Evaluation of the film room

What scenes I considered and why did our group choose our scene

When we first started the project we all had our independent ideas, I believe 12+ between us. We had to cut the amount of ideas down quite a bit. The 3 ideas I’m going to discuss is: the tent from harry potter; the apartment from fifth element and the office from mad men (something that someone else suggested) and why in the end we all agreed upon the mad men office.

First the Weasleys tent from harry potter and the goblet of fire. I personally found the tent beautiful, the amount of detail, the patterns in the fabric and the lovely colours made me want to choose it; however we decided against this as the amount of tris for the tent itself would require for all the hanging fabric was going to be very cumbersome, not to mention none of us have any experience of modelling fabric so it was going to be far too difficult in the time we had.

Next there was the apartment in the 5th element. This would have been perfect in my opinion, its small, lots of bric-a-brac and interesting shapes where would could weave quite the interesting narrative. Sadly there was so little reference to use and any we did find was incredibly low quality so we decided against it.

Lastly there was the office from mad men. it wasn’t my idea however it had quite a bit of reference and there was enough props to keep us busy and overall it was a very manageable task. However personally I wish the scene had a strong directed light, as our still looks a bit dull, but that’s just my opinion.

Why that still in particular?

We chose the still from mad men in particular because it showed the room without any people which made our job much easier. The still also shows off a good amount of assets.

Personally I would have preferred with a bit more dramatic lighting as I feel the lighting and composition of our still leave a lot to be desired; another small issue I have with it is the colours used, it’s no one’s fault but I feel there has been quite a bit of editing done to the image so we are going to really have to work quite a bit to colour correct our martials and fix our resolution.

out of the choices we had this one seemed the most logical one to do (as it had the most reference, and the objects seemed like something we could do within the allocated amount of time) as it is good to challenge yourself, but we need to also be realistic in our assumptions of what we can do and how fast we can do it.

Gathering reference

We gather all and any reference we can get our hands on. That included watching the first season of the TV show mad men, and collecting screen caps on our objects to get as many views as possible; till we can imagine it as a tangible object.

Admittedly this is difficult for a number of reasons. For example, no matter how much we will the producers to do so, they do not tend to have their furniture as a main focus of a shot, nor do they film the less seen places (like under a table or behind a sofa). This is more of an annoyance then a problem as we can just imagine what it looks like, as the matinee will most likely not see it anyway.

There are a few objects where finding reference in the show was especially difficult, this is where the shows seemingly large décor fan base came in handy. The show is subject to countless websites analysis and discussion about the room’s modernist designs, that meant clean shots of furniture for us (which was particularly handy).

For items who’s reference was unable to be found in these ways, we searched for the closest alternative to the object (for instance with the ash tray I used a stereotypical ash tray instead as I could not find any good reference, then I just edited what I made to fit in with the office.

Prop creation

I had relatively few items to make (they were an ash tray, a chair, a table lamp, a standing lamp and a table) and I had no real issues with them. They were a wide ray of different textures and types of items so I wasn’t making the same thing over and over.

I had to fix many problems and inconsistencies in my models at first; for example my team gave me the following paint over for my chair.

When I got feedback like this I immediately went out to fix it.

This is just an example of some of the changes I had to make, most of the feedback I got was by team mates sitting next to me pointing things out to change and fix in the modelling stage if things look out of place.

Texture creation and unwraps.

So, we need photo realistic textures. To make textures I travelled around my local area taking photos of anything and everything to create textures; for instance the wood texture I use is from a bottom side of a local shops balcony, this woven in with parts of other images and edited heavily to make the textures to a high enough standard. I then ran the textures I made through software like material2bitmap and crazy bump used to get a bases for roughness, normal and other maps. But due to errors, limits in the program and sometimes my own pickiness, these are only good as a base in which you later open up into Photoshop to work into it. I will say that I am a tad disappointed with some of my unwraps I could have been far more economic with my unwrapping as there is sadly quite a bit of wasted space. However its done and I believe the PBR looks very good on the objects even if they could be tweaked a bit, maybe next time I would dedicate more time to just unwrapping so I get very little distortion as possible. Also, I was tampering with my floor textures until the last minute, it started being not green enough, but now I personally feel it is too green, however there is no time to change it now. (The following is a before and after shot of the floor)

Unreal 4 implementation

Adam is our main engine guy, he holds the master file and we give him all our assets to put into engine. We may ask to tweak things and what not but otherwise he is in control. He was the main guy whom did the light box and so our objects are pretty much just put in in the place holder’s position. At this point I am thankful we did a good light box as it saved us so much time moving and arranging items.

I think it was a good idea to have 1 main unreal level instead of trying to merge multiple unreal files together as everything is ordered, we know what’s in and what’s not and it makes organisation so much simpler. I do understand it may be difficult for just one person to be in control of the unreal engine parts, so we made sure to give him more manageable assets to counterbalance the amount of engine work he is doing. We all also have input while he’s working in engine, and I know if we request to edit something or to use engine ourselves he would let us, as one of the main reasons for doing this project to learn new things.

I also created my own small test room in unreal so I can still learn how to do things in engine on a smaller scale, and I can test how my assets look in different lighting conditions. (early and late screenshots of the testing objects are below) they helped me see any more issues, for instance in the early image the lamps light maps look awful and I couldn’t pick up on that in max.

What I could have fixed things and my plan for next time

There is a lot of things I could have improved. My main point of contention was my unwraps, they are all slightly distorted (especially my chair) and there is a bit too much free space. Most of my work needed constant tweaks; this is where my group have been instrumental. (notes posted on the group written by Ellie and Adam)

We all have been keeping notes on feedback from each other, our peers and our tutors and sharing them. Mostly it was small things, for example, Adam and Harvi did a paint over of my chair highlighting mistakes in the back and pillows where the back needed to be more curved and the pillows less puffed. At the time of writing, I realise it would be a good idea to keep a list of all the fixes and tweaks I was told to do to put into this essay. At the very least I know what to do for the next project.

3d is not my strongpoint yet. but with up and coming personal projects I will fix this, the only way to improve one’s topology and general 3d skills is the same with 2d painting skills; practice, practice and more practice. I will also admit that this project I could have done more work in engine. I did set up my own little level for myself and my assets so I can learn how engine works however next time I would like to get a copy of the room and set up lighting and particle effects myself just for practice.

Lastly, I feel next project I will pay more attention to my folder organisation and naming conventions, as they seem to be slightly scattered in comparison to my team mates

Final Room

Everything was going well till the day before, every time we tried to fix an error another error popped up, then another and another. We did our best to fix things however some of our textures (including a metal ash tray and an ashtray on a stand) tuned black in engine, but the same file open on another computer showed it fine. However we did our best to fix the issue, but it’s far from perfect.

All in all, I’m proud of our group project, it strongly resembles our original image. I couldn’t make the particle effects and slight cigarette smoke smog in the end which I regret, maybe next project I can give it a go? And on a positive note, my trouble shooting and lighting ability’s in unreal have gotten better. I will also say that I need to consider my texture density, as I have a horrible habit of using excessive large textures for small objects. I would also focus more on the floor, as I feel I failed my team as the floor looks too flat. Lastly I would ask tutors to specify the brief for me, as I don’t know if the tri count includes the player character/camera/sky dome. I should have asked this far in advance; however that is my mistake and it’s not too disastrous. In the end, I know what I need to improve but I feel I did well. My team has been fantastic and prompt with replies, meetings and feedback and I couldn’t ask for a better group.

(The final render of our room)

Team Manifesto

As a team, we picked out the Mad Men office as our final chosen scene. We chose this scene over our different options due to a variety of reasons, such as the fact that there were enough assets to evenly distribute to everyone, and the overall colour scheme was warm and inviting. Some of our other scene choice finalists had cool and muted colour schemes, and perhaps not enough assets to keep us busy for the entire project period, so the Mad Men office had a clear cut advantage over these. The assets also have plenty of variety to them, with a good diversity in materials and enough large assets to give everyone main points of interests as well as smaller, arguably less important, objects. The room is a good size to have a camera flown around it, and there was enough reference available to us to create the rest of the room that wasn’t in the scene without having to spend lots of time thinking about how to lay it out. Finally, we liked the lighting, with lots of differently lit areas through the use of lamps and windows.

At the start of the project, we created an asset list spreadsheet and distributed the assets to everyone so that we all knew what we were doing, as well as being able to see what other people were doing, and their progress on it. As a team, each week we set soft goals as to where people should ideally be up to by the end of the week, for example having base models finished, or everything unwrapped. We made sure to keep up communication with each other, keeping people informed if we had a problem so we could share our knowledge to fix it, and showing each other our work so we could give constructive criticism. We made sure to meet and work as a group relatively often, as this allows for better communication. Individually, for the majority of the time, we all managed to work to meet the goals we set each week, which kept the project running smoothly and well on schedule. Engine-wise, we each individually worked on our asset’s materials, and then these were all imported into a single main engine file. This reduced the creation of a choke point, as it meant that people weren’t spending time figuring out how to correctly set up their materials in the scene file whilst other people were waiting to put their assets in. Overall the team worked well and cooperatively together, with good planning and organisation to make the project run smoothly.