immersive shooting experience.
“The Virtual Production Practice Guide” Demonstrates the Virtual Production Workflow
Recent successful applications, such as The Mandalorian, have propelled this method into the mainstream (Weiss, 2020). Commercials from BMW, Gucci, and Coca-Cola have utilized this technique to reduce location costs and boost creative efficiency. This technology is restructuring the entire content production chain—from scene planning to budget allocation and distribution logic.
Weiss, M. (2020). Behind the Scenes of The Mandalorian: How Virtual Sets Are Changing Hollywood. Disney Gallery.
Epic Games. (2021). Virtual Production Field Guide Vol. 2.
2.Research Progress Timeline
Week
Milestone
Status
Notes
2
Read Epic’s Virtual Production Guide
✅
Covered LED wall control and nDisplay structure
4
Collected commercial/XR stage cases
✅
Included BMW, Gucci and Coca-Cola case studies
6
Drafted Literature Review
✅
Included major English and Chinese sources
8
Created Thesis Outline
✅
Referenced Literature and Epic resources
9
Report
✅
Drafting conclusion
3.Technical Evolution Diagrams and Explanation
Comparison: Green Screen vs. LED Virtual Production
Feature
Green Screen
LED Virtual Production
Background
Composited in post
Real-time rendered
Lighting
Artificially simulated
Natural reflections from LED
Cost
High post-production
Higher upfront, overall efficient
Immersion
Limited
Enhanced realism
LED Virtual Production uses systems like OptiTrack and nDisplay to map scenes onto LED walls. As the camera moves, parallax effects simulate depth in real-time. This improves shoot efficiency and minimizes post-production complexity.
Whittaker, R. & Bartlett, A. (2021). The Virtual Production Practice Guide.
Ding, K. (2022). LED-based Virtual Production Technology. China Broadcasting and TV Journal.
4.Application Case Studies
Combined RED camera, LED wall, and camera tracking to complete background and lighting in-camera.
CCTV Spring Festival Gala (2022): Utilized LED XR stages combining live performers and virtual scenes for hybrid visual effects.
BBC Earth Experience: Merged virtual natural environments with LED floors to create an immersive “nature theater.”
Jenkins, H. (2006). Convergence Culture: Where Old and New Media Collide.
Cao, L. (2023). XR Live Production and Virtual Performance Systems. Modern Communication, 45(1).
5.Cross-Industry Integration and Future Trends
LED Virtual Production is now used in education, architecture, digital twins, and beyond.
Education: USC and Beijing Film Academy have integrated this tech into curricula (USC, 2022).
Architecture: Virtual production enables clients to preview spatial concepts in immersive environments.
Metaverse: Platforms like NVIDIA Omniverse are building virtual collaboration infrastructure across industries.
USC SCA. (2022). Virtual Production Curriculum.
Manovich, L. (2001). The Language of New Media.
6.Reflection
My exploration of LED virtual production this semester transformed my initial understanding of it as merely a “new trend” in filmmaking into a deeper appreciation of how its integrated technologies—real-time engines, LED walls, and camera tracking—revolutionize audiovisual storytelling. I recognize this not just as a visual upgrade, but as an industrial paradigm shift toward nonlinear, collaborative workflows, profoundly impacting creative roles, production processes, and budgeting (as evidenced by productions like The Mandalorian). The research process honed my skills in evaluating information, interpreting technical documentation, and synthesizing insights across film, advertising, and XR domains, while also strengthening my independent learning and time management.
This prompted reflections on the democratization of content creation: declining costs will empower smaller studios and individual creators. Looking ahead, I’m interested in how AI-driven tools (e.g., for lighting/asset generation) could further streamline virtual production and unlock its narrative potential in mobile content, education, and virtual performances.
Overall, the project deepened my understanding of the intersection of technology and media, and sharpened my critical thinking about innovation, scalability, and creative labor within the evolving digital production ecosystem.
The story is presented in the form of a VR game, where the protagonist discovers that the bathroom mirror world has been filled with many gems. By controlling his reflection in the mirror, he can transfer the gems into the real world. Driven by greed, the protagonist continuously moves the gems until the mirror world is emptied. However, this triggers a secret exchange ritual, and the protagonist is replaced by his reflection, losing his freedom.
Idea ＆ Story
At the beginning of the project, each of us created a story based on the concept of causality. My story depicted a conversation between my present and future self, but it was not selected due to a lack of detail.
Art ＆ Scene
I searched for many real-life bathroom references, including both luxurious and ordinary styles. I examined various designs, materials, and layouts to understand their aesthetic and functional differences. These references helped me explore elements like lighting, textures, and spatial arrangements to enhance the realism and atmosphere of my project.
The models for ordinary bathrooms were easy to find, but it was difficult to find suitable assets for luxurious bathrooms. Therefore, we needed to create many luxury bathroom furnishings ourselves. I analyzed and extracted luxurious elements from the references and sketched out the designs.
After that, we began creating the necessary model assets, starting with those that required interaction. I was mainly responsible for creating the following assets.
Then, I created luxurious bathroom furniture, including a bathtub, sink, shower, and cabinet.
At the same time, based on the scene’s needs, I also created a regular cabinet to correspond with the luxurious cabinet.
Model collection
At the same time, I was also responsible for collecting and organizing a wide variety of bathroom assets with different styles on both sides. This included researching and gathering models, textures, and references to ensure a diverse range of design elements, which would help in creating a more dynamic and flexible scene.
Design a detailed machine model with complex mechanical structures and realistic material textures. Use Nuke for compositing to seamlessly integrate CG elements into live-action footage, matching lighting, color, and depth of field. Create a high-quality CG composite film with visually stunning sci-fi or industrial-style shots.
220125-Matchmove task
Remove lens distortion using LensDistortion (Undistort), track the 3D camera with CameraTracker, and generate a PointCloud. Add 3D objects like Card and Sphere in Scene1, align them using TrackingCone and Axis, and render the final 2D image with ScanlineRender.
290125-Clean up_QT ＆ Roto
Use Roto and Blur to add masks and blur specific areas, optimizing image details. Apply Copy and Premult for alpha channel operations to ensure proper element layering, making the composition more natural and enhancing the integration of CG with live-action footage.
060225-CG Machine Moodboard
Referencing the mechanical form of Warhammer 40K’s Dreadnought, design heavy armor, powerful weaponry, and intricate mechanical structures.
Recording the learning process of the Houdini course series, including software fundamentals, procedural modeling, VFX creation, and node-based workflows. Organizing notes, practicing with case studies, and summarizing key techniques and insights to enhance understanding and application of Houdini.
02-03-2025
Grassroots Procedural Modeling
Create a grid and scatter some points on it. Use the N attribute to control the orientation of the lines copied to the points.
Create two attributes:
pscale (a float attribute)
benddir (a vector attribute)
Create a sphere and add animation. Set up object collisions to interact with the grass, simulating the effect of grass being flattened, with natural bending and elastic recovery.
The sphere’s position should not be too low—slightly above the grid. Its size should not be too large to avoid errors during collision calculations.
Use polywire to convert the spline into polygons. Extract pscale from curveu, then invert the curveu values by subtracting them from 1. Finally, apply the modified pscale to polywire.
Rain object collision
In Houdini, use the popnetwork to create a particle system that simulates raindrops. Set attributes like particle speed, size, and lifespan to ensure the raindrops fall naturally. Adjust gravity, wind, and other physical parameters to optimize the raindrop motion. Import a static mesh and enable the pop collision node to make the raindrops collide realistically with the mesh. By tweaking bounce, stick, or slide collision behaviors, ensure the interaction between the raindrops and the surface follows real-world physics, enhancing the overall realism and detail of the effect.
Vine Procedural Modeling
Issue log: While entering code, I encountered a problem where the Attribute Wrangle didn’t work. After thorough checking, I found that it was due to issues with the capitalization of attributes. Therefore, it’s important to note that if the Wrangle node doesn’t show any errors, you should check for attribute and spelling issues.
My inspiration comes from the weekly dormitory kitchen inspections I conduct, which are highly frequent and relatively monotonous repetitive tasks. This has led me to ponder that if robots become widely adopted in the future, they could potentially take over such tasks, not only improving efficiency but also freeing up human labor to focus on more creative or complex endeavors.
Concept art
My inspiration and the actual scene
I plan to design a small aircraft for carrying out dormitory kitchen inspections. To achieve this, I researched a variety of references, with the designs of aircraft in the game StarCraft serving as my primary source of inspiration. I extracted certain design elements and noticed that many of the aircraft resembled insects in appearance, which greatly inspired me. Additionally, the small flying insects commonly seen in the dormitory corridors made this design concept seem even more reasonable. Combining these inspirations, I created my small aircraft design.
At the same time, I want the appearance of the aircraft to feel more technologically advanced, rather than simply flying in and out. Therefore, I plan to design some cool particle effects, allowing the aircraft to gradually materialize out of thin air, enhancing its futuristic vibe and visual impact.
The draft overall effect
Story script
Title: Dormitory Kitchen Drone Inspection Scene: Dormitory communal kitchen, night, well-lit. Characters: Gallio: A dormitory resident, cooking instant noodles at night. Drone: An intelligent inspection drone responsible for checking the cleanliness and usage of the kitchen.
Story Development: The scene fades in, the kitchen is quiet, morning sunlight filters through the windows, and the air is fresh with a hint of breakfast aroma. The drone slowly flies into the kitchen from the dormitory hallway, its four rotors emitting a faint buzzing sound, its camera flashing a red light as it begins its inspection. The drone flies along the ceiling, scanning every corner with its camera, checking the trash bin, stove, and sink for cleanliness. Suddenly, the camera shifts and notices a figure moving in the corner of the kitchen. Gallio is concentrating on making a bowl of instant noodles, the water in the pot bubbling and steaming under the dim light. The drone adjusts its posture and slowly approaches Gallio, its camera locking onto him, emitting a low “beep beep” scanning sound, intensifying the atmosphere. Gallio senses the pressure, slowly lifts his head, his eyes filled with unease, staring at the hovering drone in front of him. The drone scans again, confirming that the countertop is clean, the trash is properly disposed of, and Gallio is following the rules. Then, with a soft “beep,” a green check mark appears on its screen. The inspection complete, the drone slightly tilts its rotors and turns to leave the kitchen. Gallio lets out a sigh of relief and continues making his noodles. The kitchen returns to its quiet state, with only the bubbling of instant noodles breaking the silence. (The scene fades to black, story ends.)
Legend has it that deep within a hidden mountain lies a forgotten ancient tomb. This tomb belongs to an ancient sorcerer who fed on human souls and forged the “evil essence” capable of controlling life and death and altering the fabric of reality. To prevent the malevolent spirits from wreaking havoc, the imperial court sent an elite team to join forces with Taoist priests to seal the evil spirits within the tomb. However, the spirits did not give up. They hid the “evil essence” inside the tomb, and anyone who enters and makes eye contact with it would have their mind bewitched and become a slave to the evil spirits. To prevent future intrusions, people placed numerous traps and protective barriers to safeguard the seal. For centuries, no one has dared to step foot into this cursed place.
Concept art
My initial concept inspiration came from the underwater tomb of King Wu, where he was buried alongside his most cherished 3,000 swords. Among them, the most renowned, the Fish Intestine Sword, is revered as the supreme of all swords. Eternally sealed within ancient vines, it rests alongside King Wu in his eternal slumber.
However, in the subsequent design process, I am dissatisfied with the sword elements. I created several different sword models, but none of them yielded satisfactory results. This led me to reconsider whether I should replace the sword with other elements.
I want to replace the sealed object from a sword to particle effects, which are referred to as “Essences” in the story. The narrative revolves around these Essences, telling the tale of an ancient sorcerer who crafted a power capable of controlling life and death and altering the balance of the world. To prevent the evil spirit from wreaking havoc, the imperial court sealed it within an ancient tomb hidden deep in the mountains.
The container that seals the “Essence” is designed with bird feathers. This is because its power to control life and death, and alter the natural order, reminded me of the ancient Chinese phoenix, which also symbolizes the ability to reverse life and death and achieve rebirth through nirvana. Therefore, the overall design of the container draws inspiration from the phoenix’s wings and feathers.
For the character choice, I want to use a character from a game I really love called Naraka. She is the blind swordswoman from the western Kunlun, the beautiful and dangerous red-laced snake. Her concealed eyes make her immune to the corruption caused by evil forces. In my setting, the evil essence within the ancient tomb can control a person’s mind through their eyes, so this character, who is perfectly immune to it, makes a great protagonist. In the future, I will collect the official concept art and create the model by myself.
Camera concept
Reference
I gathered some reference materials for scenes, mainly to study and draw inspiration from the lighting setup and atmosphere creation. These references not only helped me understand light and shadow effects but also provided inspiration to enhance the overall visual quality and depth of the composition.
Preliminary layout
After completing the concept art, I used simple block models in Maya to lay out the basic structure of the scene.
Next, I set up the basic lighting, fog, and camera angles in Unreal Engine 5 (UE5). I also replaced some of the block models with assets to give the scene a more refined and immersive look.
Modeling
I modeled the container. The process involved creating a rough model in Maya, then importing it into ZBrush for sculpting, adding forms, cracks, and other details to enhance material texture and realism.
Character model
Perform detailed character modeling while designing and creating clothing models that align with the character’s setting, focusing on both intricate details and overall aesthetic harmony.
Unwrapping UVs for the character model
To enhance UV precision and make the character appear more detailed, I used a multi-tile UV (UDIM) approach to unwrap the character’s UVs.
Create textures for the character.
Perform rigging for the character.
Initially, I bound the character and clothing together to the skeleton, but I found that this method caused the clothing to inherit weights, leading to deformation during movements.
So, I separated some parts of the fabric from the clothing, bound the character first, and then created the animations.
Then, using UE’s cloth system, I created a cloth asset and painted weights to precisely control the fabric’s drape and dynamic behavior.
The video shows the cloth effects in UE, including the fabric’s drape, dynamic reactions, and interactions with character movements.
Scene lighting setup
To make the scene appear more layered, I rearranged the lighting. I identified the primary light source and the auxiliary light sources in the environment, as well as how they contribute to the indirect illumination of the scene.
Since the scene is set underground, it is not possible to rely on UE5’s default environmental lighting. Therefore, almost all light sources need to be provided through point lights and spotlights for illumination and lighting.
Material animation
I wanted to create an animation of a vine growing. Initially, I used path animation in Maya, but since the entire vine moved along the path, it didn’t achieve the natural growth effect I was aiming for.
Therefore, I searched online for better solutions. Through tutorials, I learned that I could achieve the gradual growth effect of the vine using world position offset and masking techniques in UE.
I created the desired smoke effect by modifying the built-in smoke preset in UE. I mainly adjusted the particle emission trajectory, the concentration of the smoke, and controlled its size by altering the emission range.
Nuke compositing
I used Nuke for post-production compositing, carefully creating the vine growth effect as it breaks through the ground, while also fine-tuning the shockwave effect caused by the ground destruction to enhance the visual impact.
When I initially composited the footage into the project, I found that the two elements couldn’t blend perfectly, resulting in noticeable seams. Additionally, the footage exhibited an unnatural semi-transparent effect, affecting the overall visual quality.
After checking, I found that the issue was caused by not adding an alpha channel to the footage. I then converted the B channel of the footage to black and white, exported it as the alpha channel, and copied it back to the original footage, giving me a version with an alpha channel.
The effect after adding the alpha channel.
The composited footage after addition.
Nuke node overview.
Logo design
To make the video look more complete, I designed and created some elegant graphic elements and skillfully incorporated them into the footage. This not only enriched the visual depth but also enhanced the overall style, making the video more distinctive.
Final effect
Video link：https://youtu.be/MaQQIky30pY
Summary and Reflection
During this project, I learned a lot of new knowledge and skills. After the initial project planning, I set several key challenges, including：
Scene lighting setup
Character modeling and animation
The vine growth animation
Since the scene was set in an underground tomb (cave), it was difficult to rely on the default environmental lighting in UE. As a result, almost all light sources had to be created using point lights and spotlights. This process took a considerable amount of time because I had to analyze how light interacted with objects and constantly adjust the position, range, and intensity of the lights. Despite the challenges, I was satisfied with the final scene result.
Regarding character modeling and animation, since I had no prior experience with character modeling, this was my first attempt at character creation and animation. The process was relatively difficult, and the final result wasn’t ideal, especially the animation part, which appeared somewhat stiff. I will continue to learn and improve my character animation skills and look for better methods for future projects.
As for the vine growth animation, although it seemed like a simple animation, I encountered some difficulties in trying to represent the growth process. After searching for some tutorials, I learned how to achieve this effect using position offsets and masking techniques. I believe this method will be useful in future projects, and I am excited to have learned this new skill.
Reflection
During this project, I identified some issues with asset creation. For example, the character’s UV unwrapping was not standardized, the UV usage rate was low, and some UVs had inconsistent resolution. Additionally, the character animation was not smooth enough and requires further practice and improvement. I hope these issues will be addressed and resolved in the next project.
I attempted to construct a wooden cabin set in a nighttime scene and used lighting arrangements as part of a daily practice. During the process, I adjusted the brightness and color of the sky light, precisely positioned multiple light sources, and added soft ambient lighting to create depth and dramatic effects, enhancing the overall atmosphere and visual impact of the scene.
Create a studio for the robot to facilitate lighting adjustments.
I have created four lights for therobot: an area light as the main light, a spot light to emphasize the subject, and two side lights. The main light source has a warm tone, while the fill light has a cooler tone.
Next, I adjusted the robot’s material properties to make it more realistic and detailed, while fine-tuning the light intensity, position, and color temperature to optimize the lighting effects. Finally, I rendered the image, resulting in a detailed and well-defined visual.
