This blog will detail from start to finish, the pipeline of myself and my team’s pipeline for our final major animation project. Spanning over 8 months, I will go over my own tasks and how they are relevant to the pipeline, shared generalist roles and how our differing roles melded together.
A Separate portfolio showcasing final works will not be shown on this blog but instead on my LinkedIn account as it has many links to this blog and also because all of the final work won’t neatly fit in one post. The account can be viewed in the link below:
https://www.linkedin.com/in/timothy-magilton-3a9928269/recent-activity/all/
Updated Interactive Portfolio:
https://sketchfab.com/NewtEmporer
Team Members:
Jonathan Graham
Mark Smith
Timothy Magilton (Myself)
Pre-Production:
As any project starts, an idea needs to be drawn up. Many ideas were considered but ended up being either too simple, too complicated, too ambitious or didn’t match up with me and my team’s skill sets. We wanted to create something unique but achievable within the time frame. A good way to achieve a unique look was to mic 2D and 3D together, this idea stayed with us as a team right through to the end of the project. Theme was up next, looking at what other teams were doing, we saw a vast array of themes and wanted to split apart from the rest of the groups to have a project that stood out in its own right. It started with the American Civil War but then went more toward a fantasy Wild West theme. Eventually, we stuck with a Weird West theme, where a bounty hunter is asked by a local preacher to hunt a vampire. The cell-shaded themes worked well with the gritty plot and the dusty colour pallets. We could also utilize our shared skills in 2D and 3D as a team and use the cell shaders to make transitions easier. With the idea sorted, it is on to reference gathering, story development and asset creation.
Below is an evolution of our Miro Board and how it went from random theme image gathering to a much more organized layout:
You can view the Miro Board via the link below:
First Semester & Preparing the Pipeline:
The next step in this major project is to take our brainstorming ideas and our chosen methods of animating and implement them into an organized and easy-to-understand first-semester pipeline. Below is a design log that shows how all of this was achieved. The second semester has a completely different pipeline due to the creation pipeline evolving into an animation one. There was a lot of emphasis on making that pipeline transition as fluent as possible and working in a team significantly aided in this process. The communication, motivated drive to work and desire to work together have often times kept me on top of my tasks and vice versa. With a strong sense of teamwork established, there were some issues here and there that could have been ironed out faster than they were but it did not break the team apart and we stuck through it firmly from start to finish. Overall the sense to make a fluent pipeline and a strong dedicated team allowed for that pipeline transition to be near enough on point. The log below will show this from a practical > theorem perspective:
The First Design Log:
Jonathan: Minimal 2D animation, 2D prop Artist, 2D scenery/background artist, 3D animator and minimal 3D Prop designer.
Mark: 2D artist, background creator, 2D animator.
Tim: 3D Prop designer, 3D character designer, 3D animator, Minimal 2D prop creation and animator.
Idea:
Cowboys Vs. Vampires. Wild West Themed.
Story:
Exorcist puts a bounty out for removing curses from individuals. A well-equipped vampire hunter takes the job. He goes to the possessed person’s house but soon finds out they themselves are not cursed but are a vampire and have been cursing the town they reside in. The Vampire hunter escapes and prepares for a fight but is against the clock as the vampire wreaks havoc on the town.
Environment/Theme:
Wild West, isolated frontier rural quiet town, surrounded by desert, isolated from civilization, time set between dusk and dawn. (Horror thriller themed.)
(Leaning towards Cell shaded style.)
A mix of 2D and 3D (Main characters and props in 3D, lesser props and backgrounds in 2D.)
Characters:
Protagonist:
Vampire hunter, well-equipped, cowboy-themed, Van Helsing-inspired elements.
Antagonist:
Vampire mimicking an outlaw. Hides in shadow, cloaked most of the time.
Side characters:
Priest, to make the pipeline easier, his face is under a cloak, in shadows, surrounded by holy smoke.
Saloon owner, generic bartender attire, western-themed, owns a local saloon and the room being rented out by the vampire but is unaware the occupant is a vampire.
Assets and background creatures: (Can be sourced from third parties to save time on the pipeline:
Revolver – 2D and 3D
Repeater rifle – 3D only
Silver cross – 3D Only
Stake (Vampire hunter weapon) – 3D Only
Holy inspired crossbow (Part of protagonist attire/equipment) – 3D Only
Black Cat – 2D, 3D version optional
Desert fox – 2D, 3D version optional
Bats – 2D Only
Hats (Western themed) – 2d and 3D
Barrels – 2D
Beer steins – 2D and 3D
Tables – 2D and 3D
Bar stools – 2D and 3D
Bottles of alcohol (Various) – 2D and 3D
Saloon doors – 2D, 3D optional
Saloon piano (optional) 2D, 3D optional
Coins – 2D, 3D optional
Frontier town buildings (Built by team or sourced via third parties) – 2D Only
Tumbleweed – 2D Only
Lantern – 2D and 3D
Cactus (Various) – 2D Only
Mixture of 2D and 3D elements
3D props, selected assets and characters
2D backgrounds and less prominent assets such as creatures/filler assets (tumbleweed etc.) and scenery.
OneDrive, Discord and Team Sharing:
As a team, we decided early on that our main method of communication would be Discord as it’s easy to navigate through and is very creator-friendly. Microsoft OneDrive was used to share all of our props, characters, rigs, scenes and updates with one another and to walkways using both apps in tandem. This greatly increased our work ethic as a team and made sure we quickly fixed any errors that occurred.
First Semester Timetable:
| Week 1:
(Completed) |
Brainstorming ideas, creating rough storyboards, experimenting with styles, getting fundamental communication sorted – Discord, G-Drive, Miro-Board, Trello. |
| Week 2:
(Completed) |
Settling on a final story. Settle on a style to create assets and to animate in. (Cell-Shaded) See what everyone’s personal style is like and either combine them or find a compromise and find a specific style that we all can agree upon and create in/with. |
| Week 3:
(Completed) |
Finalize the pipeline, and start making more complex assets. Experiment with a mix of 2D and 3D to have a USP. (Unique Selling Point.) Attempt to make characters and main props ‘pop’ out of the screen to make major events in the animation more recognizable. Begin work on a short animatic, 2D only to save time and is a good backup if the 2D/3D mix is not viable. |
| Week 4:
(Incomplete but not scrapped) |
Consider music and voice lines, and finalize the animatic to paint a clear picture as to what music should be used and what tone and accent the voices should be in. |
| Week 5:
(Completed) |
Settle on a rough script and an original soundtrack. Start making the first 30 seconds of the animation in its final form, if 3D assets are viable, have a mix of Krita animatic software mixed with Unreal Engines animation capture tools. |
| Week 6:
(Completed but not the way we wanted it to turn out) |
Continue working on finalizing the first 30 seconds, push for 40 seconds to a minute if viable. |
| Week 7:
(In Progress) |
Create any remaining props and assets, provided there is time and we are ahead of schedule by even a few days. Characters MUST be completed. |
| Week 8: | Animating – First minute of animation. |
| Week 9: | Animating – First minute of animation. |
| Week 10: | Animating – End of the first minute and to reach a goal of 1 minute 50 seconds completely animated and finalized. |
| Week 11: | Begin using VFX software to completely finalize the start and the middle of the animation. Even if we only reach say 1 minute and 40 seconds then finalize it. This will prevent extra work later on and will reinforce our choices of SFX and if they need to be changed or updated. |
| Week 12: | Sound design. |
| Week 13:
(Christmas) |
Begin animating the middle and final sections of the product. |
A lot of changes occurred a lot during this period and most were to fit in what needed to be in and to shave off any dead weight, especially during crunch times which was usually Monday – Wednesday as those days worked well with all our schedules and we spent the majority of our time together during these times. However, evidence will be provided below to demonstrate our commitment to the plans we had drawn out and ended up being quite efficient. I’ll go over what I did, and how it helped the team and then go onto what changed as the transition from the creation pipeline to the animation pipeline came with its unique issues that we tackled as a group, which is important not for later.
Prop Creation:
Revolver:
This version would go through heavy changes but be never used in favour of more cell-shade-friendly designs.
Cell shading testing for the revolver, shadows didn’t appear as intended in the final render and so was remade.
Silver cross.
Never used due to another cross I made being more character-friendly.
Simple table:
The table was used but then replaced for the same reason as the revolver.
Gold coin:
This asset was never used in the final render due to reflection maps not working.
Silver coin:
The silver coin also had reflection issues.
Dirty gemstone:
Made it into the final animation;
Clean gemstone:
Also made it into the final animation.
Misc table:
Was replaced due to cell shade issues.
Empty table:
Replaced for Cell shading texture errors.
Crossbow bolt:
Another variant I made went into the final animation.
Silver bullet:
Made it into the final animation with new textures.
Gold-capped silver bullet:
Made it into the final animation.
Riffler’s pistol bullet:
Went through a lot of changes but made it into the final animation.
Holy Crossbow:
Also went through many changes but made it into the final animation but was changed heavily for character rigging.
Updated revolver:
Unfortunately, this variant never made it into the animation either as it was too difficult to get the chamber to fire through prop rigging.
Antagonists’ pistols:
These were scrapped entirely but the twin pistol idea made it into the animation for the protagonist to use instead.
Simple gravestone:
Made it into the final animation.
Military gravestone:
Made it into the final animation.
Holy gravestone:
made it into the final animation.
Bottle of Gin:
Made it into the final animation.
Dusty rum bottle:
Made it into the final animation.
Screenshot taking by Jonathan Graham depicting why cell shading was so important, the props to be completed by me were mostly re-done to accommodate this parameter and were done so in time for the next stage of the pipeline.
Semester 2:
Almost immediately after semester 1 ended, a lot was changed about the pipeline, including taking previously learned modelling techniques and applying them to the character rig process. As I took the role of modelling, texturing, rigging and exporting both the antagonist and the protagonist, I was a lot more prepared and motivated to create a lot of assets based on the old ones that did make it into the final animation. However, to see why this happened, the timetable will be provided below which we used for semester 2 and why a shift in work ethic was necessary to meet our deadlines.
Semester 2 timetable:
| Week 14: | Finalizing characters, fixing rigs and redoing and covering both characters’ mouths to stop smile line animation issues, Finalizing environment. |
| Week 15: | Researching sounds and revising the script, animating, and finalizing intro and opening shots. |
| Week 16: | Animating, Lighting. |
| Week 17: | Importing any remaining 2d Assets, finalising the intro and using After Effects. |
| Week 18: | Animating and Cel-shading effects. |
| Week | Animating and perfecting Cel-shading effects. |
| Week | Animating. |
| Week | Animating. |
| Week | Animating. |
| Week | Finalizing post-production. |
| Week 24: | Post Production and Submission. |
Making the Intro Sequence:
Although I’m mostly a 3D generalist I did want to add to my 2D experience. The entire sequence was made in Krita as a raster animation. I chose this method to animate as it gives a lot of freedom in separate animations working in a layered format. It would have taken far too long to animate every frame by hand or add effects when a transparent layer can be added to the frame transitions themselves.
Creating the first shots of the animation. Sticking to a more traditional layer setup, I relied on layers such as the left and right wheel being separate from the rest of the animation, the same with the dust and track movements. I could easily loop the smaller details while adding enough transition frames for our names to appear after an object flies past the screen. This makes it look like the train is moving at a good pace. In terms of pacing, the requirement is 24 frames per second, making this intro sequence well over 500 frames long, lasting a total of nineteen seconds. We didn’t want to have an intro that lasted over twenty-five seconds, so the fast nature of speeding train wheels made for a good intro animation and was easy to update. These updates will be displayed below:
First draft of the intro sequence. Mostly just background and foreground tests.
The second draft, this one focussed a lot more on the detail of the scene, this was later updated further to accommodate transition frames.
While not making it into the intro as its own animation, this still-shot animation was used but not animated in the final intro draft. Though this concept was used as a reference for the modelling and colouring of the antagonist character.
Although not being used in animation (as its own stand-alone) This image was re-created and used for the poster for our team’s end-of-year showcase and so was a worthwhile use of time instead of a waste which worked wonders for our pipeline as the end of year show material is usually created at the end of the semester instead of in the middle.
The final draft of the intro animation. Flows well with the style we were going for and stays within our team’s set time limit.
Remaking previous assets and character modelling:
Many of these assets made in semester one had to be quickly remade as I had now faced the task of modelling and rigging both 3D characters. The assets created previously did not fit into flood rigs and had too many skin weight problems. These processes will also be mentioned when modelling the characters themselves as I have never character-modelled before but learned a lot in a short amount of time. This injection of quick learning allowed me to tailor the assets properly for the characters’ meshs’. Despite my initial failure to adhere to character rig references when modelling the first set of assets, I learned a crucial process of an animation pipeline which is of course to make sure everything works and to do it in the right order. Realising my mistake I quickly corrected my error in a short amount of time to not impede my team’s progress. You can view the updated models below:
Re-rigged Holy crossbow:
The Holy crossbow needed to be made thinner and with fewer details. The first variant had detailed textures which was not necessary. This led me to be creative in my method of balancing mesh detail with texture detail and this method stayed with me right until the end of this project.
Twin M-C96 Frontier pistols:
The four revolvers were scrapped in favour of a pair of twin Mauser pistols. This pistol was rare during the Wild West Frontier periods but the reason this version was chosen was that both twins have strip magazines which are easier to work with as opposed to a revolving cylinder on a revolver, which needs its own keyframes and mesh controls whereas these Mausers just need a character mesh control.
Character Creation:
Starting with character modelling, the first step is to create the hands. I thought that the body would need to be created first but because the hands are an ‘end’ to the mesh, they need to be created first as their loopholes, vertices and level of geometry play a vital role in the rest of the character’s creation. A detailed hand makes for a high poly body and a low one equals a low poly body and so on. In order to make the fingers move, loopholes are added to mimic the look of knuckles. Then it’s mostly just cutting and sewing copies of the original finger to the rest of the hand.
The end result looks something like the example shown above.
Up next is the head. The head must follow a systematic pattern where the most geometry is in the middle. The top and bottom of the head both end in an edge loop totalling sixteen vertical ends, this is because the character is not overly detailed and the shoes and hands need to have a low poly-ish look to match the cell shading requirements due to only lambert textures being used. For the centre line which will connect the lower mouth, nose bridge and cheeks together, two parallel lines are put on either side of the centre line and then merged with the vertices one loophole before the edge of the bottom and top of the head, this was done to keep the sixteen vertex limit intact, otherwise, the neck will not work with a rig.
After the hands, the body was the next thing to create. The early version shown above has a mesh where, instead of legs and feet, the mesh is trousers and shoes. This is done to save time on modelling and there is no point where the shoes would fly off in the animation. The same has been done for the torso where the only biologically accurate parts of the body are the face and hands.
Like making the hands, the body started with a cube, then was sculpted into a torso with six loopholes up and down. The legs and arms are then extruded and a total of sixteen vertices will encase the neck all the way around. This is done to connect the neck to the head, the reason it had to be sixteen is because a centre line is not added until the characters mesh is completed. The wrists and ankles only had eight vertices all the way around them to achieve symmetry and were vital for rigging later on.
Once the main mesh was correctly sewn together, a reference was made to use for the second character. The first rig work also started at this time using Maya’s Advanced Skeleton add-on, which gives much more creative and logical freedoms than Maya’s base rig tools. The first step of rigging was simply to create a skeleton that matched the mesh.
The next step is to bind all the equipment to the protagonist and for the whole mesh to be prepared for skin weighting.
When I first started to skin weight, I was not experienced and was a little intimidated by the process but with the help of the animation tutors, Alec Parkin in particular, I realized that skin weighting is just applying an area of effect for the corresponding parts of the rig. Like in the example above, the chest needed to be edited a lot so that the shoulder and clavicle joints moved the arms in a fluent and natural way instead of a stiff and robotic way. In my perspective, skin weighting is just painting an area of movement onto a mesh and this way of understanding turned into confidence and greatly helped me in the antagonist’s skin weight painting process.
The final result is shown above where all static equipment such as the gun holsters, bandoliers, belts and bags will move in sequence with the character’s body, while the active equipment such as the twin pistols and crossbow possess their own control rigs that will snap to the characters mesh when drawn, holstered and fired.
The reference rig created earlier was used to create the Antagonist. Aside from different clothing and a smaller, thinner body frame than that of the protagonist, the antagonist used roughly the same mesh creation rules and methods used to create the first reference.
For the second time, the equipment and accessories had to be altered and bonded to the antagonist mesh for a smooth rigging process.
Once the equipment was merged with the mesh for a second time, another reference was created as just the base body to have an accurate sense of depth for rigging the skeleton to see where the curvature of the back and legs needed to go and how dense the rigs skin weights need to be. This method was used for the protagonist but only after the antagonist was fully rigged. I wish I had used this method the first time around but it’s all a learning curve and I’m happy I was able to learn this method properly and quickly implemented this way of rigging for both characters.
The final character mesh of the antagonist with all equipment bound to the base mesh.
Rigging a second time went much smoother than the first time as I had learned many new methods of biding a skeleton to a mesh and in much more time-saving ways such as ‘boxing’ out the character’s rig before skin weighting, only moving joints if absolutely necessary and mostly rotating appendages into place, keeping a centre line consistent even if it meant more work and applying the clavicle and shoulder joints independently from the torso to avoid the armpits, upper torso and lower neck from stretching too much from the upper torso.
Instructions on what order to use the head controllers for test animations.
Skin weighting a second time was much easier thanks to my renewed confidence in tackling its unique issues. It took only a quarter of the time, around a week, to fully rig the antagonist whereas the protagonist took me close to a month.
Once all skin weights were correctly implemented. I quickly went back and applied a lot of my newly learned methods to the protagonist. While this was not planned in the pipeline, it was a quality-over-quantity situation and quality won that predicament as making these characters as easy to animate as possible was crucial for my team to get the animations out on time.
End of Production Work:
Voice Lines:
At first, the voice lines were created using A.I. voice software but sounded far too robotic. The accent was good but the quality of the composite recordings sounded raspy and forced. So the voice recordings were re-done with my natural-sounding voice. Albeit sounding a lot less Frontier American, the quality was far greater and so we settled on using my natural voice instead of a voice A.I.
Original animation voices could not be uploaded to this blog because of file formats however they will be present in the OneDrive link,
Conclusion:
From start to finish, this project was a lot of fun, especially with a well-coordinated team. I learned a lot about various new industry-standard techniques within a professional pipeline. This project pulled me far out of my comfort zone in terms of character creation and I’m glad I took that leap, without it I would have only had intermediate knowledge in prop creation, environment design and 2D animation. But now I can add character modelling, rigging and 3D animation (proper) to my generalist skill set.
The team I was assigned to was the joint first best team I have had the pleasure of working with and while I am comfortably able to work independently and with a healthy amount of discipline, working in a team greatly helps me keep on track of my time keeping disciplines and work ethics. throughout the duration of my degree, I have managed to maintain my optimistic and high-octane but approachable personality which makes it easy for me to merge with a team of any size.
While there were many challenges we faced together as a team, we never fell apart as a unit. Making the decision to utilise everyone’s strengths at the very beginning of this project was probably the best decision we ever agreed upon. It allowed us to tackle various tasks at the same time and to keep in close proximity to one another. However, there were times when bad communications, missing deadlines every so often and needing to re-edit scenes, assets and animations would throw us off course a good bit. These issues proved menacing but having faith in our own abilities and each other and a commitment to the pipeline we had created made these errors and mistakes more like inconveniences than project enders.
At the end of this project, I emerged a lot more confident and more experienced in industry standard tools than when I started, expanding my generalist 3D and 2D skill set and maintaining a strong desire to work in a team. I couldn’t have asked for a better experience than the one I received. The last lesson I took away from this project is the same one I learnt when I started this degree, you cannot learn if you do not make mistakes, you cannot overcome obstacles without help from others or learn from others and industry learnt disciplines are far more valuable than aspiring to be the best contributor and that a good animator is not much good if they cannot first gel well with a team, whether their interests align or not.
References:
(In order of appearance) :
https://gunsmagazine.com/guns/old-west-replicas/
https://billymachin.artstation.com/projects/loNv5
https://billymachin.artstation.com/projects/loNv5
https://docs.krita.org/en/user_manual/animation.html
https://blenderartists.org/t/low-poly-cowboy/443047
https://help.autodesk.com/view/MAYAUL/2022/ENU/?guid=GUID-996AE123-BFF4-42B7-B5C8-E3F2DDFC3D12
https://help.autodesk.com/view/MAYAUL/2022/ENU/?guid=GUID-A2AAF7B8-4187-4E1B-8BC5-98F84474C0EF
https://www.animationstudios.com.au/advanced-skeleton
https://help.autodesk.com/view/MAYACRE/ENU/?guid=GUID-7D773C38-F9CF-4141-8ADD-4E0454838BB7
https://en.wikipedia.org/wiki/Mauser_C96
https://www.pinterest.co.uk/pin/449656344024756749/
https://www.pinterest.co.uk/pin/72409506504499393/