Ideation, Concept and Research
For my major project I worked with Aveen Fitzpatrick and Odhran McManus in my team. Because we’ve worked together on previous projects since first year, we felt that we’d be able to communicate more openly and efficiently with each other from the beginning of the project.
The first thing we discussed was whether or not we wanted our short film to be 2D or 3D. As a 2D artist, I was hesitant when the rest of my team said they wanted to make a 3D animation, but I wanted to challenge myself to become more comfortable with this medium as this would allow me to improve my modelling and animating skills.
Before moving on to forming the actual narrative of our film, we took about 10 to 15 minutes to discuss what 3D animated movies we admire, and which styles that those films portray do we want to influence our style. We all seemed to circle back to the films ‘Entergalactic’ and ‘Into the Spiderverse’ due to their comic book and hand-drawn appearances despite the fact that they were 3D. We were particularly inspired by the graffiti montage in the first act of ‘Into the Spiderverse’, where 2D effects (such as motion lines or annotated sound effects) were layered on top of the 3D rendered footage to create the illusion of the viewer watching a moving version of a graphic novel.
Since the spray painting scene stuck with us the most, we decided we wanted to implement this as part of the main narrative in our film. We then began to come up with our story. We settled on three protagonists; Leo, Zack, and Dani; three young creatives who meet each other in an abandoned yard through an alleyway, which Zack and Dani have decorated with graffiti and makeshift furniture and now use as a den/skate park. Leo, having not yet met the other two protagonists, sees a cloud of coloured smoke coming out of the alleyway as he’s looking out of his classroom window. Upon following the source of the smoke, after class he enters the skatepark where he meets Zack and Danni, who show him their graffiti art and offer him a spray can. Leo shies away nervously as he doesn’t feel he is capable of creating to the same extent as them, but after watching them work, he gets the confidence to try spray painting with them. After a while, he stands back and looks proudly at his work with his new friends, happy in the knowledge that he had taken a step to coming out of his shell.
After coming up with our story, we spent the next two weeks storyboarding, creating concept art and researching some of the films that we wanted to inspire our desired art style. We each took about 1/3 of the storyboard each and made thumbnail sketches for each shot.
In our pitch presentation, our lecturers advised us that the beginning scene where Leo was at school seemed like an unnecessary environment to build for such a brief scene, and we were lacking an explanation as to why Leo was initially shy to join in spray painting. After this feedback, we made some revisions to our storyboard. We instead had Leo walk past a poster for an art competition on his way to meet his friends at the skate park, sparking a drive to create a drawing to enter before the submission deadline. He sits at the park waiting for Zack and Dani (we changed the story so that they already knew each other and had arranged to meet there) and starts drawing in his sketchbook, but he is never satisfied with what he’s drawn. Frustrated, he tears out a page and throws it in the bin behind him, at which point Zack and Dani walk up and suggest street art as a new medium to use for his submission. In learning to experiment with graffiti, Leo is able to let go of his need to pursue perfection in his work, and he can create chaotically and freely. After a montage of his graffiti process, he looks proudly at his work with his friends, who take a photo of his mural and post it into a mailbox at the end of the film.
My main role in pre-production was concept art for our three main protagonists, as well as additional roles in prop modelling. Before designing our characters, I did some research into the concept art of ‘Into the Spiderverse’ to get an idea of the style of the illustrators at Sony who designed it. I also used a grid overlay on their concept designs to reverse engineer how they may have begun their drawings, in terms of eye-spacing, nose length, mouth size etc. After redrawing some of their concept art from scratch using my grid for reference, this allowed me to gain a better idea of how to draw our characters in a similar style.
For the main protagonist Leo, I designed a character with chin-length wavy hair; a style I’ve seen often among the art students on my own campus. I gave him a light grey hoodie layered under a charcoal grey jacket to suggest that he was kind of hiding himself underneath his layers. I finished by giving him baggy washed-out jeans and plain brown shoes.
The reason I kept his colour scheme so desaturated was to contrast with the more brightly coloured designs I was planning to include in the design of the remaining characters; Dani and Zack. I also wanted to highlight that Leo didn’t wear any form of clothing that really expressed his individuality, as the narrative of our story was to see him using street art to come out of his shell, so when we first meet him, his style is not particularly self-expressive.
Zack, the second male character, was more enjoyable to design as I was able to play around with colour and shape language a little more than with Leo, however, when I drew some concepts of Zack’s face for practice, I realised I had unintentionally made him look more like a more saturated and visibly happy version of Leo, so I ended up changing this to make the characters contrast more. To show that Zack was more outwardly expressive of his creativity, I decided to give him thick dreadlocks tied up in a messy bun and dip-dyed green. I have also seen this hairstyle as well as people with this hair colour while walking through the city centre, but only on occasion. I love looking at all the different hairstyles and colours of hair that I see around Belfast, so implementing something similar in my character design was very enjoyable. In the initial sketch where I blocked out Zack’s body, I used trapezoid and rhombus-like shapes as a guideline for his arms and torso to give him a well-built physique, showing that his stature and body language had more confidence outwardly than Leo. I also drew him in a T-shirt with a motif and a pair of jeans, both of which had paint splatters on them to show his love of street art.
My approach for drawing Dani differed slightly from drawing Zack and Leo, as I had to use the right shapes to block out her character so that she wouldn’t share their masculine build. I did separate studies of the concept art from ‘Into the Spiderverse’ for both the male and female characters, and I was able to then identify which areas to focus on which would portray a slimmer feminine form. I sketched out the face shape in an oval which narrowed at the bottom for the jaw, while the rest of the limbs were structured with rectangular or triangular shapes, connected at the joints by smaller ovals. The rectangular shapes I drew also narrowed at the bottom to show the smaller areas of each limb without creating the appearance of a thick muscle.
Once I had Dani’s body sketched out, I decided that I wanted to style her appearance and clothing in a way that was inspired by some of the styles that I admire or have wanted to try out myself at one point or another. I gave her light brown hair dip-dyed lilac with a loosely fitted beanie hat. I gave her a paint-splattered round-neck T-shirt with a motif of a Green Day logo on it as they’re my favourite band. Underneath that shirt was a second, longer-sleeved shirt which was tucked into a pair of long Bermuda-style denim shorts. I also added a bright checked shirt tied around her waist for a pop of colour and to convey the ‘alternative artistic skater girl’ aesthetic. I finished by giving her knee-high white socks and black ankle boots, also splattered with paint.
Pre-Production
To ease myself in to modelling, I picked some small items to make in Maya which may be seen in a skate park or abandoned park, such as cans of paint, aerosol spray paint cans and a roundabout. I also made a prototype of a halfpipe to get a sense of how we’d go about creating that shape. I was glad to have made the spray cans early in the pre-production process but could have saved time by scrapping the roundabout because by the time we had our environment set up, we realise it would serve a purpose or be shown in any of our shots. I did enjoy experimenting with sweep meshes and deformed shapes to create the curved barriers of the roundabout despite this.
We quickly realised Odhran had a particular prowess for modelling on Maya, as he was able to work through our asset list the quickest out of all of us, so after consulting my team, I moved on to begin character sculpting. I used Blender for this and was initially worried because I had only ever used Blender for one project last year, and the character I created for that assignment wasn’t particularly adventurous in terms of physique (I made a round baby penguin), so I wasn’t feeling particularly confident in sculpting three characters from scratch.
I went back to some of the pre-recorded videos on sculpting from our second-year classes, which helped me figure out where to start. I used a sphere to start off the head shape, then stretching it using the elastic deform and inflate brushes, then roughing out the areas if the face that would be more prominent- the nose, cheeks, chin and forehead. Once smoothed out, I’d remesh the modified sphere, then repeat the process to add more details while smoothing along the way. I also added a marionette line carved out under the nose which I then sculpted out to make the lower part of the jaw look more realistic. For my first time sculpting a human face, I felt it went well, and once I had Zack and Leo’s heads sculpted, I started to block out their bodies. Aveen offered to help out with the character sculpting process by making Dani’s model, which I was very grateful for, because I started to feel under pressure trying to sculpt three characters by myself. Aveen followed a different approach to character modelling, creating and extruding planes over a reference image of a face in order to create the mesh. At the beginning of semester two, we had our production progress presentation, and on my slides, I talked through my sculpting process following the tutorial videos. The feedback from my lecturers was that they were ok but they seemed to be making the characters look old with loose cheeks or lacking detail. They put it to me that I was trying to ‘learn sculpting from square one’, which was not an advisable strategy to be following weeks into the production process. This was true as I had given myself extra time to try and fix the models faces, meaning that I was falling behind on finishing the sculpt for the rest of the character’s bodies, and we had yet to start animating.
This was a difficult fact to face as I was really hoping to be able to materialise an actual model that would capture my concept art, but I felt I was letting down the team by struggling to finish a sculpt. Our lecturers suggested to us that we use free-to-download base meshes which we could then manipulate and add clothes and hair to in order to quickly create character models so that we could finally move on to the animation process. This was disheartening but we felt it was the right move in order to catch up with the schedule we had set for ourselves in pre-production.
Once I stretched some of the vertices and edges of the base mesh models for Zack and Leo, the process went much faster. I was able to use the hair I had sculpted for my initial models and add it to the base mesh, and once I had finished re-shaping their faces, Odhran took the models and expanded certain parts of their geometry in order to create the appearance of a clothing layer. Unfortunately, since we were now behind schedule, we didn’t fully get to create the layered-hoodie outfit originally planned for Leo in the concept art, so both male characters would now be wearing jeans and a fitted T-shirt. Aveen used a separate female base mesh for Dani which turned out lovely, but the time frame we were working with led to the final Dani model deviating from the original concept art too.
I was disappointed that we couldn’t achieve the look we planned for with our models, but the most important thing now was that we started animating now that Odhran had rigged our models (this was one of his main roles and he did a great job; he even added advanced skeleton so that we could use more fluid movements to cup and spread the fingers on our models’ hands). I unwrapped the models for Aveen to import into Substance painter where textures would be added. I didn’t find it too difficult as the base mesh template had already assigned a UV layout for the face and arms, and I was able to unwrap the clothes ok. I found it difficult to unwrap the hair though as it was an irregular sculpted shape, and I sometimes struggle to know where to begin in terms of choosing which edge to cut, and because we were behind schedule, I used the camera-based view on the UV map to help me know where to separate certain parts of the hair, then unfolded those segments automatically. It wasn’t the most efficient way to work, but it seemed to work ok when it came to texturing, but this is something I would like to practice more.
Production – Animation
We ended up each taking one character to animate so we could split the workload evenly, and I chose to animate Leo because the main segment of our animation that I wanted to focus on was his graffiti montage sequence. We had discussed what this scene would entail and expressed to each other our concerns on how to create the visual effect of spray paint in Unreal Engine as well as combining those shots with an inverted layer mask which would make it look like the graffiti was ‘appearing’ where certain parts of the mask were erased. We began to feel pressed for time especially after falling behind with sculpting, re-sculpting and eventually moving to base mesh tweaking and texturing those instead. Because of this, we decided to come up with a new way to show Leo using graffiti which wouldn’t complicate the effects we used as much. Instead of action shots of him doing graffiti, we decided to have Leo enter a dream sequence, where we see him floating and swimming through a starry night sky. Suddenly, he spots two cans of spray paint, which he grabs and sprays using the paint jets to propel him around this night sky. We thought that it would still allow us to create a nice scene by shooting a more ethereal sequence in which Leo could move freely, to further highlight him letting go of his perfectionism and embracing the unknown. I was also excited to experiment with using particle effects for the first time to create the jets of spray paint, and a space environment would allow me to play around with lighting too, which I’ve always enjoyed.
After re-planning the graffiti sequence, I then worked through an animation checklist I had set myself; a walk cycle which would be used in the beginning and end of the film, and an animation of Leo sitting down on a bench and reaching to grab his sketchbook. This was a rough animation, and since Odhran was working on that scene, he offered to take the animation and finish it off so he could add his and Aveen’s animations alongside it.
My main animation checklist consisted of:
- A swimming breaststroke cycle to use as he moves through space
- ‘Swimming’ on to his back and floating idly as he smiles
- A look of sudden surprise as he sees the spray cans
- Stretching out his hand to grab a spray can
- Examining the spray can and smiling as he ‘floats’ upright
- Holding out two spray cans and gliding around
(A link to these files will be included in my submission link)
I feel like I could have done a less generic walk cycle for Leo, because the frames I referenced from our ‘Principles of Animation’ module in previous years provides the keyframes for a casual/confident walk style, whereas it may have been better to have a looser or slower walk to better fit in with Leo’s lack of faith in his abilities. Even something as small as his body language being more hunched over could have helped to portray that more visibly from the start of our animation, and if I had more time, I would have spent longer on that walk cycle to make it look more fluid and less generic.
For some ideas on poses for floating and freefalling animations, I imported an FBX file of my Leo model to Mixmo, where I was able to test out some of their keyframe presets for inspiration. I feel like this helped a lot because it allowed me to see what motions were used for what actions, how the arms and legs moved and how to portray idleness using subtle movements. Being able to see a reference of my FBX in front of me as I worked really helped me to stay focused and driven throughout the animation process.
If I were to do this project again, I would love to explore how I may have been able to use cloth physics in Unreal Engine to add more dynamic clothing to the characters, because, as we had just extruded certain areas of the geometry to create fitted clothes, some parts of the mesh corresponding to Leo’s skin would jut out through the clothes depending on what way you manipulated the rig. Additionally, I found it hard to make the animations as fluid as I wanted them to look, because if you raised or rotated Leo’s arms a specific way, the geometry of the clothes would get this spiky, stretched appearance where the vertices were further apart, so I tried to limit some of his raised and upper arm movement to avoid this, which in turn made the animations a little more stiff, so that’s definitely part of the pipeline I’d like to research further over the summer to see how else we could have approached it. I really admired how the facial rig Odhran had made worked, as I had never animated a character who was able to actually use facial expressions to show emotion, which was something I enjoyed experimenting with.
Production – Environment and Lighting
After animating, I began to assemble the space environment in Unreal Engine. The first thing I began looking in to was how to create the appearance of a starry sky, and by following a YouTube tutorial, I was able to import a sphere from Blender, invert the faces of the sphere and apply an 8k milky way texture to the material, to make a kind of starry sky sphere which worked really well. Unfortunately, something in that level seemed to have become corrupted the following day when I opened the project, so I took a different approach.
I went on to the UE Marketplace and downloaded a free set of blueprints called ‘Good Sky’. This allowed me to create a sky sphere that could change light intensity and colour depending on time of day, but could also project stars across its UV tiles, which was a lot easier than trying to invert the faces of an imported sphere only to have it corrupt itself.
I used a directional light for the sort of overall ambient lighting in the scene and gave it a purple-lilac hue to compliment the starry sky dome. It was sometimes difficult to know which direction to rotate the light in terms of when Leo was added to the environment as well, because I wanted shadow to fall on him a certain way, yet at the same time I didn’t want his body to be too hard to see when I lowered the intensity or changed the angle. It was also difficult to know which way to rotate especially since there was no asset like a window or lightbulb which would be contextually generating a ‘light source’ as such.
To show the importance and other-worldly nature of the spray cans as they entered the scene, I added a point light to the level sequencer, which increased in intensity as the camera zooms in on Leo’s surprised face, creating the impression of a warm glow being emitted from the cans, contrasting the cool tones from the rest of the environment. I initially tried editing the material nodes of the spray cans by plugging an additional base colour node into the emissive colour input as I thought it would add a glowing aura to the cans, but instead it just turned up the saturation and brightness which made the can look washed out, so I decided against this.
Because of the fantasy-style environment this sequence was taking place in, I wanted the graffiti cans to spray something that looked a little more surreal than just coloured steam clouds, so I watched a tutorial on Niagara Systems and how to create emitters from scratch to make particle effects. This felt difficult to wrap my head around at first, because when creating an emitter, you have to manually clarify the particle spawn rate, wind curl noise, velocity etc. and with there being so many options for how much these attributes affect the spawning particles, I quickly began to feel overwhelmed. However, after following the tutorial on how to create a basic spark emitter, I ended up with a trail of small particles which I assigned a colour node to so that they would still show the colour of the paint while they glowed. I added a long lifespan to them so the particles would linger after being emitted, as well as a drag that would keep the trail of ‘spray’ a consistent thickness when trailing behind Leo, rather than thinning out quickly in accordance with how fast the animation was moving between its keyframes.
https://youtu.be/04k9JDx-KTM?si=_GADAdqBgE-HAJ6f
We struggled to find a way to animate props in Maya in terms of adding parent constraints between the characters and the spray cans, but when assembling my scene in Unreal, rather than having to key every movement of the prop to follow the motion of Leo’s hand, I was able to add the prop to the level sequencer and use an attach track to clip it to follow the movement of any joint in Leo’s skeletal mesh. This allowed me more fluidity in shots where Leo grabbed or held the can as he was floating; even if he only moved slightly, the can would move slightly with him, so they didn’t look like two separate assets out of sync with each other.
Using the same method, I was also able to attach point lights to the nozzles of each spray can, to make it look like the concentrated spray was glowing as it came out of the can. As for moving the animation sequence around the environment, I took a lot of inspiration from the fire extinguisher scene in Pixar’s Wall-E (2008), where Wall-E uses the extinguisher jets to propel himself around in space. Because the sprays were two separate colours, I also wanted to make Leo’s model spin so that the two spray colours would cross over one another. If I had have started testing various particle effects in Unreal Engine earlier in the production process, then I may have worked to create an emitter that expelled a higher concentration of particles, or at least did a trial run of the spray paint effect using a steam or gas preset as opposed to a spark emitter. After setting up several different shots for my segment (which totals 1min 37s out of the whole runtime of our short), I sent those shots to Aveen who added them to the Adobe Premier sequence.
Aveen and I also browsed the website Uppbeat to find some music for the graffiti sequence. I was looking for an instrumental piece that sounded relaxing and hopeful, and we found one which happened to be named ‘Hope’ which felt perfect for the scene, especially with the violins beginning to swell as Leo reaches for the spray cans.
Conclusion
Overall, I am pleased with how our short film turned out, but I think we spent too long in the later stages of pre-production re-doing things like sculpts and rigging which led us to fall behind in our schedule. Once you start to see yourself fall behind, it’s very difficult to remain confident in your abilities, which also takes a toll on your motivation to try doing things again if they don’t work out in the way that you hope. I think we could maybe have considered base meshes at the start, then focusing on how to make the hair and clothes realistic to compensate for using a model which wasn’t made from scratch. Aveen and Odhran were both amazing to work with as we had a great rapport and emphasised on communicating with each other between classes. When I felt myself starting to fall behind after the sculpts and I felt like I was letting the rest of my team down, Odhran and Aveen encouraged me and helped polish the meshes by modelling shoes and extruding the mesh to create clothes while I focused on the faces and hair. It was also a valuable experience in working with Unreal Engine again, as I had not used it since our 3D environment project last year, but the real-time render was extremely beneficial to our workflow, and it was amazing to have been able to make a particle effect system for the first time. That may be a small thing, but it’s something I never really saw myself as capable of doing until this semester, so I’m grateful we chose this software for our render.
Please find below a link to our finished short film, ‘Art Block’:
Elsewhere on my UU Blog site you can find an image portfolio of my screenshots and concept art for this project:
https://blogs.ulster.ac.uk/ornamccreesh/2024/05/12/image-portfolio-art-block/