Week 1 – 29/09/25
“CGI IS FOR LOSERS!”
Christopher waltz in a panel for Frankenstein replied to a question from the audience that “CGI is for losers”
Why did he say this?
I think there could be a few reasons for Christopher Walz saying “CGI is for losers”. My thought on Waltz’s take is that CGI is commonly perceived to be something that isn’t good looking, a cheap way of making movies. I think that CGI is misunderstood and falsely given a bad reputation.
Why I think CGI is misunderstood?
A lot of films being released are sequels or superhero films. Blockbusters that often have supernatural aspects. These films use CGI to show things that wouldn’t usually be possible on set. The CGI used in these films aren’t always good looking due to short deadlines made by corporate decisions so that the films can come out earlier. Also the scenes that are usually made by CGI are obviously not real. The audience knows that the effects aren’t real resulting in realism being taken away from the scenes.
I think CGI is often linked to these superhero films. With their spectacles and fight scenes. CGI is much more complex than just making cool fights, CGI is a great tool. There are many examples of the tool where audiences aren’t even aware of its presence. Because audiences aren’t aware of good CGI people often think of the bad and obvious moments in films, such as these blockbuster/superhero films. The result of audiences linking CGI to superhero movie spectacles could make the tool feel cheap, adding a false belief to the craft of CGI
Investigative study Module project
Topics that interest me in visual effects that I could talk about:
- VFX jobs (how hard it is to get jobs in the industry)
- Digitalising actors
- visual effects is falsely perceived
- Practical effects vs digital effects
- The future of AI in visual effects
My title?
How will Ai affect Visual effects?
Research:
AI and ML technologies are already beginning to transform the VFX industry. Here are some of the areas where they are currently being applied:
One of the biggest advantages of AI in VFX is automation. Which reduces the need for repetitive and time-consuming manual labor. Runway ML covers some of these common visual effects tasks:
-
Rotoscoping: Traditionally, rotoscoping, the process of manually isolating objects or characters from a background, is labor-intensive. AI-powered software can now perform rotoscoping with remarkable accuracy and speed, using deep learning to track and separate subjects from moving backgrounds. Tools like Adobe After Effects‘ Roto Brush 2 or Runway ML’s AI models are examples of this application.
-
Object Tracking: AI can significantly enhance motion tracking, where it can analyze footage and apply digital elements to moving objects in real-time. This includes enhancing or replacing elements in live-action shots, such as adding explosions or environmental effects like smoke and fire.
-
Facial Motion Capture and Deepfake Technology: AI is enabling highly accurate facial capture and performance replication. Deepfake can create hyper realistic digital avatars or perform face-swapping, as seen in films like The Irishman (de-aging of Robert De Niro).
These tasks performed by ai will cut costs, create efficiency and better results.
Runway | Tools for human imagination.
Polemic: The Camera Never Lies
Deepfake Queen: 2020 Alternative Christmas Message
- Write a brief description – what is a deepfake?
Deep fakes is technology manipulating a video or a camera and placing someone else’s face over another persons face in a believable way. Deep fake can be done in real real time.
- How does it work? How is it believable?
Deep fake uses AI to create deceptive images/videos. The technology relies on two components. The generator and the discriminator. The generator creates the fake content that will go over the subjects face. While the discriminator decides if the information from the generator is real or fake.
- List some examples with images.
- List keywords associated with the topic.
identity theft, fake news, deep learning, AI, photorealistic
Is it okay to replace actors faces using deepfakes technology?
I wouldn’t say yes, however I wouldn’t say no either. I think it depends on the use of the content with the deepfake. Anything sexual, political or anything that would be misleading is a big NO. Along with taking a dead or live persons likeness without the families permission is also no. Anything light hearted and amateurish is fine. Content that is obviously fake.
Unreal: The VFX Revolution
Task: Listen to the podcast and write a quick annotation summarizing what the programme was about
Notes:
Paul Franklin is the narrator who has a history of working in visual effects.
George Lucas coming off his success in American Graffiti would help change the handheld camera aesthetic in Hollywood. Inspired to make a fantasy movie, George Lucas wanted to use new camera techniques and show off stunning visual effects. John Dykstra was hired to be head of the visual effects in star wars. Dykstra was focused on new camera effects that could create motion blur and push the boundaries of visual effects.
Dykstra’s Camera Effects: The Dykstraflex
What it did:
- Allowed for precise, repeatable camera movements – very important for compositing multiple elements such as backgrounds, crowds, models in a single convincing vfx shot.
- Effects at this time were limited to still or clunky movement to maintain a decent looking shot. The Dyxtraflex gave scenes fluid motion and creative freedom to the director in what kind of shot they wanted.
- The Dyxtraflex enabled the illusion of fast moving spacecrafts through multi layered shots. This was thought to be impossible at the time
Doug Trumbull started a company with paramount that would make technological developments in cinema. Steven Spielberg had just finished Jaws which was a huge success. He was given freedom for his next project and wanted to attempt a film in Sc-fi. He was interested in using visual effects for his next film idea. However he was told the technology wasn’t there yet. Although he was convinced by Doug Trumbull that they could use parts of the technology in visual effects for digital motion control of the cameras. Here they’d be able to composite scenes with a huge amount of people or include a UFO in a scene where the cameras were moving. Which was revolutionary at the time.
Robert Black was one of john Dykstra’s earliest recruits. He was known for his experimental film and optical effects making him a great candidate for using Dykstra’s motion control system. This would be later known as Compositing.
06/10/25 week 2
Ai was used in a photography competition. A German photographer thought it would start an interesting debate if he produced and won a photography competition with an Ai created image. He in fact did win and admitted to have used Ai to win, ultimately rejecting the competition.
What are the debates that this situation starts?
- Should anti-Ai steps be put in place for these creative competitions
- Is it considered a photo?
- Should Ai have its own category
- The photo still needed an artist to keep changing the photo
Is Ai a challenge or an advantage for visual effects?
I think its both. Definitely a tool that can enhance work by being more efficiency and more accurate. I think it could also risk replacing a lot of the industry.
My Investigative study
I’m still focused on the idea of talking about ai but I don’t know what angle I should take. I’ve got a few ideas on how I should tackle this subject.
- Trying to predict what the next 5 years or nearer future could look like
- What jobs could change or be at risk in VFX
Week 3 13/10/25
My Investigative study review with Florian
Going in to todays lesson I wasn’t too confident in sharing my research template on AI. My topics are broad and the things I mention don’t match well with one another. I explain how I want to talk about many topics within AI which I think made me feel lost about how I wanted to tell my essay.
When Florian reviewed my work we agreed that I should focus on a smaller section of AI. Specifically focusing on AI roto. This was one of the first things that came to my attention when researching VFX AI. Roto is a task that is tedious and not creative. I think this is the perfect conditions for AI to take over. There is no creative freedom for the Artificial Intelligence to take over.
I can also talk about how Junior roles work could change with Ai roto scope taking over. The junior compositors do this tedious work while learning comp and building their portfolio. What work would they do instead.
Florian helped give me the idea of using AI roto as my topic, as well as including my own primary research. 2 primary research examples and a secondary research.
Primary research 1
Have an AI tool use roto on an image that I have also roto scoped. Compare the time and how accurate the two pieces of work are.
Primary research 2
Ask industry experts thought out questions to do with AI roto work.
Secondary research
(can be anything about Ai roto work…)
Overall thoughts on the review
I feel a lot more confident with my essay now. I have a thought out plan with my own primary research to talk about. It’s something that I’m interested in and something that will most likely affect me in the VFX industry. I’m very pleased with the outcome of having my investigative study spreadsheet reviewed.
Re-do of my Investigative study – this time about AI Roto
week 4 20/10/25
AI+Roto-Joel’s_Presentation.pptx
new title: How ai roto affect the VFX industry
Week 5 27/10/25 – Bibliography
- An article on VFX ai Roto by Jonathan Lam. Explores big budget movie projects that have already had Ai work integrated in the making of the films. Since I am working in a very similar topic I think it would be a great idea to refer to this article and have a think about some of the points and research Jonathan refers to.
https://elements.envato.com/learn/ai-in-vfx?msockid=2bfd374d7fd863f63f0521c57eee62c8
- One of the three AI comp websites I would like to investigate and test. I want to perform my own studies on the imperfections, efficiency and quality of Ai comp technology at the moment.
- The second AI comp website I plan to use for my primary research.
Video editing made easy and fast with AI
- Runway ML
AI Visual Effects Tools | Runway AI
Week 6 Research
researching articles on Ai tracking and rotoscoping. I looked in to three different sources.
Source 1
A conference paper on Ai tracking dancers, preserving data and providing a new way of teaching dance moves. Using Ai to track multiple dancers doing the same dance and storing that data to provide a 3D visualisation of that dance. I read a bit of the paper and I don’t think it is quite what I am looking for. I did like the paragraph explaining the data processing part:
“In the preprocessing phase, the collected videos were processed to extract frames and key points of the dancers’ bodies. Computer vision algorithms, such as OpenPose and MediaPipe, were used to identify and track the dancers’ skeletal structures throughout the video sequences. This step involved segmenting the video frames and accurately mapping the positions of joints, limbs, and facial features.“
I like this extract from the paper as it show the use of Ai collecting data from real life movements which links in to rotoscoping tracking. I am looking for something that is more technical and focuses on Ai more. However It was nice to read the implications of Ai in another art form, this being dance.
Source 2
A research article on Ai videos. I started reading the introduction and I found it interesting how the the author (David Ledo) compared the rise of Ai videos to early remixes of songs by disk Jockeys and calling the new Ai artists clip Jockeys. Just thought I would note it down. I do think this is interesting how history has repeated itself with these trends except in another art form.
The paper goes over generative rotoscoping. This is taking real footage, rotoscoping parts of the video and layering ai generated frames over the existing original frames.
Not exactly the ai rotoscoping I am after.
Source 3
Artificial animation
The paper goes over the effects of Ai in the visual effects industry. It does this with semi structured interviews from nine different visual effects artists. The research investigates the VFX artists perception on Ai. They are asked how they think it will help and what limitations they think Ai will have.
1: AI in creative and technical applications
The applicants acknowledged AI’s usefulness in repetitive or technical tasks. Which I think could be useful when talking about my own paper on Ai rotoscoping. However the artists raised their concerns on AI have the creative judgement required for high level projects. Ultimately AI was recognised as a creative assistant. Not a replacement for a human.
2: Ethical and legal Considerations
The artists raised their concerns on intellectual property. Something that the artists at Framestore also brought up when I went to visit with my class from the university of West London. They feared AI generated content might create unwanted legal battles with copyright infringement. Therefor Compromising projects. The unclear ownership of this content makes AI something to be distanced from.
3: Challenges and limitations
The artists highlighted potential challenges in integrating AI in to their current workflows. While intrigued by the idea of using new tools such as Sora they doubted the productions readiness. They also expressed how an overreliance on AI could cause a decline in creativeness and reduce opportunities for skilled artists.
Their Conclusion
AI holds promise for being an assistant tool to artists, not a replacement. Meaningful Integration of the tool will require clarifying over the ethics of the tool such as content ownership. Also a change in the industries workflow would have to be made to allow AI to make a bigger impact. AI is something to be intrigued and concerned about.
It’s interesting reading the artists thoughts on the technology but the brief mentions of rotoscoping and tasks that require repetitiveness is what made me keep reading as it relates to the type of research I am trying to find. Majority of the talk on rotoscoping was on the sub tittle 1Creative and Technical Applications.
One of the most significant benefits highlighted was the ability of AI tools to streamline and automate repetitive tasks, thereby improving overall workflow efficiency. Cameron discussed how AI is utilized for mundane tasks such as denoising and procedural generation, which saves time and resources stating: ‘In my day-to-day it’s more about using it for mundane tasks like denoising or some procedural generation stuff’, which, according to scholars, seems to be a standard approach adopted by users who perceive the value of AI in addressing administrative or cumbersome uncreative tasks ([ 2]; [ 3]; [35]; [41]; [47]). Similarly, the use of AI for rotoscoping tasks was noted as a substantial time-saver but it still required some manual clean-up. ‘We tried using CopyCat [AI tool] for that and it did speed things up. But it wasn’t perfect. We still had to go in and clean up a lot’.
This part of the paper directly links to my research. The VFX artists bring up the potential of using AI for under creative tasks. Tasks that are tedious, technical or repetitive.
Cameron notes that him and his team of artists tried integrating AI rotoscoping with their work. Despite the application not being perfect they still saved a lot of time. Since it has been over 3 years when this paper was published I would like to test the progression of AI rotoscoping in 2025. I want to see how far the technology has grown.
Week 7 10/11/25
Presentation week 11 and 12
Essay submission – week 13
Digital sketchbook – week 14
introduction –
Literature review –
Methodology –
Findings and discover –
Conclusion –
Method – An actual way of which you collect data or research your data
methodology – Why conduct your research in a particular way? Write 300 500 words on how and why you are performing your research
Case – study when you examine an aspect of VFX for a case study, you will describe how it relates to your case study.
Thanks for installing Wondershare Filmora Video Editor for Windows
Week 10 01/12/25
Research with Katie, trying to find more secondary research that links to my investigation.