Steps:
– Identify an aspect of visual effects for your research area.
– Write a short proposal with a focused research question.
– Draft the first chapter (either the literature review or the methodology).
– Design your research method.
– Do the practical work and/or reading (literary) research.
– Write the full essay with findings and conclusions.

Deep Fakes
A deepfake is a fake video or audio made with AI, often using machine learning that looks or sounds real.
Most deepfakes are face swaps, lip-sync edits, or voice clones.

The AI learns patterns from many images or audio clips of a person and then generates new frames or sound that copy that identity. Sometimes it looks strange (uncanny) and easy to spot, but the tech is getting better and can be very convincing.

Title: AI and the Future of Junior Compositing
Main Research Question: How are AI/ML tools reshaping entry‑level compositing roles in VFX, and what skills should candidates have to secure a junior compositor position (UK/EU/USA/CANADA)?

Sub‑questions:
– Which junior tasks are most affected?
– What portfolio evidence now matters?
– What skills now junior must have?
– Do studios reduce, re‑label, or upskill these roles?
– Maybe AI will bring artists more benefits than challenges

Claims:
1) Outsourcing was the first wave and AI is the second wave (and may follow a similar pattern)
– Idea: The first big change for junior comp was outsourcing of roto/paint/plate-prep to lower cost countries. AI now automates parts of the same work across UK/EU/Canada/US, so the effects on entry routes may repeat.
– Why it matters: It explains why there are two times less “gateway tasks”  first geographically and now technologically.

2) Skill shift: less “no brainer” tasks and more problem-solving, communication and decision making
– Idea: Juniors now need to show judgement (when/why a method), not only clicking tools. Soft skills + technical reasoning matter more.
– Why it matters: Hiring teams want juniors who can spot artefacts, explain choices, and take feedback.

Ideas:
Maybe interview Alexander Williams – Dean & Director of Animation & VFX at Escape Studios – He talks a lot about how AI benefits VFX artists

Resources:

1. Artificial imagination: Industry attitudes on the impact of AI on the visual effects process
Narayan, A. D., Caillard, Duncan, Matthews, Justin and Nairn, Angelique (2022), ‘Artificial imagination: Industry attitudes on the impact of AI on the visual effects process’, Interactions: Studies in Communication & Culture, Special Issue: ‘The Human and the Machine: AI in Creative Industries’, 13:2, pp. 113–31, https://doi.org/10.1386/iscc_00056_1

2. The Rise of CreAltives: Using AI to enable and speed up the creative process
Pearson, Andrew. (2023). The Rise of CreAltives: Using AI to enable and speed up the creative process. Journal of AI, Robotics & Workplace Automation. 2. 10.69554/WLDX9074.
https://www.researchgate.net/publication/372028377_The_Rise_of_CreAltives_Using_AI_to_enable_and_speed_up_the_creative_process

3.Creativity and technology in the age of AI
Pfeiffer, Andreas. (2018). Creativity and technology in the age of AI. 10.13140/RG.2.2.16400.76804.
https://www.researchgate.net/publication/338840672_Creativity_and_technology_in_the_age_of_AI

4. The Hollywood Jobs Most at Risk From AI
Cho, Winston (2024), ‘The Hollywood jobs most at risk from AI’, The Hollywood Reporter, 30 January, https://www.hollywoodreporter.com/business/busi-ness-news/ai-hollywood-workers-job-cuts-1235811009/. Accessed 23 April 2024.
https://www.hollywoodreporter.com/business/business-news/ai-hollywood-workers-job-cuts-1235811009/

5. Hollywood animation, VFX unions fight AI job cut threat
Asher-Schapiro, Avi (2024), ‘Hollywood animation, VFX union fight AI job cut threat’, Context, 9 April, https://www.context.news/ai/hollywood-anima-tion-vfx-unions-fight-ai-job-cut-threat. Accessed 23 April 2024.
khttps://www.context.news/ai/hollywood-animation-vfx-unions-fight-ai-job-cut-threat

6. The evolution of VFX-intensive filmmaking in 20th century Hollywood cinema: an historical overview
Venkatasawmy, Rama (2012). The evolution of VFX-intensive filmmaking in 20th century Hollywood cinema: an historical overview. Open Research Newcastle. Journal contribution. https://hdl.handle.net/1959.13/1047912
https://openresearch.newcastle.edu.au/articles/journal_contribution/The_evolution_of_VFX-intensive_filmmaking_in_20th_century_Hollywood_cinema_an_historical_overview/28981499?file=54351380

7. The AI Takeover In Cinema: How Movie Studios Use Artificial Intelligence
https://www.forbes.com/sites/neilsahota/2024/03/08/the-ai-takeover-in-cinema-how-movie-studios-use-artificial-intelligence/

8. Drama at Disney: A Thematic Analysis of Creative Worker Identity Negotiation and Identification in the Documentary Waking Sleeping Beauty
https://iafor.org/journal/iafor-journal-of-media-communication-and-film/volume-7-issue-1/article-1/

9. WILL AI REPLACE VFX ARTISTS?
https://escapestudiosvfx.com/2023/12/16/will-ai-replace-vfx-artists/

10. The Impact of AI on VFX: Transforming Nuke Learning, Compositing, and the Future of Film Making
https://www.nukezerotohero.com/post/the-impact-of-ai-on-vfx-transforming-nuke-learning-compositing-and-the-future-of-film-making

11. The Magic of AI in VFX: Enhancing Visual Effects Like Never Before
https://numalis.com/the-magic-of-ai-in-vfx-enhancing-visual-effects/

12. THE CONVERGENCE OF AI AND VFX: SPEED, CONTROL, AND THE FUTURE OF CREATIVE WORKFLOWS
https://escapestudiosvfx.com/2025/03/20/the-convergence-of-ai-and-vfx-speed-control-and-the-future-of-creative-workflows/

13. AI in VFX: The Future of Automated Compositing and Rotoscoping
https://arenaparkstreet.com/ai-in-vfx-the-future-of-automated-compositing-and-rotoscoping/

Week 2 – Developing a Research Area and is Generative AI the Future of Image-Making?

What was covered in the lecture:
Each student presented his idea of RQ from the previous week.

AI developments:
We were discussing AI–generated photographs. You can find a summary under the next section.

The Workshop Activity:
AI developments

A generated image winning a photography exhibition raises many questions. The “artist” even declined the prize, saying it was made with AI, but they still wanted to give it to him.
So should AI images and photography compete? Is it art? And who gets the credit? the prompt writer, the AI, or the software creators?

It also makes me ask: what is a photograph? Today, even a “real photo” can be heavily edited in Photoshop or Lightroom and it’s still called photography, while AI isn’t. Maybe the issue is that AI has no actual lens. Does that mean only traditional, lens-based work is “real” photography?

On the other hand, to get a good AI image you still need knowledge of photography such key concepts and  lighting principles. VFX also uses digital or even virtual cameras, so what’s different? In 3D software, like in real life, you must work with light and that’s the core of photography. Maybe that applies in 3D but not in pure image generation. Maybe AI is, in the end, a powerful calculator, and art isn’t something you can fully calculate.

Introduction: The Paradox of Photography
Today it’s very easy to take photos, but the easier it gets, the less we understand how it actually works. People use their phones, but the tech behind images is more complex, so it’s harder to see a unique photographer’s touch. Everyone documents everything in detail, which makes it feel like reality is captured better than ever however, is that really true?

The Weekly Activity:
VFX Research Template and Preparation
Potential area of interest – Investigative Study – V1 – Old Version
Potential area of interest – Investigative Study_V2 – New Version

Week 3 – Investigative Study template review and propossal

What was covered in the lecture:
This week we were presenting our ideas and got a feedback on it.
We started to write our proposals.

Week 4 – Student Presentations on individual Investigative Study Topics

Presentation: Link

Written Proposal – Due for 22.10:

Title: AI and the Future of Junior Compositing
Keywords: junior compositing; AI/ML; roto/paint; plate prep; portfolio evidence; hiring signals; UK/EU/USA/Canada.
This investigative study asks: How are AI/ML tools reshaping entry-level compositing roles in VFX, and what skills should candidates have to secure a junior compositor position (UK/EU/USA/Canada)?

The topic matters because many juniors used to enter through “gateway” tasks like roto, paint/cleanup, and DMPs, and these are exactly the areas now changing. In the near future, AI/ML are likely to automate or speed parts of this work inside studios, so there may be fewer simple “no-brainer” tasks and a stronger need for creativity, judgement, and communication.
I see two waves: outsourcing first (moving repetitive entry-level tasks to lower-cost regions) and AI as the second wave (automating parts of the same work inside UK/EU/USA/Canada). if the second follows the first, we can predict pressure on local entry routes and adapt training accordingly.

My aim is to map which junior tasks are most affected, show what studios and recruiters still value, and give clear entry recommendations.
My objectives are:
(1) list what tasks AI could realistically do next in compositing (roto/paint, cleanup, grain, depth/normal estimation, and DMPs).
(2) explain how the outsourcing pattern may repeat effects with AI on local entry routes and what we can learn from it.
(3) Identify the skills reviewers look for now. For example: build harmonious shots, understand colours, single shot storytelling, advanced methods where relevant (additive keys, deep compositing, smart vectors), Nuke particles when efficient, understand camera/lenses, create an efficient script and apply techniques in nuke that reduce render time in heavier software’s. Also, soft skills like problem-solving, communication, a learning mindset, and clear decision making.
(4) Define what portfolio evidence matters (this will be a focus in interviews): clean, readable, efficient node graphs, advanced skills where appropriate, attention to detail, and creative choices that make the shot appealing.
and (5) give a balanced view of benefits and risks for juniors, including the possibility that AI may bring more benefits than challenges overall.

Methodology (non-practical): an industry review based on a structured scan of academic papers, articles, and books (with Manovich & Arielli, Artificial Aesthetics (2024) as my main reference), plus toolmaker posts (e.g., Foundry on Nuke ML/CopyCat) and educator/studio blogs (e.g., Escape Studios).
I will also pull conclusions from a small number of informal interviews with recruiters, senior compositors, and educator/industry voices (for example, Alexander Williams – Escape Studios, Mark Spevik, Ariel Levental, Will Cohen) to capture current hiring signals, what portfolio evidence now matters and the overall approach to AI.

Source overview:
Manovich & Arielli AI is good at prediction and patterns, but humans judge meaning and quality. This supports my human-in-the-loop view: AI handles routine parts and the junior checks integration and tells the visual story. They also warn about “default”, same-ish looks. I’ll use this for skills and portfolio sections.
Artificial imagination, Narayan, Caillard, Matthews & Nairn (2022). Balanced view of benefits (speed/efficiency) and risks (jobs, quality): supports my benefits vs challenges section.
Foundry (Nuke ML/CopyCat) posts
Concrete task examples (e.g., assisted roto, cleanup, denoise/upscale). Shows what AI can do next in comp.
“Pfeiffer (2018), Creativity and Technology in the Age of AI”: professionals value Pros want tools that remove dull work so they can make decisions: backs my focus on creativity and judgement.
The Evolution of VFX-Intensive Filmmaking in 20th Century Hollywood Cinema: Shows how VFX became central to mainstream production, supports why fundamentals (edges, cleanups, colour, lens behaviour) remain valuable.

Expected outcome: a set of entry recommendations that tells juniors which skills they need now, what to prioritise in learning, and how to prove these in a showreel. The focus moves from shot volume to using human judgement: working efficiency, decision making, communicating clearly, and showing solid craft. If the evidence supports it, I will also state that, overall, AI brings more benefits than challenges for artists: faster routine work, less repetition, and more time for creative choices.

Week 5 – Student Presentations on individual Investigative Study Topics – continued

What was covered in the lecture:
This week continued presenting our ideas to the class.
Then, on the second part of the class we covered advanced referencing and searching for sources with Katie McNammara

Week 6 – Literature Reviews and Sources

What was covered in the lecture:
Literature Review Section:

– Describing the bigger pictures that provides background and creates space or gap for the research. The literature review will help to shoe the knowledge we have in the field. This is not just a summary, and it shows how the sources links to the argument. We present our sources based on their overall relationship to the project.

– In this section we need to introduce the way we describe things, like CG and computer generated

– We need to describe it like so: Prince (2010) and also the title when I introduce the subject. If its from a certain page, we can add the page and maybe the quote.

– Primary source – directly connect with the topic (interview, industry sources)
– Secondary source – usually published (book, academics)

Week 7 – Methodology

What was covered in the lecture:
Essay structure:
Intro 750 (done)
Lit review 1250 (done)
Methodology – 300-500
Discussion / finding section – 1000-2000
Conclusion – 500

Methodology:
The philosophy of reaserch methods
Why did I chose to approach and investigate this title like that (interviews, Primary research, secondary reaserch….)
– The method should be clear and easy to understand
– Write it as if you are designing it for someone else to follow
Types of sources (academic, industry…)
A full methodology section includes (not just method choice, but rationale, design, tools, ethics, limitations)
It should include:
– The study of research methods, the methods themselves and the
principles that guide them
– The overall research strategy and rationale for the project
Interview: Why you choose to interview in person
VES Talk: Why you choose to use this source
Secondary Research: Why you choose academic secondary research
– Why it is important to my reaserch

Qualitative approach
Content analysis, interviews or case studies (tends to be subjective, outlining the experience of something)
Quantitative approach
Surveys, tests or experiments (often involves counting data and statistics)

In my case:
Methodology (non-practical): an industry review based on a structured scan of academic papers, articles, and books (with Manovich & Arielli, Artificial Aesthetics (2024) as my main reference), plus toolmaker posts (e.g., Foundry on Nuke ML/CopyCat) and educator/studio blogs (e.g., Escape Studios).
I will also pull conclusions from a small number of informal interviews with recruiters, senior compositors, and educator/industry voices (for example, Alexander Williams – Escape Studios, Mark Spevik, Ariel Levental, Will Cohen) to capture current hiring signals, what portfolio evidence now matters and the overall approach to AI.
VES talk.

Gant Chart:
Introduction section – done
Lit review: done
Methodology: 17/11/2025
Discussion / finding section – 31/11/2025
Conclusion – 7/12/2025
Making a presentation – 7/12/2025

Week 8 – Workshop

What was covered in the lecture:
Upcoming weeks:
Week 11 – 8/12/2025 – presentation
Week 12 – 15/12/2025 – presentation
Week 13 – 5/1/2026 – submit the work

Then we had the chance to keep working on our essay. During the class, I was reading the feedback I got, after sending the first three sections of the essay for review. I started making changes according to the feedback.

Week 9 – Workshop

What was covered in the lecture:
Process:
From last week’s class and until today, I kept refining the first three sections (intro, literature review and methodology) of the essay according to the feedback I have received.
Then, I summarised the feedback I have received regarding my showreel. The only feedback that is missing it Hugo Geurra’s one, which I will receive tomorrow (25.11).
This week, I intend to keep working on the discussion section by summarising the feedback I will receive from Hugo Geurra and notes from the VES talk.

Feedback from industry artists:

Mark Spevick (FX artist at Framestore):
– Put the best shot at the beginning.
– Shot number 3 and four doesnt look good – 3 looking like a bad PS and after effects. Shot 4 needs more refinement
– They will look at the first 2 and if its not appealing they wont continue
– If I put something that doesnt look good, they will assume that I think that it looks good and therefore im not good enough.
– 4 is the maximum amount of shots

Lev Kolobov (Visual Effects Supervisor at HBO):
– Good start.
– Most of your shots in reel is day-to-night that it is very challenging. I would suggest to have few shots that you keep original footage as-it-is and add VFX (ideally day with nice light). Most of time this is what we do we add stuff to look like it was shot with camera.
– Gotham city shot – I would suggest to shoot something at real night city (even with your phone), don’t do any changes to footage just add skyscapper tower, neon sign and batman sign – and make sure it matches your footage 100%
– Also be careful of shots getting marky, washed out and “foggy”.

Ariel Levental (Lead compositor at Framestore):
– Good reel for a junior
– Add more Deep elements bc all the big studios using it
– Add one more shot of GS with hair details
– One of the shots should be visually impressive
– The impressive shot should be on the first or second of display

Hugo Guerra – Honest Showreel Review –

– It is a good junior reel and should pass the first round.
Greenscreen shot:
– Strong shot. Refine edges on the left side especially the hand on the pole. Consider using Nuke Cattery for a depth pass to apply on the plate. Gain up the man in the background.
CG integration shot:
– Excellent. Many versions exist but this one is well done. The monster breath created only in Nuke is a highlight.
Projection shot:
– Not strong because the source image is weak. Reduce ZDefocus and glow on the New Gotham sign. Adjust the Batman sign so it hits cloud layers. The card layout is good and the subtle camera move adds parallax.
Castle shot (set extension):
Main notes are fire simulation and environment reaction. Cleanup is good. Showing modelling skills is positive and shows awareness of the pipeline.

Follow up conversation:
– To stand out, show extra skills such as modelling, tracking and simulation used inside compositing shots.
– Learn the new machine learning tools from Foundry because studios need artists who can use them.
– On roto and basic tasks. These skills are even more important now. In the past much of this work was outsourced. Now more of it is returning to studios in the UK, EU, USA and Canada. Compositors need to supervise, refine and correct the tools.

VES session notes – Roto and Prep in the Age of AI (VES, 2025)
A panel of supervisors, leads and toolmakers discussed how machine learning is being used in roto, paint and plate preparation, what works today, what still needs human judgement and how this changes entry‑level work and hiring signals. The emphasis was on practical pipeline use rather than future speculation.

– ML success is not perfect. Human review is needed
– Studios use ML most for segmentation and matte assistance, depth previews, basic inpainting and speed‑ups on cleanup. These tools are treated as accelerators for first passes, not as final shots
– Labour shift –  With tighter budgets, full outsourcing of roto and prep has reduced. More of this work is done inside UK, EU, USA and Canada facilities, but now under compositor supervision and finishing rather than long runs of manual tracing.
– Some mid‑level artists also refine roto when needed.
– Juniors will be expected to supervise and correct ML passes, not just run them.

Week 9 – Workshop

At this point, I have finished the essay. On this week’s class we had the chance to start working on our presentation.