In this section I will discuss the planning and process of this visual effects project. I will make a critical analysis of the overall experience i.e. what has gone right and wrong with the project and areas that could have been improved.
I began this project by using an old model of mine; I would take the process of working with 3D VFX meaning; I would work on the; texturing; particles; animation and match moving.
I began to make a plan of what the robot was going to look like and the settings it would appear in I.e. the identity. I began to look on the internet for things that would encourage the project, my first intention was to turn to robot simulations. A robot simulation where the robot comes into the scene and fights a group or an individual.
Robocop was one; where ED-209 is switched on for the first time and Warhammer 2000 was another where you would often see a mechabot taking on infantry.
I wanted my robot to have a modern / futuristic look to it, I wanted it to look new and shiny almost advanced but something that didn’t look like it came straight out of the warehouse. The idea is to create something believable that fits into the scene.
“We’re looking for seams, we’re looking for cracks but when they make you forget all that and pull you into the magic that is the story then they have succeeded”
– ILM Samuel L Jackson
My plan is to create a glossy metal robot that is in the stages of testing. The robot will look slightly unfinished as if it is in the process of being finalised. The movements will be limited to rotations of the top half; any movements that don’t rely on the movement of the feet. (This was done because animation needs the help of ‘rigging’ a process that in unknown to me)
The robot will have a face off with an individual or group of people where there will be one victor. The scene will show gun fire, animation, mussel flashes and fast cuts to give the feeling of an action sequence.
The texturing for this project was the most lengthy process, it involved bringing a realistic 3D model together by the choice of textures and the work done with them. When I began the project I had to learn that choosing your textures is a vital role as some images will not be suitable for a design.
The use of a program called Gimp aloud me the freedom to use more textures by making normal images seamless. A texture artists relies on these images to be wrapped around a 3D object and won’t appear to have a break in the pattern. Most textures are an images that has been repeated enough times so as it will cover an entire model.
This is a mapping node that allows a better manipulation to texture placement and size on a model. I found this process needed a level of care and patience as the effects are sometimes subtle. The scaling of textures meant that difficult areas of a model would be made easier to texture or at least give a techniques where mistakes can be hidden. For example, I was having a great deal of difficulty in making a texture fit around the robot’s guns. Because I was relying on the texture mapping node it only really throws a texture over an object as best it can and can have a great deal of error. Where the problem arose was the shape of the 3D model; it had a number of side that are uneven and unsymmetrical, which can limit these nodes.
How I could have improved this process would be to create seams in a model where all the sides are flattened out so as the texture artist can have more control over the overall look; it means you can better position textures onto awkward areas and would have improved the look of the guns to the model; it also gives it a distinctive look as you can work on one side with out effecting the rest. This is known as a ‘UV unwraping’:
This here is where you may see the benefits of UV unwrapping. As you can see where my textures have stretched and distorted:
All problems with box mapping textures, as you can see I have coordinated the texture to hide the errors as much as possible. From a distance this may not be seen but an extreme close up will always put a texture artists work on show (Realism ILM quote).
Nodal Editing was a process that made my texture work come to life as a texture on it’s own can be seen as a boring image to our eyes and certainly would not uphold the design and story behind my robot. The nodal system has aloud me to bring multiple textures together which has brought the design to life.
Usually you would start with a ‘Texture Coordinate Node” this node will indetify what you are working on. For example, nodal systems are a way of translating information; telling it where to go and what it should interact with. You may need a nodal system to work with a ‘camera’; ‘object’; ‘reflection’ etc. they are also used in exporting and all undergo the same process.
I found I had to spend a lot of time realising the effects certain nodes can have on one another. For instance, when I began to work on the body of the robot I used two coordinate and mapping nodes that would be brought together using a mix node. The mix node began splicing two images together, I began to understand the effects by changing how the mix node would work. You can select it to mix in a number of different ways: By having it ‘mix’; ‘multiple’; ‘overlay’; ‘subtract’ etc. It isn’t complicated because the program does it for you although you need to make the decision on which way it combines your textures.
I reached one of the first problems here as it was clear that some textures would not go with one an other and when they did they didn’t always work in the way I wanted them to.
The next part was to use a series of colour ramps and contrast nodes to bring the opacity of one texture over the other. For the most part it was choice in textures that was vital.
This was extremely important as when working with VFX it is vital you bring the best tools before starting a design and good textures are both hard to find and important. The first thing I found was not every texture is right for VFX because of the quality of the image and other factors. I began to look for textures that had a high resolution value so as they would hold up on my 3D model, also images with reflections or shadows could not be used as the light information is given to a texture by the program and any alterations will effect the scenes realism.
I realise now at the end of the project the importance of these textures and what it would mean to go out into a visual FX project with your own textures by use of photography, and to come up with a more personal design. It would provide you with a more accessible project as you could plan and shoot exactly what textures you needed for what. For example I was having problems finding a bumped metal image, it was hard to locate, I knew exactly what I needed but had to rely on sources from the internet. I found a texture that was good but if I had given more time to go out and photograph my own texture I feel I would have had a better pallet for my design.
Not only was locating and choosing textures an important role in but choosing a pallet of textures that would work well with one an other. As you can see from the image above your eyes are drawn to areas that do not suit, so in this case the untextured areas and marble looking gun and areas of grey and pink. I found that choosing just one texture to fill the model made it look fake. This is when I began to make variations on the texture I had already built, this was done to imply some character into the design i.e. I made some areas appear like they have been changed due to construction, each area had seen more treatment than the other.
I did this effect by increasing the ‘bump’;- so it made areas appear warn and also change the colour temperature to make it appear like the robot was unfinished which was part of my initial plan.
Lighting a model and Scene
(I was pleased with the overall choice and design in textures glossy metal base and glass provides an imagination of what the unfinished guns (to the left) will look more like. and rubber recoil pads look great)
Once the textures were chosen they would be finalised against a light source. I used a HDRI of a car park to bring in the light to my design. I found that this would be the biggest part of error through the project as I managed to lose this crucial lighting information in scenes which in essence meant I lost my textures. I would have to go along with a different look than this image in the final video as some shots had already been exported.
I decided to film the project at the back of my house, it was both accessible and provided enough space for a stand off scene. I was a little disheartened that I did not come across my ideal location as this project would have worked a lot better in an industrial environment, this was all down to time management, I simply ran out of time to get hold of a better location. Other locations were looked into like a site on Brooksby Campus as well as Melton Mowbray translation although the use of a replica gun would have always limited the use of public locations.
I also only had myself to film and act in the scene as I couldn’t manage to find a film student / actor who had the time to spare. Never the less I went a head with filming / acting and pulled a series of shots together.
The matchmoving process was fairly simple as I only used one shot that involved motion all the rest were static shots as I knew the VFX would fit simply over these shots and saved on time for the project. I found that I had an error value on 2.0+ which in isn’t the best value and would cause my design to slip, again because of limited time instead of re-filming the scene I decided to shorten the clip so as that the error wasn’t as pressing.
I would need to first turn my 8 bit RAW footage that was filmed into Quick Time; I would then need to turn them into targa images which would be imported into After Effects; the VFX would need the robot to render out separately for the AO and shadows so as they are not combined over the video; I would then need to export the animation and targa from AE to Premier pro to then be exported to vimeo H2.64.
This project has given an abundance of information to me. It has shed light on the VFX pipeline making the idea of a VFX career more realistic, it has shown me the way through programs like Blender and After FX and has taught me things I never knew.
I would say my biggest issue was underestimating the time it takes to carry out a VFX project, with better time I could have brought together a number of other element that would have strengthened the project.
Also program settings has been a big issue. It is increasingly seaming vital to not only understand a VFX process but the program itself. As the smallest alteration can affect everything and if you do not understand the changes that have been made you will rarely come out of it ok. This was shown through the final design of the robot, the new unknown alterations would change the entire colour temperature of the robot, something unfortunately that couldn’t be changed without extra time.
I am happy with the finished project, I feel goals in planning and design are acceptable, the robot moves are believable and the added sound effects and particles bring it to life.