There are many important techniques, ideas and practical processes involved that enables the digital visual effects industry to achieve a desired outcome but the fundamental rules that everything on screen follows in relation to movement remains the same. From real live action footages to animations done on computer programs, the laws of movement follow the same path whether it be a character or an element. In accordance to these rules and with the help of research and practical work, the visual effects industry has been able to accomplish a lot in bringing imagined spectacles to life in a computer-generated world.  Visual effects is a process that is used to help re-create an idea which is harder to achieve using conventional ways e.g. of special effects, make up and animatronics. It can also be hard to re-create ideas on large, unusual landscapes which can be very costly and there is also the question of the safety and security of the talent and the crew involved in the film making process. For a long time in film making, the film industry has depended on the use of special effects. An example of this would be Star Wars Episode 1 (Star Wars Episode 1: The Phantom Menace, 1999) where a puppet for the character Yoda was used as the physical appearance and the mannerisms of that character would be too hard to achieve using an actor. We can see in the scenes involving Yoda that the movements are not as fluid and seamless as one would expect in today's time. The character is also not very “believable”. Today, one of the main goals of the visual effects industry is to make the computer-generated effect, environments and characters as real and visually convincing as possible for the audience enticement. This also helps greatly with explaining the films narrative. Most of the entertainment media today use visual effects an example being Marvel Avengers (The Avengers, 2012).

Coming back to the point mentioned earlier about the fundamentals of animation, in 1981 animators Ollie Johnston and Frank Thomas introduced the idea of the twelve basic principles of animation in their book Disney Animation: The Illusion of Life Johnston and Thomas (1981). These principles talk about the basic requirements of any animation, no matter how small or big. Namely these principles included: Squash and Stretch, Anticipation, Staging, Straight Ahead Action and Pose-to-Pose, Follow Through and Overlapping Action, Ease In, Ease Out, Arcs, Secondary Action, Timing, Exaggeration, Solid, Drawing and Appeal. This idea has been developed and taken into the world of visual effects as well and is not only limited to computer generated actors and characters but also to elements like water, snow, gas and much more. In this essay, I will be focusing on the process of simulation of one these elements. I will be recreating a snow fall scene. I am going to talk about how the snowfall can be simulated using the visual effects tools and will also comment on what hurdles that I faced when trying to achieve the desired result. For this project, I am using Autodesk Maya and Premier Pro in order to achieve the desired results.

​

Expanding more on the topic of this essay, digital effects simulation vs real life special effects, in one of his books, Special effects: new histories, theories, contexts, Dan (2015) argues that the river of slime scene from the film Ghostbusters II could have been simulated with the use of digital visual effects but due to the budget of the film and the availability of an inexpensive substance methocel, the film took advantage of these factors and used the substance instead. This can suggest that in the past, digital visual effects was an expensive and less competent tool to use. However, even if that was not the case, the director decided no to use the digital effects but the actual special effects substance. Some of the reasons as to why methocel has been widely used in the film industry is due to its controllable nature .e.g. viscosity and colour. Dan (2015), also states that digital simulation and special effects can very well go hand in hand to help achieve the desired effects e.g. the egestion of Jon Voight by a giant snake in the film Anaconda (1997). The snake was rendered and simulated in the 3d program however methocel was used as special effects for the reptile’s saliva. The simulation tool in digital visual effects is an old practice .e.g. the simulation of water in the film The Ten Commandments (1956), the scene involves the actor playing Moses parting the red sea in two. This involved a technique of reversing a pre-recorded footage of water being poured into a container. Another example is of the film Life of PI (2012), where the whole sea in the background has been simulated using digital effects green screen and can be seen to be very realistic.

​

As mentioned earlier for the practical side of this research, I decided to re-create a snow fall scene and compare it to two live videos I have filmed myself, exhibit A and B. I chose this method so I could simulate a real-life element, tinker with its behavior in the Maya software and then compare it to the two real life footages of the same element to decide what the visual differences are. First, I went into the software Maya and created a background relative to the desired outcome of a snowscape. I created a large snow dune and made different waves in it emulating a patch of land covered in heavy snow. The outcome was good as the landscape is quite realistic and believable. I then moved on to making the background sky which would also be the light source for this scene. I chose it to be a morning scene as we can see the sun rising up from the left side of the screen. Next up I started the main part of the practical which was the simulation of snow. For this I have created an emitter in scene and moved it to the top of the dunes out of the camera’s perspective. This was to make sure that the seamlessness of the whole scene is maintained as the goal was to achieve a snowfall scene that has realism to it and is convincing enough for the audience. Then I went to the properties of this snow emitter in the Maya software and played around with the settings. The main settings to focus on was the gravitational pull of the snow, the emitting rate, the speed and spread of the particles. At first the gravitational pull and the speed were both out of balance, I had to amend that and bring them to work together. I did so to make the snow particles would hang and float in the air for some time before falling to the ground. The emitting rate was to control how much particles were being produced and speared across the scene. I decided to go for a medium spread. In the program I was also able to control the size of the emitter but not the particles. This gave me the ability to control how much area in the scene I wanted to cover using this technique. Then I finally used the randomness option in the particle emitter setting which gave the snow particles random movement. This was a very useful tool as it added a wind effect to the whole scene and the particles. This was done all to mimic a real snow scene and bring it as close to the actual footages as possible.

​

Now I will talk about some of the hurdles I have faced in the practical research. First problem I had was with the rendering of the whole scene. The process took around two and half hours of constant system usage and of course if something goes wrong or the outcome is not as desired then the whole procedure needed to be repeated which in fact did happen a couple of times. The first reason why I had to repeat the whole rendering procedure was because the output file format was not properly selected. Instead of using the .ext at the end of the file name, I accidentally selected the .ext first then the frame number .e.g. name.ext.#. This file name format was required to achieve a video footage as I needed to take all the rendered frames to Adobe Premier pro and clip them together to create a continuous scene. Moving on, the second big issue I face and could not resolve was the watermark for the rendering system in Maya known as Arnold. To avoid this watermark issue, I decided to render the whole three hundred frame scene once again using other renderers present in Maya. However when I tried using these other renderers, although the watermark disappeared, it gave me a very unrealistic result and completely omitted the background, the snow particles and the sky  leaving me with the texture less snow dunes and a black back drop. This made the simulation appear far less convincing. Another Issue I faced with the emitter particles system was the particles being too smooth at the edges which fair enough is sometimes the case with the actual snow particles however, if I could add a bit of a roughness to the shape and randomness to the size of these particles, it would make them appear more realistic.  The particles system also didn’t allow me to make the landing snow to appear as melting in the dunes. I also encountered some crashes while creating and rendering the simulation which forced me to load the whole file again and restart the procedure at hand. All these issues combined proved to be very time consuming and I had to go back and forth to address and amend these technical problems.

​

Watching both the real and the simulated snow fall footage, I was pleased with the results. The simulated footage is quite convincing as the rate at which the snow particles falls, the randomness in the wind of the scene and the spread of the particles are all the factors that make this scene visually convincing. The camera positioning for the re-created simulation and the actual footages is also very similar as we can see in both videos, the particles are falling in the fore front and also towards the mid ground and the back ground. The actual footage (exhibit A) does have different sizes of the particles and so does the simulated snow scene but as mentioned there is a roughness to the particles of the snow in the actual footage which gives it a slight advantage. However, there are also a few drawbacks. The minor one being the speed of the particles. In the real footage, we can see the speed of the particles are all different to each other, some are coming down slowly, some faster. In the simulated version, the particle speed is all the same which gives it a slight artificial feel. For the major drawbacks I consider the watermark to be one as I was not able to remove it and the incapability of the other renderers present in Maya, as they were not able to give me the desired result.

​

I then decided to shoot another video of a snowfall (exhibit B) on a bigger scale and compare it to my simulation in Maya. In this video the first thing I’ve noticed is the disturbance in the air. The number of snowflakes due to there large quantity have made the scene very unclear, which I have not been able to achieve through he Maya snow emitter system. There is also an aspect of more randomness to the wind direction that is affecting the snow in this video. In one frame the flakes are coming towards the camera but right in the next frame they are going away from it. Again, the different in the size of the snowflakes can again be noticed in exhibit B. The snowflakes are also very rough towards the edges and very random in shape. In comparison with my Maya simulation, I was not able to create a vast number of particles as it was very limited with the options as I have selected the highest possible number of snowflakes/particles which still seems to be very less in contrast to exhibit B. I also noticed the speed of all the particles was different to one another due to the difference in their sizes and shapes.

​

Next up, I decided to investigate the industry standard special effect used for scenarios that include snow fall. The tools used is known as a snow machine. In the video provided (exhibit C), it is clearly visible that snow machines are able to re-create a snowfall very effectively. Not only the flakes itself but the machines are also capable to cover a big area and provide the snow residue on the ground depending on the needs of the film and narrative. Comparing my Maya snow simulation to exhibit C we can clearly see the use of the snow machine would provide better results as it is much more convincing and also is practically present in the real world, rather then being rendered on top of the scene. These findings of mine sound very similar to Dan (2015) where he mentions the use of methocel as slime in a number of films rather then it being generated in a computer program.

​

In conclusion, I would say there is still role for special effects in the film industry as it is evident based on my findings. Although saying this, I do believe with the proper use of other computer programs and more knowledge on the topic of simulation, snow and any other kind of simulation is possible as this can be seen in the early example I used of the film Life of PI (2012). But as a final an example to support visual effects, I have taken a small clip from the film Frozen (Frozen, 2013). Looking at this clip provides a strong argument that simulation of snow is very much so possible and can result in a very convincing effect on screen. However, as Frozen is an animated film, there can be a counter argument that the same result might not be possible in a live action film.