Anderson Shelter Scene Submission

Welcome to the submission blog post for my Anderson Shelter 3D environment.

This scene was created over the last 3 weeks as a proof-of-concept for my proposed focus area for the year. I am going to be looking at atmosphere in games and how all the different elements of a game environment come together to make an atmospheric scene, including such things as lighting, textures, audio, particle effects, detail/set-dressing, historical/contextual accuracy, post-processing and rendering and so on. To this end, I set out to research the creation of atmosphere and apply this knowledge to creating an atmospheric scene.

Here's the final presentation shots and video of the scene. Below these I have done a write-up about the project.

 

Submission Video (Note: blogger compression has made the video look awful! (Please enable audio)):

...and now for the write-up! This write-up describes the steps that I took to achieve my goal for this project, the design process and the motivation behind my design decisions.

Idea Generation:

First, I had to pin down what my scene would be. I set the following criteria:

• Realistically, it had to be something small so that I could do it in 3 weeks. It couldn't be a huge warehouse for example because it would take too long and I wouldn't be able to put enough detail in. It also needed to be an indoor scene. A WW2 Anderson shelter seemed like a good idea because its small and indoors.
• It had to be a setting with a lot of atmosphere potential. There needed to be a lot that I could do with the audio, lighting etc. An office supply cupboard would be terribly unatmospheric, for example, whereas a bomb shelter gave me a lot of options.
• It had to be something that I could do a lot of research on. There needed to be plenty of information online and in books and places that I could visit to learn about my setting. Its safe to assume that ice age cave dwellings would be difficult to research. An Anderson shelter was a good option because plenty of information, images and first-hand accounts are available, and there's still some shelters in existence that you can go and see, which brings me to my next stage...

Research:

Research was a huge part of this project. I began my research immediately after coming up with my idea. I went to the library and withdrew some books about the Battle of Britain and the Blitz and read about Anderson shelters. I created a Pinterest board and began to gather images of shelters and images depicting war-time life in Britain. I did plenty of online research. I found websites containing first-hand accounts of shelters.

In terms of first-hand research/field trips, I visited Western Approaches WW2 museum in Liverpool and I attended a talk from author Michael Foreman at the Harris Museum and asked him about life in the war and shelters. My research was ongoing throughout the project.

I decided that I wanted to go into very fine detail in the historical accuracy of the scene and that I would set the scene on an exact date. I liked the idea of setting the scene on the night of October 28th 1940 - exactly 76 years prior to the projects submission date. So I researched whether this would make sense. I delved into the depths of information about the Battle of Britain and the Blitz and found a brilliant website that had exact day-by-day action reports for the entire duration of this stage of the war. On 28th October 1940, there was a night-time bombing raid on London -

Creating the scene:

A lot of work went into trying to make the scene as atmospheric as possible. I identified a bunch of things that needed to be done/included to achieve this.

Modelling: I did all of my modelling in Maya. I stuck to fairly low poly count limits when modelling - the biggest model in the scene is just shy of 3,000 tris. I didn't want to go too low poly because I thought this would limit the scenes realism and by extension, how atmospheric it was. The texture maps are mostly 2048x2048 with smaller/less important models being 1024x1024.

Texturing: The textures are almost all PBR with diffuse, normal, specular, metallic and roughness maps. Again, this was to maximise realism. The textures are hand-painted in Photoshop but in a realistic style. I used Maya to bake normal maps and then edited them in Photoshop as appropriate. I generated the specular, metallic and roughness by taking the finished diffuse textures and making them grayscale then playing with the brightness and contrast levels until they looked right.

Scene composition/rendering: The scene is built and rendered in UE4. I added a simple dust particle affect. I added post-processing affects such as bloom, slight lens flare, a dirt mask on the camera and subtle colour tints to the scene.

Audio: I created audio for the scene using Logic Pro X. The final audio contained several layers. There is two seperate layers of atmospheric noise/drone of planes overhead. This was created by recording the sound outside of my bedroom window, I then added lots of affects to this such as distortion and reverb to make it sound like the noise of planes in the distance. One of the plane noise tracks is high frequency and the other is low frequency. I found a recording of a radio broadcast by Winston Churchill from October 1940 and added it to the audio track. I edited the EQ so that it sounded like it was playing on a radio. I also added the sound of air raid sirens. I also created the sound of bombs and anti-aircraft guns by heavily editing drum samples.

Creating the audio

Historical Accuracy: My goal from the very start of the project was to make the scene as historically accurate as possible, here's some things that I did and some things that I included to try and achieve this:

• As mentioned previously, the scene is set on 28th October 1940. I chose this date as the project submission date is 28th October 2016. A raid on civilian areas of London did occur on this night. I made sure that everything you see and hear in the scene made sense given the date.
• MKII air raid wardens Brodie Helmet. This was issued to all air raid wardens. It was dark gray with a white W on the front of it.
• Air raid warden's report forms.
• 1936 Zenith 805 cathedral radio. This was a common home radio set that would have been used to listen to war-time broadcasts such as the Churchill speech that I included in the audio.
• Puddles on the floor of the shelter. Anderson shelters were notoriously damp places!
• Propoganda posters. The scene has two posters, the blue one on the back wall is part of the series of posters released in 1939 which also included the more well-known "Keep Calm and Carry On" poster. The other is a recruitment poster for air raid wardens.
• Lighting: the scene contains a lantern and candles - two common methods of lighting a shelter.

Conclusion:

With the scene now finished, I can say that I am fairly happy with how it turned out. I think I made a good decision on the concept of the scene. Conducting the research was a lot of fun because it is a subject that really is fascinating to me and I have learned a lot. I would have liked to include more content in the scene but unfortunately the project deadline arrived!

This project has definitely allowed me to test the validity of my proposed focus area and I already think that I have learned a lot about how atmospheric game environments are created. I'm looking forward to delving deeper into this as the course goes on.