Week 06: Film Room Evaluation

The scene that we decided would be appropriate for our film room scene was none other than the black market room in movie Pacific Rim. Originally my group decided on Davey Jones cabin from Pirates of the Caribbean: Dead Man’s Chest, but decided to scrap the idea after we evaluated the scene and realised it would be too much of a task to replicate the atmosphere the scene created in the movie, not only that but the lighting was very poor hiding most of the assets in the room, this would make it difficult to accurately model what was in the cabin.

The black market room however has a glossy marble floor and lighting that wouldn’t be as hard to replicate, the colours emitted by the jars holding alien specimens in the scene makes the room stand out and if replicated properly will make a very impressive and beautiful scene.

Another scene we had in the line-up was a room from Kevin Flynn’s hideout (Tron Legacy), like the Black Market this scene had much better lighting and for the most part the textures were a lot less complex, this would make the scene easier to replicate but most of my group were uncomfortable with how simple the scene from Tron was. They must have overlooked some of the assets in the scene which I found quite interesting.

Bike Asset

Area 1

Area 2

Area 3

Overview

The fact that our chosen room has a sci-fi theme makes it a lot more interesting and unique compared to other rooms we could have chosen, it has alien specimens in jars, unique shells on worktops, sci-fi equipment and an aesthetically pleasing wood grain finish on an oriental themed room, this sets us a very challenging task with a high reward if replicated accurately.

Personally I think this still was the best we could get from the film. There wasn’t a single screen capture we could get that didn’t have the actor as the main focus as he explored the place, this inevitably left us with a still that was partially covered up. That aside the still has very good composition and a variety of complimentary colours making what could have been a perfect shot. It represents the theme of the film including most of what makes the film sci-fi and it shows at least one asset from each team member.

Creating UVs was familiar as I had to do it a lot last year, this time however I put a lot more emphasis on packing my UVs tight so all the space was being used effectively, this meant allocating the larger more visible parts of my models to larger spaces on the map, and making sure all parts of the model were unwrapped effectively. This included adding and removing seams appropriately, pelt mapping, relaxing and stitching to create a professional standard unwrap with no stretches and such. This was quite a challenge as I just managed to get my head around the basics last year.

Dremel UVW

Dremel Material Assigned

Making the materials in the Unreal Editor and creating different types texture maps in Substance was definitely a highlight in this project. I have never created a photorealistic texture with normal, roughness or specular maps before and I was impressed with the materials I had created. Once I had managed to build an understanding to how each parameter/property changed the way light affected the material surface I was able to create realistic looking materials that weren’t real, e.g. creating glass with metallic properties to add more reflection. It was also surprising to me how you could take other materials and use them to create normal maps for materials that have completely different properties, e.g. using a plastic texture with deep scratches as a normal for a metallic material, it doesn’t sound logical but as long as it looks good nobody complains.

Dirty Plastic

Material Test

We ran into quite a few problems in the Unreal Editor, luckily for me I didn’t have to deal with most of them as I didn’t spend a lot of time with our actual scene until the very end. The scene can only be accessed by one person at a time so I did most of my UE4 work in a separate scene so I could access it when I please. One of the main problems was glass. Our scene depended a lot on glass material but nobody knew how to create it, the default glass the editor supplies was terrible, nothing but a transparent material with no properties of glass whatsoever. Seeing as I was handling the creation of the glass jars I took upon myself to create the glass material that could be re used around the scene.

Glass Comparison

Light and glass test

Reflection Test

Initially my glass material was useless as it appeared invisible but after some research I was able to add a roughness map to my glass so it wouldn’t look too clean. I also created a transparent material with tweakable parameters that would allow me to create realistic light refraction, suitable opacity and colour change which would come in handy, I also created a reflection parameter using the metallic property but unfortunately no reflections could be seen initially. I actually managed to crete reflections for my glass material but only after the hand-in!

Naturally I also created the materials for my own assets, The Dremel tool, Mortar and Pestle, Hand Tools, Lamp and Jars.

HighresScreenshot00001

HighresScreenshot00000

HighresScreenshot00004

HighresScreenshot00003

HighresScreenshot00002

Lamp

Mortar&Pestle

Jar Progress

[INSERT SCREENSHOTS OF ASSETS IN-ENGINE]

I could almost shed a tear looking back at the troubles I went through to create a material that had a close representation to real glass properties just to scrap the idea of having glass jars in the scene all together. It was a bad combination of UE4 and poor lighting in our scene that ultimately made any glass material look almost transparent or plastic, so after advice from a tutor we decided it would be best if most jars were made from a marble material instead (this was then changed again a couple hours before hand-in). We made the initial change from glass less than 24hrs before hand-in, so I had very little time to create a new texture considering I had other things to finish off that I would have preferred to spend my time on. The glass didn’t look too bad on the large Kaiju tank, it did its job of being transparent enough to see the hard work Connan did on the Kaiju organ in Zbrush but was solid enough to make the glass appear translucent and dirty, unfortunately it doesn’t show the refraction properties in the glass without a fine inspection of it but we kept it regardless.

White Marble

Black Marble

Black Marble Texture

Untitled

As for a team analysis, as a whole I do believe we worked well together, there weren’t any major disagreements or arguments throughout the whole project. There were a couple members that were not pulling their weight at different stages of the project but in the end everybody did a respectable amount of work. When it came to prop distribution we all selected an asset to work on initially on a first-come, first-served basis but then decided it would only be fair to select your next asset to model once you had finished the one you were originally on, this system would help promote speed in the work process and would allow those who were quicker at modelling assets to help those who were slower. We used an online spreadsheet that all members of the team could access to store our asset list and information related to it such as who is currently modelling a specific asset, a tri count of the asset, whether the asset has been unwrapped or textured, and more.

My performance on this project was acceptable and I say that purely based on my final work. For the amount of effort and time I put into this project I could have produced better assets in another. I believe my group bit off a bit more than we could chew. I was strongly for the idea of choosing a “simple” room that we could have replicated (Tron film room idea) than a complex room that came with noticeable flaws if not replicated perfectly, unfortunately that is what we ended up with. It would have been nice to model something that was actually big enough to justify a high tri count but all my assets were relatively small so I had to keep them low poly with not a lot of detail.Time management in this project started off adequate, I had two of my assets modelled in the first week which was a good starting pace considering I was fresh back on the course, so it was basically me just warming up. After getting pass a major “blocker” in week two I was performing excellently, I was getting all my work done ahead of schedule which gave me a little buffer thankfully because the majority of week five (an important week) I was absent. Week five was when all the lighting and textures were being applied in UE4 and it seems to be the week when my team started to realise the scene just wasn’t going to work, me being there or not wouldn’t have changed the bad choice we made at the very start of this project.

Team manifesto:

Choosing a beauty shot for this project was hard as the scene always uses close camera angles to the actors, meaning that the scene is often not in focus. Shot 1 was good, however the actor takes up so much of the shot it would look a lot different in scene and it doesn't show much detail. Shots 2 and 7 are almost identical and don’t show any of the room so they were not appropriate. Shot 9 is a really nice shot, however it is largely taken up by the shell meaning it is not appropriate for the shot. We needed a shot that would show a diverse array of assets without been massively covered by the actors. Shot 6 showed many of the assets however the actors take up a large portion of the shot; shot 5, however, shows a large section of the scene and the actor does not take up much space. This is why we chose shot 5 as it displays a large array of assets and lighting information, while still keeping a focus on the scene and not the actors.

Our team decided on a system where we listed our assets and people could pick an asset to start and then once that was completed could choose another. This was a good system to use as it meant quicker people could model more things and help the slower people. We kept a list detailing all our assets, their tri count and stage of completion on Dropbox; this allowed us to update and keep track of where everyone’s assets were in production. One area that could have been improved is having a timetable or setting deadlines as this would have set hard dates for asset completion and would have allowed for more time putting the assets in engine.

Connan started with white boxing as he had experience with engine, then moved onto making assets, completing ‘Alien wall jars w/ stand’, ‘wires’, ‘large incubator’, ‘weight scale’, ‘container’, ‘draw’, ‘kaiju organ 1&2’, ‘pillars’, ‘steps’. Connan also intermittently worked in engine. Liv got us screenshots of the room to start with and quickly moved onto making the ‘square tank’, ‘trolleys’, ‘ceiling’, ‘clip board’, ‘jars’, ‘jar crate’; she also completed the first lighting pass. Chris started making assets right away and made ‘desk’, ‘magnifying glass’, ‘trays’, ‘chair’, ‘kaiju corpse’, ‘chandelier’, ‘bone fragments’, ‘scalpel’, ‘textured basket’ and worked in engine placing assets. Tashan also began on asset creating immediately and made ‘jars 1-5’, ‘lamps’, ‘Dremel’, ‘pestle & mortar’, ‘knives 1&2’. Harry created ‘shoes’, ‘cloth basket’, and later textured items that had missing textures as we quickly found he was very good at texturing assets and so textured the shells and draw. Kathryn started creating assets however did not texture two of them; she created the ‘basket’, ‘large shell’, ‘small shell’, ‘skinmite’.