FMP – Concept Refinement
This week, there was a presentation in class, and for this, I developed a storyboard, though it wasn’t fully enough to build out the entire story.
Comments from the presentation session:
- Suggestion for the transition to the final scene: It was suggested that someone could steal the protagonist’s phone, prompting them to look up and notice the character on the 3D billboard. Alternatively, an effect could take the phone away instead of someone physically stealing it.
- Paradox in the message: The main message is to encourage people not to focus too much on their phone screens. However, if the protagonist shifts their attention from the phone screen to a large billboard screen, doesn’t this create a paradox? The attention just moves from one screen to another.
In the test filming process, I realized the video was too shaky, and stabilizing it using Nuke or After Effects might be difficult. To capture more stable footage, I’ll need to consider the camera setup. I also thought about building a 3D street in Unreal Engine, but I’m uncertain whether I’ll have enough time to complete it, and achieving a level of realism comparable to real footage could be challenging. When filming both the phone screen and the background simultaneously, I found it difficult to control focus and proportion. Additionally, the phone’s screen reflection was too strong, making it hard to display clearly. One potential solution could be to film the hand and phone separately using a green screen.
There was an issue activating Nuke, but I managed to solve it after the presentation and conducted some stabilization experiments. When comparing results, Nuke’s outcome was better than After Effects’. I made a comparison between the footage before and after stabilization—while the first scene looked too blurry, the second scene came out well. Excessively shaky footage becomes blurry when stabilized, so using a well-stabilized camera like a GoPro might be a better option. Considering the potential for scaling up during stabilization and cutting the edges, filming in a resolution higher than FHD can ensure good video quality after stabilization.
After the session, since the main point of focus was the final scene, I revisited it. The character’s design for the last scene was particularly challenging. I could go with a simplified character like I did in previous terms, or I could use a realistic human model created with Metahuman. I recalled an image of Queen Elizabeth I saw at the airport, which was composed of a large number of smaller pictures. This sparked an idea: what if I created a character made up of particles representing countless social media photos? However, this effect would require an overwhelming number of images, so I adjusted the concept to include geometric shapes, data, and text instead.
After refining the character’s appearance, I asked ChatGPT for a concept image, and the result was satisfying. I then searched for tutorials that could help me achieve this look, but I had trouble finding ones that closely matched the style I wanted. Ideally, I would execute this in After Effects, but if that’s not possible, C4D, Blender, or Houdini could be suitable alternatives.