Film Room Project Final Evaluation
The first step of our film room project is to select a film scene and room that we feel, as a group, would be the most appropriate scene to replicate in the project. I put forward three suggestions.
My first idea was the cottage from the opening scene of Quentin Taratino's Inglorious Bastards. Tarantino is well know for making full use of his sets and moving around with the camera, so right off the bat this scene makes a lot of sense as it would be easy to get a good feel of the entire space, compared to other scenes where the camera limits what we can see, and therefore, recreate accurately. This scene also uses some effectively drmatic lighting, highlighting objects and spaces of relevence.
The Griffindor Common Room from the Harry Potter series was another suggestion I put forward, due to the way the room comes together with its colour theming. The warm golds and reds of the room, with a fireplace as the focus create a very homely feel that would be engaging as a 3D environment. The assests, as well, would be fairly simple to create without too much fiddling or difficulty. The only issue would be with the very pattern-heavy textures. The brick walls would give some interesting oppourtunities to work with PBR, which I have yet to learn.
The bar from Kingsman: The Secret Service would also be a strong option. An action sequence takes place here, allowing the team a broad scope of the room. In terms of how 'do-able' the room is as well, the chairs and tables allow for a few assets to be made and still allow us to fill the room well, leaving time for extra details like glasses or posters. We'd also be able to focus on the interior due to the frosted windows, and not waste time on a skybox. That said, the lighting from the windows creates a dynamic look for the room that would be engaging to explore in a 3D environment.
The scene we decided to go with was the bar fight from Kingsman: The Secret Service, also choosing a different still to use as our reference.
The still we picked depicts Eggsy, the protagonist, shot from a third person angle facing the stepfather and his goons, sitting at the far end of the pub. The shot is framed to make Eggsy seem taller and more intimidating and dominating, as this scene takes place at the end of the film. I feel this was an appropriate and engaging still to use and represent the film, as well as a practical and interesting environment to recreate in 3D. The balance of larger ‘focus’ areas of the pub such as the bar and the entrance, along with the repeated, smaller assets of the furniture meant that it was manageable to recreate, as well as easy to assign assets to individual members of the group. The large mix of surfaces and dynamic lighting also provide ample opportunity to familiarise ourselves with PBR and Normal maps and engine lighting. I’m happy with the still we chose overall- the balance of practicality, ambition and simply being an interesting still is something I’m particularly proud of, as choosing an appropriate still has proved challenging among the plethora of films and scenes we collaboratively suggested.
Creating the props for the scene proved to be simple at first, but became increasingly fiddly as time went on. The first prop I made was the Lower half of the bar- a focal point of the scene, so I made it my first priority. Working with Matuesz to make sure both halves of the bar fit, I made a simple base and worked my way up. Instead of wasting time baking complex insets on the side of the bar, I modelled a simple ‘frame’ and copied it across the body of the bar. The brass rail proved figety, so instead of making individual and highly tri-consuming loops that the rail would be fed through, I settled for modelling the loops into the bar itself, which proved much more efficient as I had overstepped my tri count.
The clock was easily the simplest model, making a curve for the clock face to give it more dimension. I wanted a rounder edge, but I didn’t want to overstep the tri count.
I used the photo still, as well as a few other key shots to make the wall table work. The most distinguishing feature of the model was the legs, so I paid close attention to get the correct shape for them. I used strip modelling for the shape and then bridged a hole in the middle of the legs, which I had to play around with a little due to strange seaming. I worked with Dan, who was working on the bench, to get the scale correct for the table itself.
The coke fountain was a bizarre asset- it’s not prevalent in the scene at all, and is slightly out of place. I had to model the tap partially off memory as well, as there was only two (badly photographed) references online and I couldn’t find a real one to photograph. I worked with what I knew though, using the shape of the classic coca-cola glass to inform what the curvature of the unseen part of the tap should be.
Texturing assets is nothing new to me, but normal maps were a first. As we were doing a real-world environment, we could use photo assets and didn’t need to hand paint much, with a few exceptions. Getting the UV layout of my models was painless for the simpler models- the clock was just a matter of pelting and relaxing at the right angle, and I was able to overlay both sides of both legs of the wall table to make optimal use of the texture space. The clock’s texture was very easy to create and apply due to it being circular, and with no awkward shapes.
The bar was more difficult, however- taking individual sides and unwrapping them was fine, but the final unwrap required me to overlay the body of the bar over itself twice to make room, which still didn’t provide a lot of space on the UV. I used the empty space to easily store the rest of the unwrap, but I wish I’d found a more efficient solution for the large body of the bar. For the texture, we shared a universal wood and brass texture amongst us that was used on both the bar and the wall table, in order to keep colour consistency within the level.
The coke tap proved to be simple enough, as I didn’t have to worry about seam issues due to the nature of the object itself being separated by a black plastic through the middle. The raised ‘coca-cola’ logo on the front of the tap was not textured on however, in favour of using heightmaps to create the effect in engine.
For the wallpapaer a tiling texture needed to be used, with a simple normal map for some noise. Tracking down an appropriate image was tricky but I eventually found one that worked.
Working in a team in the Unreal 4 engine is inherently problematic, but we handled it with as much organisation and as little fiddling as we could. With Matuesz being the most comfortable with the software, we would work with him on whiteboxing and lighting, as well as regular revisits to the scale and shape of the level and how it was coming together. Getting the scale of our assets to match up was our first real hiccup while working in engine- it was an inevitability, but we handled it as well as we could. 3DS Max registers scale differently to UE4, so almost every asset we imported, save a few, were not to the correct scale and had to be scaled by hand. We made sure to do this in groups and working concurrently, so the positioning and overall scale of the room wouldn’t get thrown off, which would be disastrous to go wrong in the middle of the project.
We also made sure to have as few engine files as possible. Hearing of the issues other groups had with keeping a confusing number of engine copies, we kept ours to a minimum by physically exchanging the file or simply keeping a computer open with the engine file already loaded if we were working in the same room. Keeping a few backups with this allowed us to keep track of our progress much more smoothly than we would have. Keeping up with each other as a team went very well, and I’m proud of how well we communicated who needed what and when to meet up during this project.
In terms of my performance this project, I’m happy overall but I’m hesitant to say I’m content with my work. I’m proud of my contributions, especially at the beginning of the project, where I really tried to bring the team together and organised a facebook group and tried to keep us all on the same page. Work at the beginning was smooth, but eventually devolved after a few hiccups that I simply got frustrated at- the bar itself was an important asset for the room as it’s what the player is naturally drawn to as the focal point of the room, and I feel like I made the modelling harder than it needed to be, especially with the panelling and the rail. The coke fountain and lack of reference also hindered my spirits and only served to frustrate me, and ended up with an asset that I’m not really that happy with. Working in engine was also an element of the project that I feel I missed out on a little. While working with normalmaps and materials in engine was a baby-steps approach that resulted in me understanding how it works, in terms of how the actual engine works and functions, I don’t think I pulled my weight, and let more experienced members handle it.
Improve workflow
Reduce roadblock frustration
Practice Engine work more thoroughly
Get over fear of asking for help
The final still of our room manages to reflect the original still quite well, in my opinion.
In terms of scale and making the environment look as lifelike as possible is where I think we succeeded the most. The 'feel' of the room is strikingly real, and is something that I immediately see as a positive. I'm genuinely proud of just how good it looks. The textures on the tables look great, and capture the light very well too. The inclusion of the clutter like placemats and bottles really help bring the room to life, the posters working especially well. I'm kicking myself a little for the bar though- it's not quite as high as it should be, and the texture of the surface looks a little too dirty, we overshot on the grunge. There's also a very noticeable gap where a higher stool and table should be, but were skipped for time. A shame, because it stands out a little. And as much as I like the lighting, we didn't quite capture the sunlight correctly, and the room looks too artificially lit.
Team Manifesto:
We chose to recreate the post-credits bar fight from Kingsman: The Secret Service because it was the most balanced scene we could find in terms of asset distribution, difficulty, lighting and mood. The still is simple, but effective in conveying a mood of subtle tension through camera angles and character body language, but the scene itself is fairly simple. Light coming through the windows in a ‘god ray’ style evokes and evokes a very different mood to what the scene implies, giving the impression of a warm, welcoming British pub. The still we chose also allowed us to take full advantage of the fact that the pub was home to a fight scene, with the swooping camera and multiple angles used giving us a clear indication of space, as well as a view of all the assets we would need to make, meaning very little would need to be guesswork or left out. The still also offered opportunities for us to practice creating many different kinds of surfaces and textures, with wallpapers, woods, plaster, glass and upholstery all available as materials we were able to recreate.
As a team we were required to evenly distribute out assets and make a work plan for the coming weeks, something we immediately established as soon as the project began. Our planning began by organising a facebook group where we would establish meet-up times, progress, updates and general discussion of the project. We knew this would be essential to keep track of everything so this was a very early step. Another early group organisation tactic was the asset list- we went through the stills we had of the scene, individually picking out assets and putting them into an excel document, where we would decide tri counts for each asset. While doing this, we also organised assets into a priority list based on size and importance. For example, the benches were going to be higher-poly models, finished and put into engine first before something like placemats. This allowed us to create the ‘feel’ of the room first and make sure that the scale was correct, as well as ensuring that if we stumbled during development the essential, room defining assets would be done in time. During this, we set up a time frame for what needed to be done and by when, aiming for the essentials to be completed by week four, ensuring a two week long polish, and giving ample time to deal with any inevitable road blocks. During the project, our team would also lend our models to each other in order to keep a consistent scale for assets that sit closely together or work in tandem, such as chairs, benches and tables.
Bibliography of textures: Bear Woods, (2015), Clock Hands 1-7/8" - Black [ONLINE]. Available at: http://www.woodparts.ca/media/qh-style1.jpg [Accessed 30 October 15]
Bear Woods, (2015), Clock Hands 1-7/8" [ONLINE]. Available at:http://www.woodparts.ca/media/ss_size1/qh-style6.jpg [Accessed 30 October 15]
Bear Woods, (2015), Clock Hands [ONLINE]. Available at: http://www.woodparts.ca/media/ss_size1/qh-style6.jpg [Accessed 30 October 15]
Katie Blevins, (2015), clock face [ONLINE]. Available at:http://decoart.com/img/projects/files/880_download_clock_face.jpg [Accessed 30 October 15].
Imagetextures.com, (2015), MetalBare0103 [ONLINE]. Available at:http://www.textures.com/download/metalbare0103/13384 [Accessed 30 October 15]
Imagetextures.com, (2015), WoodFine0026 [ONLINE]. Available at:http://www.textures.com/download/woodfine0026/20606 [Accessed 30 October 15].
Rath & Doodeheefver, (2015), Vintage retro baroque wallpaper [ONLINE]. Available at:http://www.vintagewallpapers.be/wp-content/uploads/2014/03/393.jpg [Accessed 30 October 15].
