with Nigel
This week’s materials focused on progressing towards the thesis proposal, and I continued reflecting on the feedback received during the tutorial. Since I had started reconsidering my original direction, I spent more time researching possible alternative topics and comparing the availability of academic sources related to each area.
I also reviewed the uploaded guidance documents and thought about which kinds of topics would allow for deeper analysis and more flexibility in research. During this process, I continued gathering references and reading around several themes that interested me, while gradually narrowing down potential directions.
with Ting
This week I revisited my previous lip sync dialogue shot and continued refining it based on the feedback I received earlier. I refined several facial poses to improve the clarity and readability of the expressions, making the performance feel stronger and more intentional overall. After adjusting the facial performance, I also cleaned up and clarified the lip sync to create a better connection between the dialogue and the character’s acting.
During this process, I realised that my initial approach to lip sync was too literal. At first, I was keying a distinct mouth shape for almost every individual sound and letter. However, after studying my reference more carefully, I noticed that natural speech does not require every phoneme to be fully posed. Trying to animate every sound separately made the lip sync feel overly busy and visually cluttered. Instead, simplifying certain transitions and focusing only on the most important mouth shapes made the performance appear much clearer and more natural. This observation helped me better understand the importance of selectivity and readability in facial animation.
Alongside this, I officially began blocking for my advanced animation shot. Since this assignment is entirely self-directed, working on a shot that I personally planned and designed felt especially motivating and enjoyable. I started building the main poses and timing for the pantomime performance while experimenting with more exaggerated and cartoony acting choices. Additionally, this shot required the use of technical elements such as parenting and object interaction, which gave me the opportunity to further practice an area that I previously felt less confident in. It also helped me better understand how to use the camera more effectively to support the performance and staging of the shot. Applying these technical setups within an actual performance-based animation helped me better understand how they contribute to creating cleaner and more controllable workflows. Compared to previous assignments, this process felt much more creatively open, allowing me to focus more on personality, appeal, and entertainment within the animation.
with Serra
This week focused on experimenting with RealityScan and photogrammetry workflows. We explored the process of scanning real-world objects and environments and converting them into digital assets that could later be integrated into Unreal Engine. The process was more challenging than some of the previous workshops, particularly when trying to achieve accurate and clean scans, but it was still very interesting to observe how physical objects could be translated into digital environments. Seeing the results gradually improve throughout the process made the session engaging and rewarding.
Alongside the practical workshop, we shifted our focus from conceptual research to technical experimentation. Our first objective was to establish a reliable motion-tracking workflow using MediaPipe with a webcam input. We also tested connecting an iPhone camera through OBS Studio, allowing us to achieve a higher-quality live feed for tracking and providing greater flexibility for future development.
Once the tracking system was functioning successfully, we began deciding which movement parameters would best communicate our concept. We analysed the available tracking data and selected the body points that produced the most precise and noticeable motion, ensuring that the visual output would respond clearly to user interaction.
Using these selected parameters, we created our first line graph visualisations, translating body movement into real-time data. This marked an important step in the project, as we started transforming human motion into simplified numerical representations, directly supporting our concept of AI reducing complex human behaviour into measurable data. The technical tests also gave us a stronger foundation for developing the following stages of the interactive experience.
We also put together a project statement and submitted.