Matthew Roberts

This comedy song succinctly sums up the ‘morally ambiguous’ section of my (still not completed) film

It’s also very witty

The Sputnik Scene

The cold war game scene in the video shows a variety of significant events that happened in the cold war, the arms race and space race, and ties them all together into a fluid 18 second sequence (based on a video game).

One of these scenes will (for 1 and a half seconds) - show Sputnik in orbit around the earth.

This is a rendering of the earth, a test. In the scene, there will be a quick cut to a simplified animation of Sputnik in orbit. 

n.b. It takes ages top create a scene like this, and in the film it only appears for a moment - just long enough for the eye to acknowledge it, then it moves onto Sputnik in Orbit.

Making It All Work 2

This video tracks the American character (Player) as he falls behind in the cold war race scenes. 

In the final production, I will cut between the Russian and American cold war players. The purpose of these films is to build the environment, and get the motion in and functioning.

All the buildings (which are at the same z-plane as the characters), are linked to a null object which scrolls through the scene. This means that the buildings only exist on the timeline for the time they are visibly seen the the virtual camera.

This creates the illusion of a huge environment, which are players can run through. The characters mostly remain in place, unless they jump up a level or onto a platform. A virtual camera tracks their movements up and down levels, whilst the environment is scrubbed to create movement.

At this stage the characters are framed wide, but will need to be framed tighter in the final shots. This is easier to do once the action is all in place, and can be done by creating mutiple comps with differemnt virtual camera zooms and framing, then cutting between the action for a good cinematic effect.

The other advantage with this approach, is that once all the action is worked out, I will know exactly what needs to be designed, how long it will feature on screen, and will be able to judge how it looks based on the following criteria:

1. Is the action too fast?

2. Will it look ok at a smaller screen size (for example, on a web embed).

3. Can I get all the elements into the background that help tell the story

4. if not, work out a way to do this (by adding more time, creating space to get significant visual elements in) 

5. What details can be cut - whilst pushing g the narrative elements to the fore?

and so on.

MY PHILOSOPHY ON WHAT I HAVE DONE ON THROUGHOUT PART 2 OF MY M.A.

It is essentially a process of doing, and trail and error - and I feel by doing this, I am intuitively learning how to do it. This is, ultimately, the only way to learn, Mistakes need to be made. 

There are tutorials for many things on the Internet, and they are useful, but they only show how to copy a technique. There are no tutorials for projects like this. I also believe that there is far more to learn by doing. I always find you have to get things wrong many, many times before you can get them right.

I also find that to really learn something, you need to do it many times, repeat the process for it to sink in and become useful in future. The learning and doing has not been a rapid process here, but is the most important part of my MA. There isn’t much to show by getting things wrong, getting stuck, trying again, but it is vital for future success.

Making it all work

Creating the middle section of this film (the cold war platform game scene) has been very challenging.!

It is the most high energy, action packed p[art of my film, and the one that posed the most problems. I had previously decided to attempt all the hardest scenes forst. These are all the scenes that involve things I did not yet know how to to do. I have never done any character animation, rigging or any of the things in this scene before. This meant learning the following prior to getting to this stage:

1. Character Design.  - I attempted to design all the characters myself from scratch, but had difficulties achieving a consistent style over multiple characters, many of which needed to have turnarounds (i.e. been drawn front, 45º, side-profile and back). I did all the research extensively, and knew what the characters out to look like, and gathered all the reference material.
2. Character Design collaboration - I have limited drawing skills. This means I can do most things, but I did struggle a little with the designs for characters (as outlined above and here). I spent time over 3 weeks trying to get these done, before contacting an illustrator friend to create simple line drawings, which I could build into full digital characters. (see other post on character design). I managed the illustrator’s workload, and supplied a brief for each character. This meant I had to really give each characters story, so he knew how to approach the design. This meant drawings would come back and forth (over the internet) - please refer to paper file.
3. Once some decent characters were ready, they all needed to be digitised. This meant I redrew every character in Adobe Illustrator, in all poses. This meant that characters would change from pencil drawings into fully fleshed out colour characters.

Here is the rendered version:

These characters would all require rigging. This is also something I have never done, and didn’t know how to do. The rendered character needs to be broken into all the ‘puppet’ parts that would cover every limb to create the movements needed. This is:

1. left foot
2. left shin
3. Left thigh
4. Right foot
5. Right Shin
6. right thigh
7. hips
8. torso

etc. I could so on! 

These all need working joints, so they can move the way they need to. This is very complex if the characters have a stroke around them. It means the fill part needs to be on extra layers in front of the stroke, or it looks wring when the joints move. This process needs to be repeated for every character who needs rigging, and in every position - e.g. front, 45º, profile, rear.

rigging example:

I felt that it was important to know all this before I designed the backgrounds, because the characters would need to be animated later, and I wanted to be able to work out what would need doing, to prevent a stop/start workflow later on. I did some Inverse Kinematic tests with a skeleton to practice my first walk cycles.

Once I was certain I could do walk cycles, I then rigged the cold war race characters (the Russian Red Guy and the American blue guy). I realised at this stage that I could animate them, but I needed to know where they would be, when they would move, where they were going, when they would jump, stop, start. This lead to the building of the sets.

THE COLD WAR SCENE

I have never created a scene like this before, and for some time did not know exactly how this would work. It needs to feel like a platform video game, in which the players travel through a scene.

I could not find anything on this particular issue. I spent a long time trying to find out how this would work, and after an extensive search I asked several pro designers who had done things that involve a lot of space, where a camera moves through a scene. Just Katz, of a Flock of Pixels simply advised me to test things out, and to “do whatever works”, whilst another friend, Andy Olley said it would be best to avoid a massive comp in After Effects, as it wouldn’t render well, and the computer would freeze up.

I realised at this point that I just needed to try things, so I drew the whole scene out on paper (about 30 feet worth)

From here, I tried the big comp in After Effects. This was basic shapes, without features in. I discovered from this process that the idea I drew out didn’t work as I hoped. These were the problems:

1. The amount of buildings and details I envisioned were too many. The eye could not see the details.
2. Many of these things were essential to the narrative in this part of the film - It was important to be able to see what was going on.
3. The general motion of the scene was too fast. It looked wrong!
4. The process of trying to work out how to shrink the horizontal area to get this speed right was not intuitive at all. If I were a brilliant mathematician, I may  been able to work this out somehow, but I knew by now this would still be a bad solution to this problem.

There are problems with building a huge scene digitally, and scrolling through. There are limitations on widths, and the computer does not like having to deal with a huge ‘comp’.

I knew that this had to get fixed (which resulted in these tests).

The set up is driven by a big comp, which is set to the proper narration. The elements (blocks which represent buildings) in the scene are parented to a null object (this is an invisible later), which is scrolling. This means the layers are only as long as they appear to be , but are all arranged in an order. Once they are off the screen, they no longer need to remain in the timeline, and the comp is just a simple 1280 x 720 pixel screen shape. A virtual camera is set up, and that tracks the movements of the players as they move through the scene (however the players are not moving forwards, just up and down, jumping ‘on the spot’ - the background is scrolling). The camera also zooms on certain actions.

There is currently two version of this scene, as I will edit between the two takes to show the race scene. It would also be relatively easy to add more shots once everything is in place. This will help the narrative and home in on important details

It is a lengthy process (hence why it is the most complex part of the film).

.

(NB - the action is currently framed mostly with a wide shot. This looks ok when viewed at full size, but the characters look small when it’s embedded at a smaller size, like it is here. To get around this, please click the You tube link in the video (bottom right), change viewing size to 720p, and click to view in full-screen mode)

I also found difficulties working out how to take my ideas, storyboards and scripts and turn them into real scenes. There are a great many tutorials avilable via the Internet which will show you many techniques, but these tend to be isolated to certain effects, looks, or methods to do a single idea. My film involves a multitude of ideas, all working at once. These are:

1. Character Animation & Rigging. (this is after all the characters were designed)

2. Audition, selection and recording of voiceover narration (for the entire film).

3. Building of a large scene/world for the characters to race through.

4. How this might physically work - how to build the environment.

5. testing how this works, blocking out action.

Further posts will break down what was done, in order, with details of what I tried, what works, what didn’t work, in the order that I tried them.

NB - The characters here are static, but have already been rigged for animation.

FYI  -This means breaking the character into puppet parts, so that there are enough working joints to animate them. (Feet, calves, thighs, hips, torso, upper arms, lower arms, hands, neck, head) - plus separate features like eyes. There are also a load of replacement elements, such as different hand shapes, blinks, different mouths.

Some of these characters also need ‘turn arounds’ - so the above puppet process needs to be repeated for side views, 45º turns, rear views 

This requires a lot of work - so it is wisest to work out exactly what you will need doing so that time isn’t wasted rigging characters which don’t feature much (etc)

Here is an example of a rigged character. This is a skeleton rigged with inverse kinematics.