Practice 1: Developing Skills

[selinasharmapractice1]

Taking one of the assignments further - Neutropia (part 5): Exploration with AI to take concepts further (attempt 2)

In my previous AI exploration, I made reference to copyrighted IPs in my text prompt which the Netflix AI guidelines (2025) note is something to be avoided to prevent potential plagiarism. Further to this, ChatGPT has since updated their image generator. I decided to take the AI exploration with this project further to suit the Netflix guidelines and to test the capabilities of the new image generator.

Using the same prompt, I removed the references to the copyrighted IP. This time the results were less dramatic in its colour grading - this suggests that it may have previously heavily referenced the IPs that I had mentioned for the grading. These results are better creatively for the grading process as it is relatively flat in colours, so it can be given to a colour grader to creatively alter the image.

The new image generator is better at photorealism and does not appear to have artefacting - the words in the image are comprehensive and legible, including even on the signs and the vending machines in the back. The previous image generator had some artefacting with the letters and signs, and with the assets themselves. This is seen a lot less in the results of this new ChatGPT image generator.

Figure 1 - first output

The first output followed my render closer than the previous outputs, but it forgot to have doors for the barriers and made the station platform too wet. While I think it would be wet because of the weather conditions and river stream, I imagine drains would prevent this to make the platform safer. I asked ChatGPT to make the floor slightly less wet, turn the windows into sliding doors, and to move the boat closer to the platform so that passengers can logically get on/off. I also asked that there be a bucket where the cones are to signal leaks to the viewer as this was an asset the AI missed from my render.

Figure 2 - second output

The second output completely removed the wet floor rather than making it 'slightly less wet' - this highlights how the newer AI model still struggles with control (like making specific changes). It also removed my wet floor sign and cone assets from the scene completely, recontextualising the image despite me not requesting this. It did however, move the boat closer and even allowed room on the platform for when the doors open for passengers. It also included the bucket I requested, but it created a weird asset in the roof for the leak to come specifically out of. This logically does not make sense. It is clear that, while artefacting is less common and photorealism is more achievable, the AI still lacks control and reasonable judgement, which requires human input.

I requesting that the cones and wet floor sign be brought back into the scene and that the floor around that area be wet. I also requested that the strange where the leak was coming from to be removed.

Figure 3 - third output

The third output successfully brought the requested assets back and removed the unwanted asset. Overall, it has done a good job at following the brief, albeit a few iterations were required. Even still, this was a speedier process than traditional texturing and rendering.

Reflection and Evaluation

The issue of control is still present with ChatGPT's newest image generator. It makes creative decisions itself to remove assets, change positioning of assets, or create new ones. This makes art direction difficult to manage when the AI cannot stay consistent or follow simple change requests. However, for post-vis or look-dev, this is a decent workflow to get an idea of how final shots could look. Using them for final outputs can be problematic as a result because it will struggle with consistency between shots if it cannot be consistent within one shot. Furthermore, where the AI gets it information is a concern as it has made this station 'Northwood' in the image despite this in fact not being Northwood station. There was no mention of Northwood in the prompt, yet the AI made the decision that it would be. This highlights the practical issues of using AI for final outputs, despite its technical advancements in photorealism and preventing artecating.

[selinasharmapractice1]

Taking one of the assignments further - Neutropia (part 4): Exploration with AI to take concepts further

My time with the project began to run slim for texturing, lighting, rendering, compositing and grading. Thinking back on Andy McNamara's talk on 'Generative AI and The Future of VFX' (2026), I remembered how he suggested using AI as a tool for visualising what untextured renders could look like. Since this is the stage I have reached with the project (Figure 1), I wanted to take this as an opportunity to explore the capabilities for AI currently to understand how it can be used in a workflow (not to replace workflows or to be used as final outputs).

Figure 1 - raw render of my scene

I started by writing a prompt in ChatGPT to explain the art direction for this scene. The prompt was as follows: "Taking inspiration from the mood boards attached, I would like this plain render to be visualised into a station 100 years in the future where, instead of train stations in the UK, we have 'boat' stations', an adapted mode of transport that has become the norm due to extreme climate change. The weather in the UK 100 years from now is severe rain, flooding what used to be train stations. I have modelled and laid out a 'boat station', where the 'boats' are on water. They mechanically work in a similar way to how water rides work at theme parks. They move along a stream of water, where tracks would have been when it was a train station before. There are barriers with sliding doors that only open when the 'boats' have arrived. Take inspiration from how Elizabeth Line stations like Tottenham Court Road have done this in the current day to inform the barriers and sliding doors in this image. Have adverts on the barriers in sensible places that reflect the rainy society 100 years from now. Some examples could include: insurance for weather damage; waterproof sneakers; designer rain coats and umbrellas. The adverts should be futuristic as the society is 100 years from now, so imagine holographic imagery. In the back of the station, have vending machines for umbrellas, rain coats and rain boots. The cones are surrounding a leak in the station. There is a wet floor sign to bring caution and awareness to this. The inside of the station should overall look futuristic (but in a mundane way rather than a sci-fi/apocalyptic way), while the outside of the station should look worn down slightly for contrast. This can be done by having wear and tear to the brick wall outside, greenery growing as a result of rainfall, and leak stains. The image should aim for photorealism in style, looking indistinguishable from a photo. The outside should be foggy/misty and gloomy. There should be rain hitting the glass barriers and the river stream. The inside of the station should be futuristic yet bland, having a mundane approach. The atmosphere should take inspiration from films like 'The Batman' and 'Blade Runner 2049' in colour grading and in tone."

The mood board I am referring to here is the one I created during the planning stage (in part 2).

Figure 2 is the result:

Figure 2 - ChatGPT's interpretation

The atmosphere and style is what I was envisioning; however, the logic is not perfect. The AI has created rain in the station which does not make sense. I asked ChatGPT to amend this. Figure 3 is the result:

Figure 3 - ChatGPT's amendment

Figure 3 is better logically now. However, there are still issues with this image:

the 'boat' is not a safe or sensible distance away from the platform

it moved where I placed the cones in the scene

my seat design has been changed

When I asked it to make small changes to fix these issues, the entire image would alter slightly (Figure 4). This demonstrates the lack of control currently with Generative AI.

Figure 4 - the result of asking ChatGPT to move the boat closer to the platform barrier

Overall, it has done a good job interpreting the prompt but it is difficult defining details and having control of the details. Artistic direction becomes difficult to control. This is why AI is best used to brainstorm and visualise rather than to use as final output.

To take visualisation further, I took the output and asked Sora 2 to simply 'Make it rain outside'. Sora 2 was able to do this, although it completely removed the water stream outside and had a train run inside the station which is completely nonsensical (Figure 5).

Figure 5 - screengrab of Sora 2's first attempt

This time, I specified to have the train be outside (al

Figure 6 - Sora 2's second attempt

Reflection

I approached using GenAI following the guidelines of Netflix and the teachings of McNamara. However, I realised in hindsight that I did forget Netflix’s guidelines makes it clear that reference to other IPs in prompts should be avoided.

Bibliography

McNamara, A. (2026) Generative AI and the Future of VFX. Lecture delivered online, 11 March, 14:30.

Netflix (2025) 'Using Generative AI in Content Production'. Netflix | Partner Help Center. Available at: https://partnerhelp.netflixstudios.com/hc/en-us/articles/43393929218323-Using-Generative-AI-in-Content-Production (Accessed: 1 April 2026).

[selinasharmapractice1]

Andy McNamara - Generative AI and The Future of VFX

Introduction

Generative AI is starting to be implemented into the industry in explorative ways, making the future of the industry and Generative AI hard to predict. McNamara, Gen AI Lead/Creative Director at Cinesite/Tech X, presented a talk to us which gave an insight into how they are experimenting and using Gen AI. He comes from a background in 3D, VFX and CG for 35 years, so his research and opinions come from an informed perspective. This talk provided insight into Gen AI specifically in the VFX industry.

The Talk

McNamara believes that curiosity, a progressive mindset, creativity and technology are what drive innovation. Approaching GenAI this way, therefore, can innovate workflows and processes. Innovation is important as it gives a competitive edge. He expresses that AI and machine learning can accelerate and support (not replace) a VFX workflow.

With AI being a 'buzz-word' at the moment, McNamara notes that there are many opinions about AI. He suggests 'cut[ting] through the noise' to put experimentation first.

Using a hybrid workflow of traditional VFX and AI can accelerate the mundane workflows allowing focus on the creative process. It can be used in some cases for final outputs, but this is in specific use-cases. GenAI as a final output is currently disruptive as a workflow because you do not have control over it.

At Cinesite/Tech X, they are having incremental adoption of AI, keeping artists in the loop. They do so in a way that allows their creatives to 'do their best work'. One GenAI model being heavily explored is ComfyUI. This allows more control than text to image prompting as it is a node-based image generator. This workflow will be familiar to traditional compositors who are used to using Nuke for node-based compositing.

McNamara outlines a few use-cases of AI:

text to image (worded prompt)

image to image (sketch as input)

text to video

image to video

video to video (using AI as a renderer)

Using AI as a renderer, he notes, has the most potential for the future as it could replace traditional rendering. Providing a screenshot of a 3D shot in the workspace to AI can provide fast outputs that could replace the long render times of render engines. However, I can see that there would still be the issue of control in terms of getting the textures and lighting intended through this method. With exploration of ComfyUI, perhaps this is something that could allow more control if intending to do this for final outputs.

Image/video editing can be useful for: removing objects/subjects from plates (speeding up paint and roto); image in-painting; video in-painting; video out-painting.

LLM's are useful for editing existing imagery using 'conversational editing'. This allows a back-and-forth where quick iterations can be formed. In the future, conversational video editing could be a potential innovation. AI can also upscale resolution, vary the angle of a shot, and do image restyling. Nevertheless, it is still hard to control.

Most useful to the workflow is 2D AI assistance at the moment. It helps roto, masking and segmentation. While it is not perfect, it does a good majority of the work that speeds the workflow. Human input is still required to do the 20% that AI will miss, and to also clean up any mistakes that the AI could make as AI uses computational estimates rather than human judgement. AI can do the AOV outputs similar to a 3D rendering engine (e.g. z depth, cryptomattes). This can be useful for stereoscopic workflows, although it may not be perfect as it does estimates to make the outputs. For crowds, AI 2D crowd replication can also be useful to workflows.

The current challenges of AI include the rapid rate of progress; the quality of it (technical and creative); control; data security (where does the data inputted go?); ethics and copyright (how has the AI been trained?); environmental concerns (data centres); and perceptions of AI (of clients, creatives and consumers). Although, client perception is starting to change as they are becoming curious and interested in AI's capabilities.

Ultimately, AI is not going anywhere. It is another toll in the creative toolbox. As a result, tools and workflows are changing. The creative roles and skillsets will therefore evolve. However, what always be valuable is solid art and design skills. As the skillsets of artist roles evolve, artists must be kept in the loop and there must be authenticity. This is hugely important. To keep up with the changes, we should be curious and stay informed. Job descriptions will likely have an expectation of an understanding of AI tools, so it is important to future proof your skills.

While AI is still in its early stages in terms of workflow implementation, we are yet to see a 'Maya' or 'Nuke' of AI - but it will come. AI will encourage generalist mindsets and smaller productions. AI will especially support Production. The problem with AI here is that it will potentially take away junior roles, making the industry even smaller and harder to break into than it already is. The other perspective of this is that new roles may replace the old roles as pipelines evolve. Therefore, it is important to understand the different types of techniques you can use with AI. Diving into ComfyUI can be useful into understanding the technicalities.

The result of AI on filmmaking could change the skillset and scope of filmmakers. Although, quality will always stand on top and slop will always surface to the bottom. There is still a skillset to filmmaking whether you are using AI as a tool or not. If AI is used sensibly with a keen eye, this can accelerate concepting.

The rate of AI's evolution is extremely fast. We are already starting to not be able to tell the difference between AI and photographic imagery/film. What will still be a point of contention is the controllability. The evolution of AI is at a much faster rate than the rate of legislation. There will need to be legislative placement on AI to prevent the misuse of it.

McNamara predicts that there will be an AI literacy expectation potentially in the future, like how computer literacy is an expectation of most jobs today. Still, core skills will always be necessary. The software may change, but fundamentals remain important. Things get called AI that are not AI, which confuses that public perception on AI's capabilities even more. He concludes that AI is ultimately a tool of creation.

Reflection

This talk was delivered March 11th 2026. Reflecting since then, many of his predictions and opinions seem to be coming true. I am starting to see AI roles appear that ask for previous experience in 3D/VFX/Maya/Houdini/NUKE etc. I am also starting to see less specialised roles and more '3D/AI generalist' roles. Further to this, I am seeing people using ComfyUI to turn their 3D scenes into renders without using traditional render engines like Arnold. It is clear that the industry is undergoing an AI revolution slowly. Since this talk, AI has evolved even more and gotten even better at image generation. This makes me keen to explore how I can use AI in my own workflows too, perhaps trying the capability of AI as a renderer.

Bibliography

McNamara, A. (2026) Generative AI and the Future of VFX. Lecture delivered online, 11 March, 14:30.

[selinasharmapractice1]

Netflix's guidelines to using AI

Introduction

GenAI is a new technology that is starting to be experimented with and implemented into workflows. Its use-cases are not fully realised yet as experimentation is ongoing with it. There are also legal and ethical grey areas with it as we cannot be sure how the AI has been trained, and it has an environmental cost.

Netflix has released an article last year laying out their guidelines on GenAI and how they’re using it. This blog post details Netflix’s approach to AI. It is useful to look at how big studios are using GenAI as it provides guidance and sets a precedence.

Netflix's guidelines

In their introduction, Netflix signposts that they see GenAI tools as ‘valuable creative aids when used transparently and responsibly’ (Netflix, 2025).

Acknowledging the the legal issues that can arise with GenAI, Netflix sets out the following considerations when using it in creative workflows:

1. Outputs do not recreate identifiable characteristics of copyrighted material.

2. GenAI tools do not store/reuse/train on the production data inputs/outputs.

3. GenAI used in enterprise-secured environments to protect inputs.

4. Generated material is not part of the final output.

5. GenAI is not used to replace/generate new performances without consent.

The approach of Netflix prioritises protecting data and creative rights when using GenAI. This means GenAI tools used should not be trained off copyrighted material nor should I be trained without the consent of artists. These raise legal concerns.

As GenAI trains off of material fed to it, it cannot produce anything substantially new and risks created things that already exist (hence copyright concerns). Because of this, Netflix insist that GenAI should not be used when making ‘main characters, key visual elements, or fictional settings that are central to the story’ (unless given written approval). Making reference to copyrighted material should also be avoided as it can risk plagiarism. Ultimately, GenAI should be approached with caution and sensibility just as you would with traditional production.

Taking an ethics approach, Netflix ask ‘to consider both the intent and the impact’ of GenAI content. They note that GenAI can be misused to mislead or misrepresent content. They also stress the importance of protecting union-represented members to not replace/impact work done by these members without their express consent.

Plot-aiding assets or prominent elements should be treated with greater care when it comes to GenAI. Smaller props that do not contribute to the story can be regarded as ‘incidental’.

Netflix suggests using GenAI to ‘explore ideas or mockups’ but that final versions should ‘involve meaningful human input’.

Reflection

Something to make note of is that the article often sets out guidelines of what not to do, before then following with the condition ‘unless’ (given written approval, for example). This suggests that these guidelines are still being experimented with and are subject to change as technology, workflows, and laws change.

Despite this, the Netflix guidelines give a good insight into how big studios are approaching GenAI and what its strengths/limitations are. If I experiment with AI, I will be keeping these guidelines in mind to inform my workflow and process,

Bibliography

Netflix (2025) 'Using Generative AI in Content Production'. Netflix | Partner Help Center. Available at: https://partnerhelp.netflixstudios.com/hc/en-us/articles/43393929218323-Using-Generative-AI-in-Content-Production (Accessed: 1 April 2026).

[selinasharmapractice1]

Taking one of the assignments further - Neutropia (part 3): 3D Modelling

After planning the scene, I began blocking out the layout of it and where assets will go.

First, I blocked out where the platform barricade doors would go (Figure 1 and 2). I started with a cube that I scaled up and positioned to be at the edge of the platform as a safety procedure.

Figure 1

Figure 2

To make the scene feal photoreal, scale is important. Therefore, I put a human reference into the scene as a scaling reference (Figure 3 and 4). I scaled the human reference to match the humans in the original photographic plate (Figure 5). The human reference was useful to then plan the size of the benches which I could block out with cubes. I then cut edgeloops planning where the doors and panels would go, before extracting the faces for the doors to make them separate objects (Figure 4).

Figure 3

Figure 4

Figure 5

I then examined the scene in perspective view so get an idea of how the whole scene looks as a whole outside of the shotcam (Figure 6 and 7).

Figure 6

Figure 7

Back in the shotcam view, I examined the scene in a wireframe view to see the tracks behind the barriers against the scene (Figure 8).

Figure 8

Happy with the block out so far, I made a new maya scene to build the assets in a clean environment. I imported the human reference so that the assets could have a scale reference as I model.

As I was making the models, I made a moodboard to get ideas of what 'futuristic design' could potentially look like taking inspiration from what designs currently exist (Figure 9). From here, I began modelling.

Figure 9

Figure 10 - bench process

Figure 11 - bench

Having finished the bench, I began modelling the contextual assets that will fill the scene: a traffic cone for areas of leakage; a bucket to catch leaks; a caution sign for wet floors; a vending machine that supplies umbrellas, raincoats and rain boots; a drain; an umbrella; an an umbrella helmet which I thought would be adapted into everyday life to deal with the weather conditions like how face masks were worn by society during the COVID pandemic (Figures 12-16).

Figure 12 - cone

Figure 13

Figure 14

Figure 15

Figure 16

Once I was happy with the assets, I brought them back into my scene and placed them throughout (Figure 17 and 18).

Figure 17

Figure 18

Next, I UV mapped my assets to prepare them for the texturing stage (Figure 19).

Figure 19

Before moving onto texturing, it is good to light the scene so you can assess the textures against the scene lighting. I created mesh lights in the station and had an skydome for the outside (Figure 20).

Figure 20

Reflection

I enjoyed modelling the assets and populating the scene in this way. The next stages would be to texture the assets in the scene, create narrative lighting, and colour grade for atmosphere.

[selinasharmapractice1]

Taking one of the assignments further - Neutropia (part 2): Mood-boarding and Planning

To help steer the project in a cinematic and believable direction, I made moodboards of both film references and real life references to base my inspiration on tangible existing things. Collating such references can help ground the project to a society we are familiar with rather than going too futuristic and going off of sci-fi tropes.

Figure 1 - Overall look at the moodboard

On the left in Figure 1, I have made a moodboard putting real life references and inspiration for the project. These references will help the project to feel based on reality. On the right in Figure 1, this is the moodboard for creative projects that I can take inspiration from: stylistically, contextually and compositionally. These references will be a reminder to make my project visually interesting.

Figure 2 - LINE Mobility (Art&Graft, 2026) on the left; Heavy Rain (Quantic Dream, 2010) on the right.

The LINE Mobility project (Art&Graft, 2026) presents a unique new system of travel, where people travel in pods along tracks, being a combination of the car and the train. Studying how they approached presenting a new method of travel can help with my approach in this project that combines boats and trains as a method of transportation.

Heavy Rain (Quantic Dream, 2010) is a good reference for rainy conditions and rain patterns. Rain going in different directions, falling at different speeds and different sizes will all contribute to believable rainfall.

Figure 3 - Blade Runner 2049 (Villeneuve, 2017)

Blade Runner 2049 (Villeneuve, 2017) is a great cinematic reference for a rainy environment set in the future. The mood and atmosphere set by the weather conditions adds to the storytelling. Further to this, the way they think about lighting in this film is a great reference on a technical level and a contextual one too. The lights would be bright and colourful to be seen through foggy and misty conditions, and the light would travel through the fog/mist creating a halation effect. These are effects I would like to implement in my project.

Figure 4 - The Batman (Reeves, 2022)

The Batman (Reeves, 2022) provides great reference for rainy conditions and it is used atmospherically.

Figure 5 - Death Stranding 2 (Kojima Productions, 2025)

I really liked the rain on the camera lens in Death Stranding 2 (Kojima, 2025). This heightens photorealism as well as immersion for a viewer as it is mimicking how a camera lens would pick up rainy conditions. This is something I would like to implement in my own project if I do a camera angle that is sat outside of the train platform.

Furthermore, the greenery resulting from rainfall is something to remember to implement in my work where it makes sense and it plausible, like an outdoor brick wall for example.

Figure 6 - Photorealistic flooded subway (Pepe Baze, 2021)

This artist Pepe Baze is a perfect example of a photorealistic space that has been flooded presented in a cinematic way. I will be drawing heavily from this artist's work in this project.

Figure 7 - Roller coasters with water

These water rides have a mechanical system where boats are carried along water streams. This could be useful when thinking about how the actual boat system would work mechanically.

Figure 8 - Primary photography

Figure 8 is a photo I took on a misty morning which provides reference of how light looks through a camera lens in these conditions: bright, exaggerated, halation. This will be a good reference when trying to get a photorealistic look.

Figure 9 - Flooded stations

I imagined the boat system replacing train stations from flooding of the tracks creating bodies of water. This is something that is already happening, especially in the UK. When searching for references on Google Images, simply searching 'flooded stations' brings up examples that are clearly in the UK, and even local examples that are in Hertfordshire. These examples provide a grounded reference to base the project off of.

Figure 10 - Rainy conditions paraphernalia

Smaller details that tell the story of an extreme rainy environment would be drains to prevent flooding, wet floor signs, wet/muddy floors indoors, leak buckets, advertisements capitalising off of the rainy society etc.

Some advert ideas that can be scattered around the environment subtly to contribute to the environment's storytelling:

charities for families affected by the weather conditions

shoe adverts boasting 'anti-slip' or 'anti-wet' technology

fashion adverts promoting designer umbrellas or raincoats

'TFL Boats' (company name I am assigning to the travel system) encouraging premium 'first class boat' upgrades.

vending machines with umbrellas, raincoats and rain boots.

dehumidifier adverts

Figure 11 - Train platforms with sliding doors

Finally, to prevent people swimming in the water streams and to prevent drunk people from falling in and drowning, the stations will have these sliding doors to ensure safety, as well as to prevent water splashing onto the platforms and causing slip hazards.

Evaluation

Moodboarding has helped me to think about the reasoning behind my decision making. It has also helped to give me a better idea of the cinematic look I am going for in this project. My next blog posts will make a start on the project, continuing on from where I previously left the Neutropia project.

Bibliography

Art&Graft (2026) LINE Mobility. Available at: https://artandgraft.com (Accessed: 1 April 2026).

Baze, P. (2021) Photorealistic flooded subway. ArtStation. Available at: https://www.artstation.com/artwork/1835we (Accessed: 1 April 2026).

Kojima Productions (2025) Death Stranding 2 [video game]. Tokyo: Kojima Productions.

Quantic Dream (2010) Heavy Rain [video game]. PlayStation 3. Paris: Quantic Dream.

Reeves, M. (2022) The Batman [film]. Burbank, CA: Warner Bros.

Villeneuve, D. (2017) Blade Runner 2049 [film]. Los Angeles, CA: Warner Bros.

[selinasharmapractice1]

Taking one of the assignments further - Neutropia (part 1)

Reflection on Neutropia

Having completed all of the assignments and reflecting on them, I will be taking Neutropia further. After having a tutorial with Saint, he recommended thinking more about the cultural logic behind images. For example, he suggested taking inspiration from stations along the Elizabeth Line that use barriers to prevent people falling/jumping onto the tracks. He also suggested a hierarchy boat system (like how we have first class and standard, for example) where different boats can represent social status. Finally, thinking about the small everyday mundane things can add so much to the storytelling of an image.

Figure 1 - how I left the Neutropia project last

Due to time restrictions, I will look into kitbashing different types of boats using pre-existing assets. This way, I can still make something using my creative vision while also saving on time.

My plans for this scene so far:

kitbash a boat in 3D.

add more greenery to the brick wall to signal to the viewer that the weather conditions have promoted unintentional growth.

Add barriers in 3D to the platform to provide logic for preventing people jumping/falling into the water stream.

Have the inside of the station (waiting platform) appear maintained in comparison to the outside of the station.

raincoat/umbrella/rainboots vending machine perhaps.

Adverts promoting gadgets or charities to help with the extreme weather to provide context of societal needs.

buckets placed where there are leaks through the roof.

A couple of other stations to show a variety of what stations look like in this society.

Bright lights and signs to help people see through the fog and mist (potential light pollution in the sky as a result).

Grey overcast and mist

Potentially rain simulation

Some things I can take inspiration from to help guide the project:

Rollercoasters with water (Figure 2)

Art&Graft's Line Mobility project (2026) (Figure 3, which I got involved with during work experience with them)

Heavy Rain (Quantic Dream, 2010) (Figure 4)

Elizabeth Line stations with platforms.

Figure 2 - an example of a ride with water

Figure 3 - Line Mobility (Art&Graft, 2026)

Figure 4 - Heavy Rain (Quantic Dreams, 2010)

In the next blog post, I will begin mood-boarding and planning further.

Bibliography

Art&Graft (2026) LINE Mobility. Available at: https://artandgraft.com (Accessed: 1 April 2026).

Quantic Dream (2010) Heavy Rain [video game]. PlayStation 3. Paris: Quantic Dream.

[selinasharmapractice1]

HBO VFX Production Coordinator Bootcamp with CAVE Academy - part 1

I was selected to participate in training with HBO and CAVE Academy focused in VFX Coordinating from January to February 2026. The training involved learning the technical process behind VFX (delivered by J Amin), onsite visits to studios, and industry talks from coordinators, supervisors and producers (client side and vendor side).

The first week of the Bootcamp was delivered at CrewHQ, Warner Bros Studios Leavesden. Following the first week, the rest of the sessions were delivered online 3 days every week.

This blog post details some of what I learnt in the technical sessions.

What is VFX [from a coordinator perspective]? - J Amin, 12/1/26

Important to know each stage of the pipeline and do the schedules for everyone, so feedback can be passed on to the artists and vendors/clients.

You do not want to be in the position where you do not know what you are doing, or you do not understand what needs to be done.

Always start with good preparation in pre-production so the effects can be executed to the best level. Get VFX involved as early on in the process as possible.

VFX is the process where imagery is created/enhanced in filmmaking.

on-set crew to get all the references so that there is less guess work in VFX.

Sometimes VFX team will do SPFX too (anything that can be photographed).

SPFX and VFX work in collaboration with stunt team.

Most successful shots require both VFX and SPFX.

SPFX can also create VFX elements that are photographed during production -> data capture e.g. HDRIs.

VFX complements SPFX's work -> does not take SPFX's work or job.

Communication, friendliness and working together is super important.

Green lit show process: hire a VFX Supervisor and VFX Producer -> determine the types of VFX involved -> Storyboarding/Previs -> Budgeting

A VFX Supervisor will guide a project creatively from concept to finish, bringing technical skill and a specialised skillset.

The types of VFX is determined from doing a script breakdown, which will be done by the VFX Supervisor and VFX Producer. They determine how it will be accomplished and the cost.

With Storyboarding and Previs, this can be 2D or 3D. 3D is better for providing concise answers and accurately budgeting. 2D provides good visual reference for how VFX can execute each shot.

At the budgeting stage, the VFX Producer (in discussion with the VFX Supervisor) will assign cost to each potential shot. After budgeting, a production may need to scale back on amount of VFX shots or their complexity.

Figuring out scheduling, budgeting and payment terms.

Pre-viz, tech-viz and post-viz to help with planning.

Previs is when you take your storyboards and turn them into what they would look like through a camera.

Techvis figures out how to do it and tests camera speed. Important to also do it for health and safety concerns.

Postvis allows for changes with existing plates e.g. if a shot needs changes. Postvis also plans for the compositing stage, being the 'slap-comp' stage.

After the breakdown of the script and budgeting the cost, you have to put it in a spreadsheet. Have to then massage things to look good for a client.

Rebids may be necessary.

For bidding, best to be upfront and honest about the work with about 20% leeway. With rebids, it may be a case of assigning money elsewhere on paper to look better for a client, even if it is not done exactly as pitched, redistributing where it is going while still paying the artists the same rate as intended.

Important not to undersell as artists deserve to be paid for their work. VFX industry notorious for being undervalued.

Never offer free work as this is a bad business practice.

Early stages will be guess work until getting to production stage, so proper planning and budgeting is crucial.

If clients go off contract, then you do a rebid.

With shot distribution to vendors, smaller VFX facilities are less expensive but cannot usually take on large scale shots. Meanwhile, larger VFX facilities are more expensive, but they can create highly complex shots at great quality.

Consistency in workflow so that there is consistency between shots.

Data must be stored in folders with the correct naming conventions.

Outputs need to be consistent for handing over in pipeline.

Onset data needs to be standardised so it can be extracted and organised efficiently.

Witness cameras show you how the lighting is done on set, capturing data that the HDRI may not capture/make clear.

Factors that will affect distortion: what lens, what camera, depth of field, focus, prime, zoom.

The Life Cycle of a A Shot - J Amin, 13/1/26

The workflow of the input and output of a shot matters.

RAW has the most data and range. Allows you to change colour and contrast later on.

RAW is usually linear colour space.

RAW -> VFX -> Graded (typical workflow for a shot)

Want to preserve the RAW for as much as possible through the pipeline and workflow.

LUTs (Look up tables) used for temp looks. These will be used until instead of baking looks. Less destructive workflow.

Making changes to an LDR (low dynamic range) image will cause banding, as opposed to having an HDR (high dynamic range) image. Reduces clipping (loss of information) by using an HDR image. LDRs have less latitude, hence clipping and banding.

Ideally, you want to take HDRIs for each type of light (and if there's even more time, the types of shadows) so the lighting team has everything they need.

Size of the lights affect the shape and strength of the shadows.

Problem with shooting panos is the distortion on the edges of the image. For large scale sets e.g. buildings, have to shoot straight-on and shoot incrementally so that you do not get distortion.

You also need to capture the different brightness levels when capturing HDRIs (bracketing).

Reflection

The sessions have built on what I have already been learning in University but it has contextualised it with a Production perspective. I now understand why VFX works in specific pipelines and workflows.

[selinasharmapractice1]

VES London – The Splatting Shift: Innovation in 3D Processing (Visualskies Studios Battersea)

Gaussian Splatting is a new machine learning tool that is being implemented into VFX workflows. The Visual Effect Society (VES) hosted a Gaussian Splatting event at Visualskies Studios which I was keen to attend to find out from industry professionals how they are using Gaussian Splats in their workflows. This blog post details what I learnt about Gaussian Splats during the event.

Figure 1 - schedule hand-out from the event

Visualskies is a studio that focuses on 3D scanning and data capture, sitting in the space between production and post-production. Splats tend to feed into pre-vis and virtual production.

Rather than replacing traditional 3D capture, splatting is being used complementary to these traditional methods. It is best to use whatever combination of tools that best suit the requirements. Foliage heavy environments are where splats are best-used.

From a Producer perspective, Splats have a faster turnaround time which allows for faster decisions. This allows for extra time and flexibility for later in the pipeline. Splats also avoid needing to recapture if you capture the set on the day with splatting. Therefore, it prevents data loss. This makes their implementation into workflows appealing, especially in pre-vis and in virtual production.

As well as this, Splats are also starting to be implemented into final deliveries. The data captured can be navigated in a 3D space to give multiple perspectives.

Combining splatting with realtime rendering, Unreal Engine at this current point struggles with. The data is far too heavy and dense for realtime rendering. In this case, a hybrid approach is best:

Geometry for the depth

HDRI for accurate lighting

Gaussian Splats used for texture information.

The ‘friction points’ of using splats include:

Messy spreadsheets

Endless whatsapp groups

Overly complex naming conventions

However, Gaussian Splats can help with winning clients as they can easily visualise what they are getting with splatting. Clients are usually unwilling to pay the processing fee unless they know exactly what they are getting, so Splats make this slightly easier.

Problems onset has come from using apps like Polycam that have not been approved. The great thing about such splatting apps are that they democratise scanning through tech being more easily accessible. The downside to these apps is that nobody knows where the captured data is going which can be a data concern. To deal with this issue, Visualskies have made their own scanning app so that there can be safe and secure scanning.

Another way Gaussian Splats can be used is to make splats from CG, which prevents rendering angles when you can generate splats and put it at any angle in realtime.

Nuke 17 allows for splat implementation using the USD format. You can then roto things out of the splat directly in Nuke, and can directly colour grade the shadows out in Nuke.         This can be combined with CG or 2D plates, allowing for a hybrid approach in Nuke. With Nuke 17, you will also be able to export splats out of Nuke as a PLY (the splatting file format).

Gaussian Splats are also useful for set extensions.

The current limitations of Splats are the inability to change the lighting.

Finally, the talk concluded by expressing how accessibility to the tools is most important for innovation, exploration and creativity.

Reflection

I now have a greater understanding of how Gaussian Splats are being implemented into VFX workflows. The technology is a useful tool for achieving photoreal results. I am interested to see how this develops as it gets implemented into more pipelines, and how the tool itself might develop.

Bibliography

Visual Effects Society (2026) VES London – The Splatting Shift: Innovation in 3D Processing. Event held at Visualskies Studios Battersea, 29 January.

[selinasharmapractice1]

Assignment 4: Memory Palace - Style Frames

Results

Style Frame 1 - Metaphase

Style Frame 2 - Anaphase

Style Frame 3 - Telophase

Style Frame 4 - Cytokinesis

Reflection + Evaluation

Going back to the philosophy of infographics, these renders could be enhanced by having graphical annotations that takes a snapshot zoomed into the nucleus and showcases what is happening to the chromosomes (containing the DNA). These renders would immediately make sense to someone familiar with biological processes and mitosis, but it would not make sense to the average person because there is such little information to explain what is happening. Therefore, if I took this project further, I would think to include such context in a digestible way.

Further to this, I would give context to the renders to make sense of why the process of mitosis happens and why it is necessary. Having an environment to give the cells context would help.

[selinasharmapractice1]

Assignment 4: Memory Palace - Workflow

Moving on from planning into making, I made two files: one for the DNA replication process, and another one for the cytokinesis (division) process. I made two projects as I was unsure which one of the two ideas to go for.

DNA Replication project

Figure 1 - DNA replication project in Cinema4D viewport.

Figure 2 - DNA replication project in Cinema4D viewport.

Figure 3 - DNA replication project render.

Figure 4 - Elements making up the helix

This helix visualisation was intended for a DNA replication breakdown project, part of the interphase stage before mitosis. I was enjoying making a jelly-like glowing material for this helix. However, I realised that I had been using the standard Cinema 4D renderer and materials. I wanted to explore Redshift but I liked what I had already done in this project. Since I was already stuck between 2 ideas for a project, I took this as an opportunity to explore Redshift materials and renders in a cytokinesis project. As a result, I put this helix project on hold.

Cytokinesis project

I used cloners to give the illusion of the cell separating. I did a simple animation of the mitosis process so that I could then render out the key beats as style frames. Figure 5 shows a viewport playblast of this.

Figure 5 - Cytokinesis project viewport render

Figure 6 - Cell setup

As I had explored a jelly-like texture before with the standard Cinema 4D shader, I wanted to try make a jelly-like texture in Redshift. I did this by adding noise to the bump map. This was a similar process to what I did with the Cinema 4D standard shader in the helix project.

Figure 7 - Material setup

Figure 8 - Material preview in the viewport

I am familiar with using Maya as a 3D software and Arnold as a renderer, so it felt familiar using the nodes to make materials in Cinema4D (Figure 7).

I made the outer jelly coat of the cell translucent so that the nucleus on the inside could be seen, as the nucleus is important to illustrate and bring attention to when it comes to the stages of mitosis (mitosis involves DNA replication, which happens in the nucleus).

I chose to do a translucent jelly texture for the membrane to signal to the viewer that this is a cell.

After textures, I inserted a skydome and put a HDRI on the skydome to give lighting information to the scene for the texture to catch. I then set up area lights in a basic three-point setup. The renders of these can be seen in Figures 9-11.

Figure 9 - Cytokinesis project render; metaphase stage

Figure 10 - Cytokinesis project render; anaphase stage

Figure 11 - Cytokinesis project render; telophase stage

From using Redshift renderer and Redshift materials, I can see how it produces photoreal results compared to the standard Cinema 4D materials and renderer which feel dated stylistically. This makes Redshift a good renderer option for photoreal renders.

Reflection and Evaluation

After exploring the Cinema4D software and the Redshift renderer, I can see why they are the top choice of renderer and software in the motion graphics and scientific animation space. The modelling approach in Cinema4 is non-destructive and Redshift produces photoreal results. I would choose to use this software in the future if I chose to pursue motion graphics.

Bibliography

daisukexdark (2020) Cinema 4d Cell Modeling & Animation Example. Available at: https://www.youtube.com/watch?v=4zMeSSFfjw4 (Accessed: 27 January 2026).

Daniel Danielsson (2024) C4D TUTORIAL | Realistic DNA (and Unrealistic DNA) 🧬. Available at: https://www.youtube.com/watch?v=ZStz4UGiN44 (Accessed: 27 January 2026).

Pluralsight Courses and Tutorials (2013) Top Tip: Use Cloners to Create a Dividing Cell in CINEMA 4D. Available at: https://www.youtube.com/watch?v=zMeNgWs6krs (Accessed: 27 January 2026).

Tutorama Brasil (2016) Timelapse/Process: Projeto DNA com Cinema 4D. Available at: https://www.youtube.com/watch?v=IKc_CEs58oU (Accessed: 27 January 2026).

[selinasharmapractice1]

Assignment 4: Memory Palace - Concepting and Planning

Inspired by medical animation, I wanted to take explore this further during the assignment. Having studied biology at A-Level, I thought to present a biological process in 3D.

A biological process that first came to mind was mitosis, which is cell division. To refresh my memory on the stages of mitosis and to get inspiration from others, I conducted some research by looking at explanatory videos and of the process and 3D visualisers.

From watching the Amoeba Sisters' video on mitosis (2016), I was able to sketch out and storyboard the key beats of the process (Figures 1-4).

Figure 1 - basic beatboard for mitosis, focusing on a narrative to put the process in context

Figure 2 - another basic beatboard for mitosis, drawing more focus to the DNA replication part of the process, again focusing on narrative to contextualise the process.

Figure 3 - quick sketch considering focusing on only the DNA replication part of the process

Figure 4 - Beat board of the mitosis process only without going into the details of DNA replication and without context

These sketches explore the different ways that I can show the process. At first, I thought it would be best to show the process in context to demonstrate to a viewer why the process happens. Figures 1 and 2 storyboard how mitosis aids repair of cuts and injuries in our skin.

However, I realised that this could actually complicate the process for a viewer. Instead, I thought it could be better to delve deeper into specific stages of the process. Figure 3 considers focusing on DNA replication which happens in interphase (the stage before mitosis getting ready for cell division). Figure 4 on the other hand considers focusing on the actual division process, where the cells split.

Once I had a better idea of what to focus on, I began to mood board the process. Figure 5 is a screenshot from the mood board where I have annotated the stages that I want to present in 3D.

Figure 5 - planning which stages to show in the 4 style frames.

Figures 6-8 show the rest of the mood board where I have collated visualisations of the different elements of the process to help guide the project stylistically.

Figure 6 - mood board section visualising the chromosomes

Figure 7 - mood board section visualising the cells

Figure 8 - mood board section visualising the stages of mitosis

Finally, I decided to focus on showing the later stages of mitosis: 1. Metaphase; 2. Anaphase; 3. Telophase; 4. Cytokinesis. (Highlighted in Figure 5).

Having researched medical animation previously, I found that Cinema4D is the software primarily used as it is best for motion graphics. I have some experience in Cinema4D from doing work experience as a 3D Artist at the animation studio Art&Graft, so I will be using this assignment as an opportunity to delve deeper into the software.

Bibliography

Amoeba Sisters (2019) DNA Replication (Updated). Available at: https://www.youtube.com/watch?v=Qqe4thU-os8 (Accessed: 26 January 2026).

Amoeba Sisters (2016) Mitosis: The Amazing Cell Process that Uses Division to Multiply! (Updated). Available at: https://www.youtube.com/watch?v=f-ldPgEfAHI (Accessed: 26 January 2026).

Curious Refuge (2015) Cell Dividing: Adobe After Effects Tutorial. Available at: https://www.youtube.com/watch?v=9aBw5LqUT5w (Accessed: 23 March 2026).

daisukexdark (2020) Cinema 4d Cell Modeling & Animation Example. Available at: https://www.youtube.com/watch?v=4zMeSSFfjw4 (Accessed: 23 March 2026).

Thom Leach (2022) Mitosis Animation. Available at: https://www.youtube.com/watch?v=7ybxaYhRpIA (Accessed: 26 January 2026).

yourgenome (2015) DNA replication - 3D. Available at: https://www.youtube.com/watch?v=TNKWgcFPHqw (Accessed: 26 January 2026).

[selinasharmapractice1]

Assignment 4: Memory Palace - Medical Animation Research

Studio Research

Animated visuals are a popular way of illustrating scientific processes and information that cannot be seen or are hard to explain. This makes medical animation its own genre in 3D and VFX. Having previously studied Biology in my A-levels, I relied on such animations to process information.

The nest phase of my research involved looking at studios. I looked at Random42 and Helix Animation, studying how they explain scientific processes in the 3D medium. Their work is abstract yet photoreal, making this type of work a creative problem-solving task by having to imagine and execute what the human eye cannot see. Most of all, their work is clear in explaining scientific processes.

Figure 1 - example of Random42's work

Figure 2 - example Helix Animation's work

This avenue of visual storytelling of information is one I would like to explore.

Practice

These type of abstract visuals are new to me in 3D. To feel more comfortable with this type of work, I practiced using a simple tutorial by VideoFort (2014) as a reference. These were the steps I took:

In Cinema4D, create a platonic. Hit c to make it editable. select the polygons and inset. Turn preserve groups off. Add an offset value of 10 and hit apply = creates an inset on the polygons.

matrix extrude the inset polygons = creates tentacle-like arms.

create a sub-division modifier and drop the platonic into this modifier.

put phong tag on platonic so that the shading is smooth.

drop everything into a cloner, setting the mode to grid. Increase the size (2000x2000x2000 in this case) to spread the cloned models out from each other.

with cloner selected, go to mograph and select random = so that the positions do not look uniform. Can then play with the rotation in the random modifier to also do this.

set the aspect ratio in the render settings to panavison for a super-widescreen look.

create the camera and set the focal length to 50 = flatter

do the camera animation as you wish -> think about the framing and composition.

disable cloner view (easier to work with one in the viewport)

create new material -> set the colour to a darker colour with low saturation. Disable specular and enable luminance. Luminance is where we will have the pulsing effect. Set the texture to gradient, then jump into it and set the gradient type to 3D spherical. Set the gradient to have a stripe of white against black on one end, and then set key frame for the gradient, then move the white stripe to the other end of the ramp and set a key frame 30 frames in the timeline. Can then animate this later by making it repeat to give pulsing effect. Turn start to 0 and radius down so that you only see one stripe on the model pulsating through rather than multiple stripes. Can change the white stripe colour to a different colour.

turn the luminance texture to layer, and add a fresnel on top of the gradient by going to shader. Set the colour you want it to be glowing.

copy the fresnel channel, then shader and paste channel. Half the saturation. Increase the banding in the ramp. Set the blending mode to be screen so its a lighter edge. Set the gradient to screen and put between the two fresnels layers.

Figure 3

14. go to bump and set the texture to noise. Dive into the texture and set the noise type to FBM. Have 2 octaves and set the space to be UV (2D).

Figure 4

15. Make adjustments where necessary e.g. to the colour or noise.

Figure 5

Figure 6

16. Go to render settings and enable depth of field. I also set glow.

17. Select the camera, go to the details section and select DOF Map Rear Blur - blurs the background while keeping the closest option in focus.

18. To repeat the pulse animation, select the material, go to window -> timeline, select the gradient keyframes, then set after to repeat. Can bump up the repetitions if necessary.

Figure 7

Figure 8 - final render

Reflection and Evaluation:

I should have set the camera interpretation to linear so that the camera animation would stay constant throughout rather than easing in and out.

With this type of infographics, I feel inspired to create something more complex to convey information.

Bibliography

Amoeba Sisters (2019) DNA Replication (Updated) [video]. YouTube. Available at: https://www.youtube.com/watch?v=Qqe4thU-os8 (Accessed: 26 January 2026).

Amoeba Sisters (2016) Mitosis: The Amazing Cell Process that Uses Division to Multiply! (Updated) [video]. YouTube. Available at: https://www.youtube.com/watch?v=f-ldPgEfAHI (Accessed: 26 January 2026).

Helix Animation Portfolio | Medical Animation & Scientific Storytelling (no date). Available at: https://www.helixanimation.com/portfolio (Accessed: 26 January 2026).

‘Hepatitis Delta Virus Infection Scientific Animation’ (no date) Random42. Available at: https://random42.com/scientific-animation/hepatitis-d-infection (Accessed: 26 January 2026).

Thom Leach (2022) Mitosis Animation [video]. YouTube. Available at: https://www.youtube.com/watch?v=7ybxaYhRpIA (Accessed: 26 January 2026).

yourgenome (2015) DNA replication - 3D [video]. YouTube. Available at: https://www.youtube.com/watch?v=TNKWgcFPHqw (Accessed: 26 January 2026).

[selinasharmapractice1]

Assignment 4: Memory Palace - Data Visualisation Research

Science Behind Data Visualisation

Data Visualisation is all about making data into clear and easy visual information. I began my research with Maven Analytics' video on The Science of Data Visualisation (2024). Below are my notes on the topic:

The 10 second rule -> presenting data that can be easily read and understood in 10 seconds

Data visualisation isn't all about pretty charts

Using simple visual attributes brings emphasis and commands attention from the audience -> simple design keys

Brain built to see clear patterns and visual cues to analyse data

Raw data -> data visualisation -> data storytelling

What type data are you working with?

The type of data you're working with often determines which type of visual will best represent it.

Financial

Textual

Survey

Time-series

What do you want to communicate?

Comparison

Compare values over time or across categories

Bar chart

Column chart

Line/area char (time series)

Composition

Break down the component parts of a whole

Bar chart

Column char

Tree map

Distribution

Show the frequency of values within a series

Histogram

Density plot

Heat map

Relationship

Show correlation between multiple variables

Scatter plot

Bubble chart

Heat map

Correlation matrix

Prioritise simplicity and clarity over aesthetics and design

Who is the audience and what do they need?

The analyst

Wants details. Will respond well to tables and combo charts.

The manager

Wants summarised data that leads to clear insights. Charts and graphs.

The executive

Needs high-level, crystal clear KPIs to track business health and topline performance.

Don't use wrong chart type for you data

Don't have noise and distraction

Don't have lack of focus in your data visualisation

Don't prioritise aesthetics over function

Don't have misleading narratives, exaggerating data to make it appear more significant

Don't omit context

Context is key

Data vis -> graphic representation

Data storytelling -> creating narratives to communicate insights

Figure 1 - Infographics (Maven Analytics, 2024)

I then watched David McCandless' talk on The beauty of data visualisation (2012). Here are some key points from his talk:

Using visual cues to sees patterns and relationships between numbers

Turn data into a visual landscape

Design about solving problems and providing elegant solutions

Reflection

This is a good starting point for understanding visualisation. I plan to take this information forward in my research by implementing the teachings in my assignment.

Bibliography

Maven Analytics (2024) The Science of Data Visualization [video]. YouTube. Available at: https://www.youtube.com/watch?v=4cq6oUjdPRs (Accessed: 26 January 2026).

TED-Ed (2012) The beauty of data visualization - David McCandless [video]. YouTube. Available at: https://www.youtube.com/watch?v=5Zg-C8AAIGg (Accessed: 26 January 2026).

[selinasharmapractice1]

Assignment 3: Trompe L'oeil - Compositing

Compositing is essential in VFX to put all of our shot elements together. The final shot should have everything integrated seamlessly to achieve photorealism.

Rendering

I started by separating the objects into render layers. This allows for greater control during compositing (Figure 1).

Figure 1 - render layers

The pink bottle has transparency. To have what is behind the bottle rendered without having the rest image plane visible to the camera, I turned 'primary visibility' off in arnold -> visibility (Figure 2).

Figure 2 - attribute editor in Maya

Figures 3 and 4 are the results of separating the elements into render layers. As we can see in Figure 3, the plate is visible through the bottle but the rest is not visible to the camera.

I rendered them out as EXRs.

Figure 3 - pink bottle render layer

Figure 4 - white bottle render layer

With these elements prepared, we can bring them into a compositing package to combine with the backplate (Figure 5).

Figure 5 - backplate

Compositing

I started by bringing in all the elements into After Effects (the gaussian splatted shoe bottle, the 3D pink bottle, and the 3D white bottle). From here, I positioned them where I wanted them in the scene and started planning what I needed to do (Figure 7).

Figure 6 - bringing the elements in After Effects

Figure 7 - planning how to integrate the assets

My plans were to:

sharpen the edges of the shoe bottle (gaussian splats are good for baking lighting and colour information, but they are made up of point clouds which can give a fuzzy look to edges).

soften the edges of the pink bottle and the white bottle (3D renders are mathematical which can give harsh edges, which we do not often see in photos).

colour correct the assets to fit the environment (the shoe bottle is slightly on the darker side, the pink bottle is washed out, the white bottle is over-exposed).

do some paint and roto work for the overlapping assets (the shoe bottle behind the translucent pink bottle).

I began with colour correcting the assets using levels to isolate the RGB channels and match the values (Figure 9). I also adjusted the:

brightness

contrast

hue

saturation

lightness

to colour correct the assets, matching them as closely to the environment as possible.

Figure 8 - before adjusting the levels of the pink bottle

Figure 9 - isolating RGB channels and adjusting the levels for each asset

Figure 10 - after adjusting the levels of the pink bottle

When it came to the white bottle, I noticed a harsh divide in colours on the white of the bottle. This was due to my topology of the mesh, which was not spaces out enough in that region, therefore creating a banding effect rather than a smooth effect. Rather than going back into Maya and fixing the topology, I decided to paint it out (Figure 11). This is more practical than sending it back in the pipeline, although it is best to ensure topology is good before sending assets down the pipeline to prevent extra work.

Figure 11 - colour adjusting and paint work on the white bottle.

Next, it needed to make the shoe bottle visible through the translucent pink one. I did this by duplicating the shoe bottle and rotoscoping the missing shoe part (Figure 13), then placing it on top of the pink bottle and changing the blending mode as well as the opacity. To give the effect that the roto'd element is being viewed through a translucent object, I added camera lens blur to the object (Figure 14).

Figure 12 - shoe not visible behind pink bottle

Figure 13 - roto work for the show to be visible behind the pink bottle.

Figure 14 - Blurring the roto'd work through the bottle

Smaller details included blurring the 3D objects to give smoother edges (Figure 15), trimming the soft edges of the shoe bottle, and adding shadows beneath the assets.

Figure 15 - softening the harsh edges of the 3D with camera lens blur

Evaluation + Reflection

The final result can be seen in Figure 16.

Unfortunately the Jawset Postshot logo is watermarked as I do not pay for the full license. For future reference, I can buy the full license or I can model and texture the shoe bottle. Ultimately, I wanted to put my studies of gaussian splats into practice with this exercise to see how practical they can be. The results are photoreal because gaussian splats bake light and colours.

The 3D assets can be improved by adding scratches and dust to the textures. Furthermore, the white bottle should have more specularity as it is meant to be a shiny bottle but looks matte here. It also should have had some subsurface scattering as light is able to slightly pass through it. These factors would have added to photorealism.

Figure 16 - Final image

I would like to take the time to learn Nuke as it is the industry standard for compositing in VFX. I would also like to try 3D integration directly in compositing packages rather than just dealing with renders to allow for greater flexibility. I tried bringing the 3D assets into After Effects, but they would come in without textures because I had created them procedurally in Maya. After Effects can only bring in textured 3D assets if it has image textures (texture maps).

Bibliography

Adobe Video (2025) After Effects Guide: Advanced 3D Edition | Adobe Video x ‪@filmriot‬ [video]. YouTube. Available at: https://www.youtube.com/watch?v=9LxKSsftUS4 (Accessed: 20 January 2026).

David Bittorf (2023) Maya Color Mangement sRGB vs Raw (maya tutorial) [video]. YouTube. Available at: https://www.youtube.com/watch?v=qt_TR-7HcGM (Accessed: 20 January 2026).

Import and add 3D models to your composition (2025). Adobe. Available at: https://video.tv.adobe.com/v/3425759?end=nothing&learn=on (Accessed: 20 January 2026).

Masd*Lab (2024) How to Import 3D Objects into After Effects [video]. YouTube. Available at: https://www.youtube.com/watch?v=B8B7tr7ProQ (Accessed: 20 January 2026).

ProductionCrate (2025) Learn VFX Workflows - After Effects and 3D Tutorial! [video]. YouTube. Available at: https://www.youtube.com/watch?v=sGjODH6e2BM (Accessed: 20 January 2026).

ShiveringCactus: VFX (2025) How to Import 3D Models into After Effects | OBJ, GLB, FBX [video]. YouTube. Available at: https://www.youtube.com/watch?v=CvpXU7TIDfw (Accessed: 20 January 2026).

WooshinsEdits (2025) How To Import 3D Models Into After Effects (Easy) [video]. YouTube. Available at: https://www.youtube.com/watch?v=echnA5skm9w (Accessed: 1 December 2025).

[selinasharmapractice1]

Assignment 3: Trompe L'oeil - shooting the HDRI and texturing the 3D models

HDRI capturing

To make 3D models fit in a plate, you need to get the lighting information that was in the environment (part of data wrangling). This is best captured using an HDRI.

Using my previous HDRI tests, I did the panorama method and then planned to stitch them together (Figures 1-5).

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

The problem I have found with this method is the auto-adjustments that the phone camera makes to change the lighting. As you can see in the panos above, the ones that capture the walls of the environment auto-adjust to be darker. This causes inconsistencies in the lighting information as you stitch them together.

Seeing how this will cause issues, I decided to try HDRI apps on the iPhone. Polycam has a 360 scanning feature (Figure 6) which can then be projected onto a Skydome in a 3D package (Figure 9).

Figure 6

I tested this feature and found that it is unable to capture all details in an environment (Figures 7 and 8), so it substitutes lost information with AI generated imagery to fill in the gaps. This is good enough for pre-vis, tech-vis or post-vis purposes, but for final outputs this is not the best tool to capture an HDRI as it will be missing lighting data for lighting artists. However, this technique works better than the previous method of stitching iPhone panos together as you actually get more consistent lighting information.

Figure 7 - Polycam results

Figure 8 - Polycam results

Figure 9 - Polycam results put onto a Skydome in Maya

The best equipment for onset data acquisition will be (Figure 10):

Canon 5D (to shoot 7 bracketed exposures)

8mm Sigma fisheye lens (to capture a 180 degree circular image when used on a full frame camera)

Ninja Nodal 6 panoramic head (to incrementally rotate the camera around)

Grey ball (18% grey ball (mid grey) to match exposure in post)

Chrome ball (to give a clear idea of where the lights were onset)

Macbeth chart (for colour calibration in post)

Slate (to give a reminder of when and where the HDRI was taken)

Manfrotto Tripod (for shooting under difference environments)

Shutter release cable (to fire the camera without touching it, allowing you to minimise camera shake)

Batteries, memory cards, gaffer tape, etc.

(Amin, 2020)

Figure 10 - best equipment for capturing an HDRI (Amin, 2020).

More instant equipment is the theta360 and the insta360, which are devices that capture the data for you. However, they are also not the most accurate either (although their results are good enough if you are short on time onset).

For my own demonstration purposes, I did not have a Macbeth Chart but I had a make-up palette lying around that I used as colour reference (Figure 11). This won't be as accurate as a Macbeth Chart, but I thought it could be decent reference for this exercise.

Figure 11 - Using a make-up palette like a Macbeth Chart for colour reference

Texturing

I will be treating the Trussardi bottle as two separate models because of their different texture requirements - the metallic bottle cap as one model, and the white body as another model. The third model in the scene is the pink bottle.

With the HDRI in Maya, the 3D models can be textured in the scene. Having the HDRI and lights set up make texturing a more accurate process as we know they are going to be in that lighting environment.

A HDRI alone does not give a photoreal look to objects (Amin, 2020). Therefore, lights need to be set up manually too (Figure 12). Having a grey ball and a chrome ball reference is best for this step (ibid). The chrome ball tells you where the light is coming from in the scene, while the grey ball tells you about light intensity and temperature.

Figure 12 - manual light set up in Maya

The pink bottle is a glass material with a pink gradient. The top of it is a metallic gold material. Understanding the material information, I set up my textures in Maya accordingly (Figure 13).

Figure 13 - pink bottle texture nodes

Figure 14 - pink bottle textured

The white bottle has a logo on it, so I UV'ed the bottle and recreated it in Photoshop (Figure 15). Meanwhile, I textured the gold bottle cap simply by making it metallic and gold (Figure 16).

Figure 15 - white bottle UVs in Maya

Figure 16 - textures in Maya

Figure 17 - Arnold render

Figure 18 - isolated channel (R)

Figure 19 - isolated channel (luminance)

The exr render from Maya (Figure 17) shows that the pink bottle's textures go well with the environment's lighting and with the plate. The white bottle on the other hand looks out of place against the plate. Isolating the channels, we can see the values and get an idea of what is causing the issue. The judgement I can make here is that the bottle model is too bright against the plate, and the metallic cap model on it is too saturated when compared to the other metallic caps in the plate. We can also see that the pink bottle is too dark against the plate when we isolate the luminance (Figure 19), which we couldn't catch at a first glance of the beauty pass. This is why it is good to isolate channels to check the values of your models against the environment.

Furthermore, the models are too smooth in texture. This makes the models less photoreal. I will attempt to fix this in compositing by adding noise and grain, but next time I should add these textures to the materials.

Evaluation and Reflection

When it comes to compositing, I will adjust the models' values to fit the plate and make them look more like they belong in the environment. I will also add grain/noise to the models so that they look less mathematically perfect.

Bibliography

Amin, J. (2020) ‘Data Acquisition Training – CAVE Academy’, Shooting a HDRI Onset for CG Lighting, 22 May. Available at: https://caveacademy.com/wiki/onset-production/data-acquisition/data-acquisition-training/shooting-a-hdri-for-lighting/ (Accessed: 17 January 2026).

Maya Help | Model glass or metal with the Standard Surface shader | Autodesk (no date). Available at: https://help.autodesk.com/view/MAYAUL/2024/ENU/?guid=GUID-426580A4-6079-46A2-B651-E3CE38A5DEB8 (Accessed: 20 November 2026).

[selinasharmapractice1]

Assignment 3: Trompe L'oeil - 3D Modelling

Gaussian splatting a transparent object can result in artefacting which can affect the photorealism in a shot. This can be observed with the pink perfume bottle in Figure 1.

Figure 1 - Gaussian splatted pink bottle with artefacts

If there is obvious artefacting, this can throw off the believability of a shot (see Figure 2).

Figure 2 - Gaussian splats inserted. The pink bottle (left) borders unbelievability when we notice the artfecating.

This is why, while gaussian splatting is a very powerful modern tool, it is important to still know how and when to model objects for the demands of a shot. Therefore, I decided to model the pink bottle in 3D. The modelling process can be observed in Figures 3-6.

Figure 3 - image plane set-up in Maya

Figure 4 - 3D block-out of object

Figure 5 - detailed model

Figure 6 - wireframe of final 3D object

The next object to model and insert was Figure 7. The shape of the bottle was simple, while the cap of the bottle required more precision due to the organic yet planar design. To achieve this, I tried to keep it low-poly to achieve the planar effect. The modelling process can be observed in Figures 8 and 9.

Figure 7 - Trussardi perfume bottle

Figure 8 - Trussardi bottle 3D model

Figure 9 - Trussardi bottle with cap 3D model

The 3D models were finally ready to be textured. I will be using the Maya Hypershade to texture these objects as they are very simple.

Figure 10 - The 3D models

Figure 11 - wireframes of the two 3D objects

Reflection and Evaluation

Next, I will be using the HDRI I shot on set by importing it into Maya so that we can see how the textured objects react to the lighting scenario if the environment. While the gaussian splatting method created artefacting that affected photorealism, the difficulty of using 3D models will be how perfect and smooth the objects will look. The key to photorealism is not perfection, but instead immersive storytelling. I will have to find a way to break the cold computerised textures to tell a story of a perfume bottle that has been used frequently so it will be a bit sticky and a bit dusty.