Pain, pain and more pain. Now with a week to go, I’ve had to stop animating in order to set up the scenes to render. Now is a matter of staying overnight and getting technical stuff to work without failure. I’ve been functioning on little sleep, so I’m trying to fight through the increasing fatigue to make sure we’re in a comfortable position to have a completed product for hand-in. It won’t be polished to the extent we want, but we believe it’s at a suitable position to call it day… for now…
I’ve been continuing my journey into the world of Cinema 4D by trying to learn “tags” so that cameras, lights and meshes are in position and look like they would if we’d rendered with Maya. The biggest difference with Cinema 4D is that meshes important as hard edges, so I needed to figure out how to convert them into a smoother mesh.
Beth said that it worked with the insertion of a “Phong”, but in my case, it needed seem to aid the situation. I did however link the cause to the number of faces within the meshes, leading me to an online forum that told me go towards adding a “subdivision surface” tag. This way, any mesh inserted below it would smooth via additional subdivisions (edges) added to the mesh. Initially, I did run into the problem of having one mesh that seemed to have a crazy number of subdivisions and it wasn’t linked to the tag, hence it was struggling to render the scene. I did eventually uncover a “subdivision weight” tag within the mesh, and by deleting it, my problem was solved.
I also reduced the number of subdivisions added to the mesh in order to avoid having too much to render within the scene. This way I wasn’t wasting precious polygons and render time, just adding it to the necessary meshes within each shot or visible to the audience. It came back to the need to optimize and be diligent about how much detail can be added before it becomes chaos to render.
I added in my mesh, wall and ground textures in addition to the new paint overs that Dan did of the wardrobe and drawers. Going back to the that time texturing the bedroom, I stated that I was going to back on the foundations to Dan so I could move onto rigging and fixing animatic problems. With that collaboiration, Dan and I were able to produce the bedroom in the style we wanted to achieve based on the short film The Unlikely Hero (2016).
Going off the feedback from our last presentation, I made the necessary additions to the lighting to get the nice haze effect coming through the window, in fact, I added an extra feel omni light with falloff to get some extra volumetric fog. As you can see from the figure below, I wanted to get a pass render of the extra fog to give it the camera lens flare glare from the original render test. This in effect would give us a substitute for creating particles in post-production given we just don’t have time at this point! I made the spotlight come in from the side to position the sun accurately and make the light on the bed come through more dynamically than dramatically positioning it in the centre. I also used a key point light in the centre of the bed and reduced the intensity of the brightness so that I could get soft shadows for added definition to the room. I ended up reducing the density of the shadows because they were too dark and non-realistic to the mid-morning lighting rig I was aiming to achieve. There was also an area lighting coming from the stairs so that I could give the impression that there is a downstairs given that there is nothing below the mesh. I had to create the illusion that this scene was connected to the living room below, thus maintaining continuity within the film. Knowing the accurate depictions of light sources in a naturally lit environment linked back to my research revolving the lighting rigs in The Last of Us.
In the figure below, you can see my approach to lighting. In previous years, my issue with lighting was defined by my lack of informative justification. In this project, I wanted to overcome this flaw by being careful with my lighting rig. As stated above, I was basing it around where natural light sources form from. Considering in this scene we only have one main key source of light (the window), I idea was to defuse and disperse the central spotlight. In the figure above, you can see how the spotlight has an infinite spread of light, creating an intense beam at the edge of the bed. On it’s own, it do not illuminate the room enough to give the impression of morning/daytime and the character’s expression will be unclear to the audience. As a result, the “fill” light above the bed serves the illumination provided by the spotlight and acts as a way to keep the character lit and in focus to draw the audience’s attention in the shot. Like in The Last of Us’ lighting rig research, it serves as a bounce so lights aren’t isolated and instead have more spread around the room. I created a falloff for the fill light to maintain the natural impression and avoid drawing attention to the artifical set up. This is then counteracted by the area light illuminating from the stairs to avoid making a ‘ring of light’ that isolates the bed. The idea is to create the illusion of a real house lit by real sunlight, hence I didn’t want to be appear dramatic and found the sense of isolation when it’s already established in the defuse of the lighting, the cold colours of the textures and of course, the composition.
This was a contrast to the living room, which needed more brightness given we had multiple sources of light from the windows, as well as my dad’s architecutral point that the front room should generate the most natural because it’s the most commonly used room.
In this case, the challenge came from being less in control of the light, but ultimately after trail and error, I began to notice that my 3-point lighting set up was the recurring pattern, so don’t fix what’s not broken I say. I used the front window as the key light and once again like the bedroom, used a spotlight to serve as the direction of the sunlight (in the image above, you are looking at a sunset, hence the source of direction is moved). Above the room is another fill light at serves to help illuminate highlights and give the audience a brighter image in which to viewer the room. I actually ended up changing it’s colour to adapt to the day/night setting. In day, I used white to keep the bright summer’s day alive as I found yellow just looked too artificial. At night, I used a very subdued blue to make the room feel cold, continuing the emotional aspects of the scene and implying the character not paying his heating bill in the same way he doesn’t pay his electricity!
I was struggling to use the side window on the left of the scene because I realised it was out of the direction of the sunlight, so it wouldn’t necessary project light. That said, my dad said you’d see have light making it’s way through, although it would be softer and by night time, it would be non-existent.
The reason the lighting set up became quite strict to like into my animation of shot 13, where the scene transitions from day to night. Effectively, in the animatic, the transition is from one extreme (day) to the other (night), so it doesn’t account for sunset. This got me a light excited because I wanted the light to be the artistic element of the shot. Turns out I had to brush up on my lack of astronomy knowledge to get the positioning of the lights correct. I researched how the sun transitioned to the moon, and through the coordination of lighting, if the sun is at it’s highest peak at midday, then the spotlight would be aiming downwards, but for sunset (at it’s lowest), the lighting would be centred along the horizon. By night, we once again reach the peak but the direction of the spotlight will have shifted. So in the video below (ignore the quality, still coming to grips with Cinema 4D), I basically animated the spotlight to move in a arcing manor reminiscent of the sun/moon position. Yet, that wasn’t entirely the challenge complete, I was careful how I chose the muted colours. Keeping white as day to keep it conventional, I wanted sunset to be calm, reflected in the warm orange, while the same emotional tone remains in the soft cold blue of night. In a way, despite the colour contrast, the emotional tone remained the same as the scene is supposed to transition the character from his frustrated and defeated state to being as ease and self-reflexive of his depressed state.
Lastly in the setup was simply getting the animation work… fun times. OK, so my work placement was in games development for Iglu Media, so when transferring animation into Unit in real time, I was baking simulation on the mesh because the rig doesn’t go onto the software. IT WAS CHAOS ASSUMING THAT THE RIG WOULD WORK FINE FROM MAYA INTO CINEMA 4D! I literally hide to follow a similar process to Unity when I wanted to avoid the pain and just have baked animation. Turns out (to be honest, I should have known this), Cinema 4D and Maya have different rigging systems, hence the rig doesn’t transfer. Makes sense, so to overcome that, a few forum searches and I found that if I exported the animation as an “Alembic” file, it would bake the animation to the mesh and all I’d have to do was import the animation into Cinema 4D… and there I was thinking it would be that simple!
Well, once I got to Cinema 4D, I thought if I matched the scale import settings with the living room, things would be smooth. For some reason, the animation was scale significantly smaller (by 100 meters). It looked fine until I moved the frames, then the animation didn’t work.
I founded my research to figure a technical solution, but couldn’t seem to source an answer. I did however remember that the mesh was in separate components. I actually thought if I was to combine the entire body mesh into one and then exported an Alembic, it would work. It did thankfully get the animation to run smoothly, but the scaling issue can down to a matter of estimating size the animation fit coherently to the model. At least we can be in a position to render.
This setup was a strategic means to have our scenes ready to go at any point. Now with all the technical details such as smoothing, lighting, import/export animation – we can put the technical stresses (bar the water simulation) to the side and focus on producing quality animation and textures in the short time left. I will be continuing to ensure my animations can complete whilst keeping the scenes in shape. This allows us to do render passes – if Dan or I finish a piece, we just slot it into the scene and hit render. It calls back to my experience learning Cinema 4D and the reasoning for doing so. I wanted to relieve stress from Beth by being trained in an area no other member was familiar with. With Beth free to make good simulation and sky paintings, we are not letting technical tasks go incomplete or forgotten, and I can focus on getting the render passes done and edited, letting Dan prioritise the animation and Nathan on textures. It was clearly a stressful experience that payed off for the team and for me personally. I learnt an entirely new software from the ground up, overcame recurring technical challenges, and with the time saved, we can looking to have the project complete for hand-in, eradicating the growing concern there was too much to do. Each team member is focused and the flow is there, and we are communicating round the clock to assist in every key aspect of the project.
Aron Matschulat Aguiar (2016). An Unlikely Hero. Available from: https://vimeo.com/175446099 [Accessed 10th October 2016]