Esme’s Blog

Over the past 23 weeks, I was able to, I feel, successfully make a documentary which utilises and adapts the animated form to express itself more subjectively. Since animation is a medium I feel should be used in ways that go outside of what can be achieved with the live-action camera, I felt this to be particularly important when portraying interviews adapted from narratives of real people’s lives.

Final Film

Below is the final film, which presented 6 people’s experiences during the coronavirus lockdown in interview format. It uses visual metaphor to encapsulate how the home became an extension of ourselves during this time period by anthropomorphizing the interviewees as architecturally location-relevant houses.

Showreel

Below is the showreel, which delves into some of the processes that helped create this film and highlights the different key stages of production.

Critical Reflection

Retrospectively looking at my film, while I feel it achieved what I had initially set out to do, there are various aspects that I would change in terms of technique and visual presentation. My main regret and focus with the film is that the animation could not be as detailed as I wished, with a lot of sacrifice of secondary animation due to rigging technicalities and time constraints. With more time, I would have loved to explore exactly how each of the different characters would have moved and created individual walk cycles/ character animation tests to see how they are all personified through movement (i.e. blinks with curtains shutting etc.). Secondly, I feel there was a lot of potentials to explore more rendering capabilities and texturing. As mentioned some weeks prior, I wanted to explore how the toon shader may have looked with bump maps and different mesh deforming techniques to try and get the impression of further material, which is something I intend to explore in further depth going forward with my own personal modelling. I feel the blend of using a toon shader and additional texturing techniques is what really sells it within the style, and I look forward to exploring this further. For this film, I also feel I could have researched further into 3D lighting for Arnold as I struggled to adjust each individual environment to portray their time of day effectively. I feel this is particularly highlighted in the scenes at night, as I could have done more with the inclusion of lights with the stars and the moon.

Narratively, the piece seems to work together in humorous aspects, with a few hidden references to other characters and the idea they all exist in the same world; however, in future, I really wish to push for detail in my project, working far more extensively on creating detailed models to see environments. This included researching more into simulation techniques such as X-gen, as I was unable to render ‘wind’ simulations and collisions due to time constraints. I would have also liked to have done more research into particle simulations for the clouds and smoke, and seeing if this can be paired with more stylised aesthetics.

In order to research techniques and considerations further, I have created a list of books which may assist in furthering my knowledge in 3D for future projects.

Further reading:

• John Wiley and Sons Inc. (2013) Maya Visual Effects The Innovator’s Guide. Autodesk Official Press.
• Abishek, K. (2021) Immersive 3D Design Visualisation: With Autodesk Maya and Unreal Engine 4. New York: A press.
• Beane, Andy. (2012) 3D Animation Essentials. J. Wiley and Sons: Indianapolis.
• Giambruno, Mark. 3D graphics and Animation: From Starting Up to Standing Out. New Riders: Indianapolis.

shot list

Nearing the end of all of my shots, I feel the use of this shot list proved immensely useful, especially in keeping track of renders and render times. This proved especially useful as several shots had to be re-rendered or are currently being re-rendered. One such shot is shot 13, in which there is a slight collision between the floor lamp and the wall. While this adjustment will most likely not be fixed in time for submission, I hope to rectify this discrepancy for my film going forward.

For the final shot of the film, I really wanted to utilise the lighting and time of day to act as metaphorical implications that support the spoken word. The last words of the film are “because if we’ve been through this, there is no way we can’t get through more”. By making the setting a morning Dawn scene, not only does this add more visually interesting implications that the character has been abducted all night is is now returning home by morning, but it also symbolises a new beginning and the end of a period of time (in this instance, lockdown itself). I wanted to end the film in this lighthearted and hopeful way to leave viewers optimistic about the future. 

Due to the fact I have spent the past few weeks animating these characters, When approaching this shot, I felt very familiar with what to do and how to clean it up so that the process went a lot quicker. Initially, during the descent from the alien beam, I wanted to make the character flip upside down and fall on the ground. However, I did not have enough frames within the audio to sell that convincingly therefore decided to make the character land quite gracefully with only an impression of the weight returning after the beam had been retracted.

Trying to get a Dawn break look within the lighting of the scene proved to be quite tricky, with the initial sky dome light appearing more like a sunset in its darker, more orange saturation. Looking at references of the morning, I noticed that there is generally more of a “pink” saturation to the sky, with strong orange rim lighting on clouds that reflect from the sun. I feel in order to get the dramatic last shot, I will need to add a strong orange rim lighting effect to make everything blend more effectively in the scene. 

As highlighted in the playblast below, I attempted to experiment with how the sun rays would move and affect the lighting; however, going forward, I feel this will be much too complicated due to the fact that animating the mesh lights creates a lot of render times that will prolong my complete the whole film in time for the exhibit.

Since the camera pans back to the sky, I wanted to put the plane that has been frequenting the scenes into the shot once again to make a direct reference to the previous shot’s characters. Below indicates the finalised animation, including camera motions.

Now that all of my shots are animated, I put any remaining shots within the university-associated render farm. While I was initially having issues surrounding the use of it, which was prolonging my rendered content from being completed, in the last week, I got assistance and was able to get it working and ready to go for the heavier shots such as this one.

During the editing process, I explored different effects to try and make apparent the forced recorded nature of the interview and bring that into three-dimensional space. Typically during Amateurish interviews (such as my own), there is a lot of camera shaking due to a lack of tripod equipment and visual indications of the camera used to record (such as borders on old bus cameras). I also feel with the inclusion of phones into everyday social relevancy, there is a lot of video footage that would not be done on professional cameras. Trying to explore this within the realms of my one film, I added additional borders to accentuate the technology used to “record” and “capture “ these anthropomorphised houses. However, I felt these effects were too jarring from one another and distracted viewers from the main content and animation, therefore scrapped the idea.

Since the film intends to be displayed in an exhibit, I wanted to enable accessibility options, so in order to do this, I added some subtitling so that the content is clear in case any hard-of-hearing people wish to view the piece and have a contextual understanding. I also feel this will be good for exhibit purposes as only two people can use headphones per screen, and it allows space for more people to watch and understand the animation without inherently needing to hear the sound. While the sound does enhance the piece a lot, especially with the use of the Indexical voice of all my friends, it is not necessarily needed to get the idea across. 

Shot 16

Moving onto shot 16, the animation involves the ‘lack’ of gravity, so to speak, so the character is floating in mid-air.

The main thing I wanted to get right before starting the animation process was getting the environmental set-up and lighting correct since this is the key feature of the shot. My initial step was to create an Arnold area light that would act as a spotlight. As similarly utilised in my skeleton project last academic term, I adjusted the exposure and the spread to make it have a smaller focal point on the character. However, due to the lack of flooring and mesh for the light to fall on, this effect does not appear very obviously like a ‘spotlight’ therefore, I returned to the process I had created earlier in paring this light with a cylindrical cone that has heavy transmission applied to it.

Putting this cone in, paired with the light, created exactly the effect I was looking for and that I used in the previous shot. While I want it to look stark against the background, I feel the lack of anything in the back does not add to the effect and will need an additional AI skydome and directional light to create a strong rim lighting effect. I also wish to add stars as mesh lights similar to that of shot 11.

When Initially using a ‘transparent’ aistandard surface, it created a warping effect which made the house look like it was in a cylinder of water rather than a beam. Therefore I changed the material to a lambert as this proves more simple and effective as a ‘light’ beam.

When previewing the test render, I noticed that the strong directional light was casting a shadow onto the cylindrical shape, which is supposed to represent a strong beam of light. I feel this effect would bring viewers out of this perception slightly and ruins the overall look. In the cone meshes attribute editor, I was able to untick cast and show shadows in render, which removed this effect quite simply. During the render process with the shots that contain the “alien abduction” sequences, I will be sure to look carefully and bare this in mind. 

In order to get the area light to follow the cylindrical shape when panning across the screen, I used a parental constraint so that it would not get left behind and ruin the lighting in the render.

The next most essential step was making the whole environment look convincing as a night sky, with layers of clouds and stars that would add an overall depth to the shot. In order to get the mesh lights to reduce noise, I had to ensure that the samples for each individual light were up to about 5-7, so that render glitch discrepancies didn’t happen.

When modelling the stars, I wanted to add a little bit of variety once again to the purely round spherical mesh lights, so I created a ‘North Star’ Style of the model to help stylise the background further. I also feel with the addition of extra lighting will add reflections and additional lighting that will aesthetically enhance the render.

As previously mentioned, since this shot did not have much gravitational weight for the character, I had room to work with how I would stylistically make the character move. To get the feeling that they had a slight lack of control, I wanted their legs to hover in the air, almost as if they were swimming. I also wanted them to float up and down as if the ‘laser beam’ was fighting to work with gravity.

Improved render

Revisiting the first shot, I noticed there were several aspects of the render that needed improvement. First of all, during this render, I did not add rim lighting to the AI toon textures, which caused them to look bland and not blend into the overall scene. As you can see in the image below, on the right is the older render, and on the left is the newer render, which I feel fits much more cohesively together as a film. I also removed the secular from the roof as I felt that unless wet, roof tiles do not generally glisten to that extent, and It proved to be quite distracting from the rest of the animation.

Shot 17

The last Shot I did this week was shot 17, which is the second to last shot of the whole film. In this shot, one of the characters who were previously down the hole uses a ladder to climb to the other side of the cliffs, in which the two characters stare at each other across the gap.

Since this section relies on somewhat comedic timing, I felt that getting the camera pan correct was the initial step, as the camera needs to move in sync when the dialogue says “from a distance” to help exemplify said distance between people.

Below is the finished shot, which I feel works successfully in achieving the comedic timing of the additional character reveal. However, I feel if I had additional time, I would work more on the motion itself, as I did not have a lot of time to clean the animation in the graph editor, so for the most part, it appears very jerky and rushed. Going forward with the animation in future, I will rely less on Maya interpolative feature to create smooth animation, even under time constraints.

Shot 12

Beginning shot 12, the only key elements within this shot I needed to model were the cardboard boxes. I Aimed to keep this low poly to speed up render times due to the lack of time constraints and also to upkeep the simplicity of the visuals. I made the box with the ability to open the different ‘flaps’ of the tops so that I had the potential to add secondary animation also, Arnold’s counter renders filter will display the toon shader lines, giving the impression of the model being a cardboard box and not a smooth square.

In order to sell the height and dramatism of the shot, I wanted to make sure the camera moves from panning upwards to panning downwards to get the impression that the tower is taller than it is. I also wanted to give the impression that the tower was very unstable and insecure, much like how the situation over lockdown was. Especially since this part of the dialogue details trying to find makeshift workspaces to work from home in.

Much like the earlier shot involving the globe, I wanted to have a very simple background that accentuates the foreground. I also used the same colour background as the earlier shot (Shot 7) so that these two would match aesthetically and tonally. The key part about the lighting was to make sure that there was a strong directional light that would make the rim lighting visible on the edges, so the foreground assets would blend more seamlessly with the background. I also needed a strong light so that the facial expressions were visible.

Below is the finished shot; I think the concept for the shot works well and gives the impression of cluttered chaos. However, there are several aspects of this shot I could improve. My initial comment would be that the motion of the lower boxes is too stiff and uniform in motion. Looking at the reference video below, there is a slight offset, which I feel is due to the uniform way I parented the boxes, I was unable to create this effect.

Shot 13

Similar to shot 9, shot 13 is located on the interior of the house, and I followed a similar method and used the doors at offset times to accentuate different parts of the speech in a more visually dynamic way. For example, when the interviewee says “Its just finding the routine to do” I used the left door, and when he says “um” quite quickly after that I used the opportunity of having two doors to stress the jaw bonces and not have the ‘mouth’ move very quickly open and closed like it does in some shots which looks unnatural, even for a talking house. Improving on this shots animation however, I feel I could have done more with eyebrow movements because I feel at times they moved to robotically and uniformly, even though the implied material of them is bricks. I think it is important to find a middle ground between natural expression and the context of the implication of material the expression is being carved out of (I.e skin or Bricks).

Re-rendered shot

As highlighted in the previous shot, the lamp in the original render was slightly clipping through the wall, so to fix any of these discrepancies, I re-rended this with the lamp moved to a more appropriate location. This will aid in improving the host quality and selling the animated world in a more professional light.

Shot 5

For shot 5, I really wanted to explore the environmental terrain by keeping an aesthetic consistency of using spherical shapes. This was to accentuate the juxtaposition of child-like ‘safe’ looking aesthetics with quite darker, more serious tones of people discussing their experiences with the lockdown. In the initial environmental model, I thought it looked quite bare in spite of the use of X-gen, and I felt like it needed the background to further accent the key foreground elements.

Looking at the Tuscan landscape, since the house itself is based on the architecture in the area, many of these rolling hills feel very gentle and green. Translating this into the environmental design, I tried to create this smaller, enclosed area utilising the reference imagery to accentuate the isolation and feeling of confinement this particular interviewee experienced.

As seen below, the initial render test had a similar colour scheme to what I wanted to achieve, but there is no colour in the background, and the frame is highly overexposed in a way that washes everything out. Due to this lack of colour in the sky, I am not able to create rim lighting that would blend all the mesh and textures into the scene together. After the animation process is complete, I will then work on adjusting the lighting of the scene to be reminiscent of a warm Tuscan location.

After creating the environmental model, I moved on quite early to the animation stage, and the first process of lip sync was the ‘jaw bounces’. While In Kenny Roy’s book it states that you know, think about the way words are being said rather than just simply moving eh mouth in time with the audio, I still wanted to keep a level of stiffness to sell that they are houses (2014).

Below is the finished rendered animation. Due to the heavy use of X-Gen, lighting and mesh the render time took several days before it was complete. However I allowed contgency time for the week so that I gauge how long an intense environmental scene will take, especially going into shots 11 and 19. Shot 11 and 19 also have the inclusion of heavy lighting that is replicant of a ‘laser’ beam. Retrospectively looking at my animation in this shot, I feel that I could have spent more like smoothing the motion out as it appears very jerky at points.

Shot 8

For Shot 8, I wanted to replicate an environment that would be rememniscent of side scroller layouts in video games, especially that of the super Mario franchise. By using this association, especially with the kinetic sign that points towards a ‘hole’ it feels symbolic of the distance many people experienced from reality during lockdown, and also using a bright colourful aesthetic to distract and juxtapose from the spoken content.

Using the previous tutorial of creating X-Gen Grass, I made a new description to enhance the ground more and to match aesthetically the previous shots. I also feel this will help accentuate the ‘underground’ area.

Lighting

Since I wanted this scene to be brightly lit and to have a ‘sunny’ appearance that would contradcit with the darker, dingy looking ‘hole’. Since one of the interviewees states how they were stuck down the “hall” with a ‘random guy’, however since the audio quality isn’t the clearest and the interviewee has a slight accent, it also sounds slightly like ‘hole’. playing with this is decided to make a visual joke out of it, which led to the inspiration of the environment.

With the addition of the rim lighting, the character and environmental assets blended into the scene much more seamlessly, and due to the blue saturation of it, achieved a ‘cherrful’ sunny look that I was aiming to achieve.

When initially lighting the underground area, it was very dark and nothing in the scene was visible. In order to get the space to be enclosed to get a top down lighting effect, I had to create an enclosed piece of mesh that wasn’t open like a side scroller of a ‘dolls house’. This meant that I had a slightly limited space to work with the camera however added to the overall lighting.

Due to one of the characters being a ‘skyscaper’, I thought I would utilising interior lighting implications to add an additional light and glow to the scene, to make it stand out more. I also feel this utilsies the character designs to a better extent.

Below is the finished shot, which I think has the inclusion of a well timing camera move. However, reflectively, I think that the walking animation is not very neat and does not quite match the timing of the audio and therefore the camera. At times there are also noticeable knee-pops which indicates I should revisit the graph editor and ensure everything is smooth. I also feel i could have played more with the weight of the upper body, as it is massively disproportionate to the legs and could have been a really good opportunity for a more interesting, weighted follow through. I also feel that some of the motion is too ‘floaty’ and doesn’t convey the right amount of weight in general. Since this shot has a very long expected render time, and it currently still not finished, I will add these improvements to be added provided there is additional time for me to work on it.

Shot 11

Using the same environment as in shot 5, I changed the environment to create a night time setting that would appear more striking and allow for more contrast when visualising an ‘alien abduction’.

Trying to ehnace the setting and accentuate the time of day, I created some stars in the sky and turned them into mesh lights with a very high exposure so they would appear to be ‘glowing’. While I considered and testing making the beam of light a mesh light, It would not have created a transparent mesh, making the character not clearly visible.

A key part of this scene is the lighting aspect , since I wanted to add this very dramatic green hue that would illumate the scene. In doing this I added a very high intensity are light and used the spread so that it covered essentially the entire grass. The green environment also helped to enhance this overall.

In the video below is the completed animated shot, in which I think the eye darts and eye animation was quite successful given the lack of things to make character expression with. However, again I feel some of the motion is a bit too stiff and also in some parts too floating without clear character motivation. If given the time, I aim to improve this along with the earlier shots.

References:

• Roy, K. (2014) How to Cheat in Maya 2014: Tools and Techniques for Character Animation. Focal Press.

This Week, I Planned to create shots 7, 10, 14 and 15, the most animation-intense week Planned. Due to the sheer amount of animation needed to be created this week, I started modelling environmental assets right away,

Shot 7

Beginning shot 7, I knew I needed to model a globe that gave vague impressions of countries. Due to the simplistic visuals of the film, I did not want to perfectly recreate a real globe with all the individual countries and instead used big pieces of mesh to imply a land mass. However, In order to successfully achieve this, there needed to be a variety of shaped mesh. Since the character who will be present for this shot resides in Canturbury, I wanted to give a vague impression that they were situated in northern Europe without modelling too much detail to keep the low poly effect.

When applying the toon shader to this model, I feel it really perfectly matches the low poly aesthetic of the film, and the shaded lines really help accentuate the colour.

While I said I wished to place the character in a rough northern European location, I feel that due to the nature of the camera pan I have planned, it needs to be on a larger plot of land rather than a smaller island so that surface gives the impression of being flatter ground without revealing the globe model until its zooms out.

Looking reflectively at the previous render test, I also felt that it would benefit from the rim lighting effect in the toon shader attributes to make it stand out from the flat, monotone blue background. This should also assist in accentuating the 3D nature of the assets.

Due to the low poly assets and referenced model, the rendering process went very quickly and smoothly, and as showcased below is the finished shot. Looking at the animation for this shot, I feel it is the most successful I have finished so far purely because of how expressive it is and how well it matches the spoken content. If I was to revisit this shot, I would focus again, particularly on one eyebrow movement that seems to snap down slightly too fast in a way that feels unnatural.

Shot 10

Moving onto shot 10, Since I already had a clear layout for this due to the previous shot last week, the main difference I had to make in terms of setting up was adjusting the lighting to change the time of day. To add some more dynamic visual shots, I wanted to create some visual difference when audiences are brought back to the interviewee and also give an indication that some time has passed in their location. I also wanted to use this as an opportunity to explore different lighting sets up and how I will create these within 3D space. I also feel different lighting and time of day match different tonalities of speech. For example, as this house, in particular, gets more serious.

For the facial animation of this shot, I followed the reference footage quite closely, which I feel is clear in the expression and timing of the character of this shot in particular. I am especially happy with the eye movements, as I feel they time very well with the overall speech.

Reflecting critically on this animation, however, I feel that, again, the eyebrow movements are slightly off and, at times, float a bit uncomfortably due to a lack of adjustment in the graph editor. In order to fix this, I will ensure that there is a flat tangent when trying to hold a particular shot so that it appears still.

Shot 14

Moving onto shot 14, the main asset I needed to create was a small dumbbell so that the character could pretend he is under a lot of strain despite its small size of it for comedic effect.

Since the original environment I created is quite bare, I decided to re-create the X-Gen grass with the newer, more effective method I have learned, and hopefully, it will cover more surface area since it will be generated in metres rather than centimetres.

As is quite visually obvious, the newer testing looks much more effective in general, and the render with the toon shader applied still does not wash out the overall colour. Since the contour that is being rendered within Arnold for this texture is a dark green, they all appear to blend into each other in a way that does not bring attention to the outline.

Beginning the shot animation, I wanted to make sure there was a conveyance of heavy weight when the character is holding it, and to visually empathises with this, a puff of smoke comes out to be symbolic of a deep breath and energy exertion. I also used the eyebrows to accentuate the weight of the object but make it appear like he is struggling to lift the weight up while speaking. While it would have been funnier, I feel if the interviewee sounded out of breath, I feel it still worked nonetheless.

Since I had already expanded and improved the environment for shot 1, instead of remodelling it, I referenced the finished animation of shot 14 into the Maya scene, in which I was then able to render the model in the same place, making the shots look more consistent.

Below is the finished shot in the new environment; I feel that animation-wise, while it conveyed the general idea, there are several weight transference issues in the animation. I feel this is particularly highlighted when he leans to the side slightly, and the weight doesn’t give a heavy appearance due to the way it balances and shakes off the side of the roof. There are also several issues with the look of the render; for example, the cloud clips into one of the trees, and also everything appears much darker and saturated than in the test render. When revisiting this shot, I will re-render it will less specular on the roof tiles, a brighter situation and more consideration of cloud placement within the shot.

Shot 16

Following on again from the previous, I followed exactly the same process of following the reference footage as a rough guide for the eye and eyebrow motions and used the body language of the interviewee to base the motions of the building.

When rendering this shot, I initially had some stars in the sky however, even rendering with a maximum number of samples on all the lights and in the render settings, I was getting a glitch where these white specs would appear sporadically on the screen due to overexposure of lights. Due to this, I created a render in which the only mesh light is the moon, which I will enhance in post-editing to appear like it is glowing. I also feel the render overall is quite dark and will need lightening in post-production to allow for facial expression clarity. Given the chance to re-render the shot, I will also add rim lighting that adds a glow to the edge of the buildings to make the moon’s light look stronger and also separate the foreground from the background.

The finished animation, I feel, is also quite successful, however, I feel that, again some of the eyebrow motions could be timed with the rest of the motions more successfully. Going forward with this rig I will be more careful and try to allot more time to perfecting the expressive animation.

Shot list

Continuing with the animation process, this week, I aimed to complete shots 3, 4 and 6. Since 3 and 6 are both set on a Floridian beach, I thought it would be practical to do both of those shots within the same week as, in theory, everything should be set up in preparation for it. Shot 4 is also quite minimal and simplistic in its approach, so it will not prove too much of a challenge in regard to modelling.

Beach Modelling

There are several assets required for me to make a convincing beach, such as palm trees, deck chairs and, of course, the ocean.

Palm Tree

Looking at images of palm trees, I wanted to look into ways to create some lower poly trees, to match the visuals of the previous. Models and also to reduce render times as much as I possibly can. In order to achieve this, I found the following tutorial, which was very useful in creating the overall effect I wanted.

The first step in this process was to create a ‘tower’ from extruded polygons which will act as the base of the palm tree.

The next step was to create a bezier curve, which will act as a base shape that the palm tree will conform to get the curved shape.

However, When I attempted to use the cure warp deformer tool, which would allow me to utilise this effect, it did not work for me as It did within the tutorial. However, When considering a solution for this, I recalled that due to the nature of beaches being windy, the trees would require some passive animation so that they would blow gently in the breeze.

Problem Solving:

Going forward with this idea in mind, I recalled some methods taught in tutorials over the postgraduate course in which we rigged a tube to bend and warp in a uniform way. Applying this, I created some hierarchically affected joints so that they would move in a curved more circular manner together rather than stiffly one at a time.

After creating the joints, I created a very simple skin binder that allowed me to shape the tree trunk to the joints in a way that did not overly stretch and confront them to appear unrealistic. In doing this, I used a closest joint hierarchy bind.

Since the tree currently had no leaves, following the tutorial mentioned earlier, I created these two low poly leaves, in which the multi-cut tool was applied to create little ‘cuts’ into the mesh to create the impression of large leaves.

After using the vertices to snap these to the top of the tree, I combined all sections of the mesh so that it moves as one and added additional joints to the leaves so I would be able to manipulate these in correspondence to the rest of the movement of the trees.

Using the reference to a traditional beach deck chair, I wanted to create a more visually stimulating background than the white plastic ones that are typically seen in more beachy resort hotels.

After creating the two ‘wooden’ sections of the chair, I used the boolean tool to create some cylindrical indents into the wood to give the impression of the chair height adjustments. I feel that a lot of the key elements of 3D modelling encapsulate the small details in a simplistic way, and I felt this to be very effective.

After finishing that section, I created a flat polygonal plane and used the soft select tool to shape it around the existing mesh to give the impression of a cotton-like material.

After applying the AItoon textures, I feel this model encapsulates exactly the reference material I was following and also aesthetically matches the material banners on the side of the beach cafe character who will populate the scene.

Water Simulation

The most essential part of creating the beach scene, I feel personal, Is the water simulation. In order to once again create a low poly beach to consider render times early on, I followed a tutorial from the same creator who very efficiently explained how to go through the process.

After using the texture deformer on a polygonal plane, in order to get the water to be as low poly as possible, the next step was to triangulate the mesh.

after this, by using the poly reduce tool, I was able to turn down the polycount by 90%, having much more minimal surface faces to render later on after the animation process.

Since a key part of my films, aesthetic considerations were to get this toon-shaded look as distinctly as possible, I attempted to get this look onto the water simulation. This, I feel, was particularly effective as the motions of the waves are gently implied by linework rather than entire parts of the mesh that present themselves in a way that matches the world I am creating.

When combining all these elements into the scene, I feel this had a very effective overall look. I think reflecting on how this environment could be pushed even further aesthetically would be considering how to make different ‘sand’ simulations rather than soft selecting a single plane to create a slightly uneven surface. However, I feel for the simplicity I am trying to achieve overall, It works quite well in regard to visual identity and cohesive appearance with the film as a collective. Since a lot of the other scene have the inclusion of X-Gen grass, I feel adding the water simulation helped balance the detailing in the scenes in regards to passive animation and environmental depth.

Now that the environment has been considered, I placed my character rig in the scene and did a render test to gauge how much lighting/ sampling would be needed for the render later on. When producing this test render, I also noticed that the glasses indicated how bare the environment was on the other side of the camera, and I wanted to rectify this. In order to better self the universe, I added some of my other rigs into the scene to give the appearance that they were ‘recording’ the character as they were discussing their experiences. I feel this adds another level of detail to the shot that improves the overall quality.

Shot 3 Animation

Moving onto the animation process for this shot, I wanted to make sure that, despite the character’s very short design and square body, there was enough movement to symbolise and enhance expression. In reference to the classic flour sack animation task that is referenced in The Animators Survival Kit by Richard Williams, in which expression and movement are attempted to be displayed through minimal human appendices and facial elements. Considering this, I feel that the same principle can be applied to my houses as they only have short legs and minimal eyes, often without eyebrows.

In regards to the tree motions and passive animation, I essentially Keyed all the trees simultaneously and then used the dope sheet to offset them, so they moved at slightly different times to get a more naturalistic look. I feel this was quite effective; however, in order to improve the look, I would create a rig in which I could control each individual leaf and time them correctly in a very specific direction to give visual impressions of wind movement.

When considering environmental assets such as towels, I used the N cloth simulation asset to make the towel fall in a natural way around the ground mesh so that it would shape by itself without me having to spend a lot of time soft selecting it in a way where the mesh would not collide. When going through this process, I did find that the first few tests collided quite poorly with the floor plane, so in order to fix this, I used two poly planes to layer on top of each other to give the illusion of thickness. While I could have created a Ncloth simulation out of an extruded plane rather than layers of the two pieces of mesh, I feel this method created the effect I wanted. After copying the mesh in its final simulation stage and deleting the history regarding its simulation, I had a cloth that sat evenly on the surface.

When perceiving the animation a few times, I noticed that the sky was quite barren, especially due to the very sunny weather implications and lack of clouds. I felt this would be an appropriate time to add the plane from earlier (and used later) in the film to fly across the sky. I also feel that this use of planes gives audience members associations with holidays and beaches and therefore has a location contextual placement.

Below is the final rendered shot, which, overall, I am quite happy about the outcome. However, there are several factors I would improve upon. For example, the coconuts in the trees sometimes detach slightly from the mesh in an unrealistic way and make the weather appear slightly more severe than I had intended.

Shot 6

Since I already had the entire scene set up already, I moved straight away onto shot 6, Which followed a very similar process of animation. However, I removed the plane from the background as I felt that would be a bit too much external animation and would distract from the main content.

My critique of this animation is similar to that of the first shot. However, I feel I was able to time the ‘mouth’ with the spoken content much more successfully, and I also feel the timing of the body movements was more effective. However, the eye movements perhaps moved too quickly across the face of the building as it is a larger surface area than that of a normal eyeball. In future, I will consider the keyframe timing between each eye dart and the space it has to cross in without looking too drastic.

Shot 4

Shot 4 only primarily needs about 3 assets, a table, a cake and character 4. Due to this, I got started right away on completing the shot.

Shot 4 Modelling

When googling birthday cakes as a reference, there were a lot of very complicated designs that had candles with ages on them and lots of patterns drawn with icing. However, to keep the cake very simplistic and to keep the age of the interviewee anonymous, I settled for a very minimal design that colour-matched the house model it corresponded to.

For the lighting of this scene, I wanted to create a very dramatic bright spotlight to accentuate the loneliness of the character and the isolation he felt on his birthday.

Exploring the lighting for this shot, I also wanted to create a subtle light from the cake that would imply there are lit candles. In an ideal world, I would love to explore with fire simulations, but simply due to time constraints, I will try to explore with mesh lights and key framing the explore to create a ‘flickering’ effect.

Since the table by itself looked quite bare and plain, I decided to utilise what I had learned about N Cloth materials. I created a plane that I could wrap around the table. Once It reached its final settled position on the timeline, I created a copy of the material and cleared the history to ensure it was still not attached to the simulation. In order to keep render times down, I did not want to add additional things to crowd the scene. However, If time allows, I will go back and re-render the simulation.

Replacing the place-holder cake with the completed model, I began to create mesh lights that resembled small flames to begin to create the effect I wanted.

My initial idea was to create spherical mesh lights that would surround ‘flame-shaped’ mesh. However, all this did was create an odd effect in which the light would not project around the mesh but was instead obstructed by it. From this test I decided creating the flame model directly as the mesh, but making it visible, similarly to the process of making stars, I was able to get a better effect overall.

From this point, I key framed the intensity of the individual lights to create a flickering effect to the candles so that they would give the appearance and implication of movement.

When producing a test render, I explored different coloured rim lighting, and by adding the same orange hue to the roof rim lighting, it created an upward light effect which gives the impression he is sitting above the cake. The same process will be applied to the additional textures of the house.

During the animation process, I wanted to make sure the eye darts looked nice and clean in contrast to the movements mentioned for shot 6 earlier in this blog post.

I feel this shot is more successful with the eye darts, and the motions are not as drastic in appearance. I also feel the body timing matches that of the spoken content. However, there seems to be a slight disparity between the eyebrow movements and the rest of the animation. The eyes also attempts to be less drastic, also seem more ‘floaty’ and less concise. Going forward I will have to find a midground in which the eye movements look more reminiscent of real eye motions.

Shotlist

Going into this week, my primary aims were to start the animation process and keep a clean and structured plan about how and when each shot will be completed. This was very important to me as I did not have a large period of time to complete my entire film, and I thought it essential to keep on schedule.

Shot 1

Beginning the process of animation, since I had already established how to use X-gen to create grass and worked on some additional environmental assets, I started straight away with shot 1. Initially, I added the X-gen descriptions so that the grass would generate, and I could add some dynamic constraints that would create the ” wind-blown” effect.

While I initially tried to import the original X-Gen file I created to see if it would work referenced, there were issues as I scaled it in centimetres rather than Metres which is what the environment scenes. So following the prior method, I recreated the X-Gen grass by creating a new description, changing the taper, coding bend displacement and applying an Ai toon shader. 

A key cinematic element I want to achieve is having all of the characters (for the most part) sitting centrally on the screen so that the transitions between the differing shots are smooth. For this idea, I took some inspiration from Director Wes Anderson, who always used very symmetrically wide shots with little to no camera movement for a stylistic approach. 

When starting the animation process, I first referred to the source interview material to look at of the facial expression, such as eye darts and jaw bounces, as these are the key elements that will successfully transfer onto my rig. Due to the inherent simplicity of my Rig, the animation process, while shorter than prior works I have done, felt the most free-ing and open to exploration and interpretation. Since it was not possible, or my intention, to copy the footage as animated mimesis, I felt free to see how I best think the movement would be adapted onto an unrealistic character rig.

Using the graph editor to smooth out the initial facial blocking, I was able to create smooth eye motions that feel reminiscent of the original video. Expanding on this within Maya’s dope sheet, I was able to slightly offset the eyebrow motions to add a level of realism. As stated by Kenny Roy, Eyebrow movements are generally times with the motions of blinking (2014); however, my character rigs do not have this ability, making it quite difficult to know when to time them.

Character Animation

Since this is the official first shot of the film, I have animated the introduction and title sequence of the film, as the starting camera position is dependent on its location. When beginning this process, I wanted to create the same feel as the opening of ‘the Simpson’s as it provides an iconic establishing shot that brings attention to the subject before introducing heavier environments.

X-Gen Simulation

Due to the inclusion of simulation within this shot, the render time will likely extend into next week, especially as I foresee issues surrounding the overall render in terms of noise reduction and simulation caching.

Asset Modelling

Plane

Since one of the shots In my animations is a plane model flying overhead, I started looking into different types of planes that might fit aesthetically with the visuals of my film. Trying to accentuate this child-like “plastic ” world, I found a tutorial online I’m which I could model a plane very reminiscent of children’s toys. I would like to make the seating big enough to fit the taller, more slender character rig into it so that he can appear to be flying the plane. Given additional time, I could like to add a cloth simulation of a scarf around his next that would blow in the wind as he flies; however, this idea will have to be put on the back burner in order to prioritise animation completion. 

Critically reflecting on this model, I feel it works well aesthetically with the piece however, in order to push this to a higher level, I would rig the propellers so that I could animate them in a way that is manipulated when the global control is moved. Going forward, if time allows, I will go back and make this addition.

Car

Trying to keep on theme with this cartoon stylisation, I found an additional tutorial that highlights how to model a car that I feel would fit aesthetically very well with the film. My initial considerations for modelling a car in this world I am creating was to be square and have roofing with tiles and exhausted pipes that make reference to the different characters’ chimneys. However, since I only really have a month to complete the film, and the car is only within a shot for 4 seconds, I thought this would be detrimental in terms of production productivity. 

Below showcases the final model, with the ai toon shader applied so that it bends more effectively into the world. Since the car will be primarily used in shot 2, I thought I would colour-match the car with the red shade of the roof tiles to allow for aesthetic consistency. In spite of the car overall may seem a bit small for a world where ‘buildings’ can ‘drive’, I feel that if given the additional time, I would make one of the other characters ‘drive’ past the vehicle to make further world-buidling advancements. Similarly to the plane, If given the time, I will rig the wheels of the car to move in relation to the main controller so that it gives more realistic motion.

Scene 2 Asset Building

Since scene 2 is one of the more detailed environments within the film, I wanted to start the modelling process and get it to look right early on to leave time for the other environments. One of the initial things required for this scene is a street lamp, as, throughout the different cuts back towards this clip, the time of day changes slightly to add some more visual dynamics between shots. This entails that different lighting will be required for different shots.

Starting to compile the scene, I was able to predominantly recycle the character model to start building the ‘town’, so to speak, by changing and adapting the features slightly to give it some individualism. However, I still felt that in order to really progress the environment, I would need an additional building to add depth to the environment.

looking further into Tudor-style buildings, I feel the reference image below gave me a lot to work with in terms of selling a ‘town square’ or a communal building and provides a lot of interesting shapes to adapt and model in 3D space.

In my model, the main thing I was taking inspiration and reference from was the roof and ‘spire’ that feels reminiscent of the bell tower in a church. I think this works well within the general setting of canterbury as it is a very traditional religious English place, hosting the archbishop of Canterbury, who is the principal leader of the Church of England.

Using the same tiling technique applied to my character models, I began to populate my model with additional details, such as wooden beams and uneven tiles, to accentuate the age of the building.

Since the ‘town’ has both modern and old elements, I wanted to create a road that juxtaposes with a cobbled stoned looking street. In order to do this, I added an additional plane and extruded random squares to give a more uneven surface appearance.

Animation

After assembling the environment for shot 2, it was time to start the animation process for this character and their introductory dialogue.

As always, I initially looked at the source video material of the character’s inspiration and the spoken content. Due to Marianna’s expression with hand gestures, and very clear facial expressions, I was looking forward to adapting her speech into animated content.

The main thing I really wanted to encapsulate from this reference is the eyebrow motions and how they affect the timing and expressivity of the eye motions. I feel this works particularly well with this rig due to the use of the wooden beam in place of an eyebrow, which I rigged so that each individual end can be manipulated.

Due to the recorded audio having an interruption of a car horn, I decided to utilise this in the animation. Something I have tried to add to help enhance this animation slightly is the use of secondary animation on the weathervane to sell the distribution of weight. I think in order to improve this shot, I could work and refine the timing of the eyebrows, as they seem ever so slightly off from the eye darts and body motions. I could also go in and adjust the timing of the door opening and closing in time with the words, as there is a very slight delay is some cases.

Rendered Shot

References

• Roy, K. (2014) How to Cheat In Maya 2014. Focal Press

One of the later characters that required rigging proves to be the most complex as it is important to understand how rigging works in relation to the use of n-cloth and the overall pipeline between the two.

The initial struggle to overcome here is understanding how to combine the mesh. In order to have my previous rigs function without lag, I combined the non-animated parts of the model to create one mesh that can be easily manipulated during the skin weighting process.

https://forums.autodesk.com/t5/maya-animation-and-rigging/attaching-clothes-to-rigged-character/td-p/7517901#:~:text=You%20would%20place%20the%20clothes%20in%20default%20position,for%20it%20to%20deform%20properly%20is%20all%20%28%3A

Rig Scaling

Despite earlier mentions of the rig and scale-related issues, I still found that despite comparing the measurements between rigs early on, I still had some that were much larger than others. Recalling a technique I learned during the collaborative module, in which if the controllers and skeletal hierarchies and grouped together under a locator, the entire character rig can be scaled accordingly.

As displayed here in the video above, the rig scaling works successfully, and I will apply this to all of my models in the following week so they can be adjusted environmentally. While I understand this is not always the best industry practice, especially when there are more complex rigs being used, initially, there would be modelling guides with size parameters to follow instead. However, for a quick problem-solving solution for now and for this project, it should work fine.

This also led me to complete my final rig, which is a bonus character that will not speak or express any emotion, and the sole purpose of this character is to fill humour spaces and references and to act as a ‘random stranger’. Due to this, the rig only really needs to move up and down and move the eyes from side to side to indicate where he is looking. I did not want to rig the eyebrows and the mouth as I felt the expression was perfect as it was and completed the role I wished this character to fulfil.

As displayed in the video above, there was a slight skin weighting error in which some of the windows were being influenced by eye movement despite being flooded to the centre joint. After looking into this, I found that the upper windows were not properly skin binded to the skeleton, and after selecting everything and binding it with the skeleton again afterwards, the results worked just as expected.

Door rigging

In my earlier rigs, aspects such as the doors were not rigged correctly, therefore leading to an issue regarding the facial animation. Learning from prior rigs completed after this, I followed the same process of using a NURBS curve and changing the pivot translation to replicate that of the door hinges. However, despite this working for two of the rigs, I was having translational issues in which when I moved the hip controller, the door would move separately from the character mesh and in the opposite direction. To rectify this, like with the rigging for the eyes, I created a skeletal join that the door could be influenced too so that after parental constraints, it remained part of the character mesh uniformly.

Next Week’s Goals:

• Begin Animation
• Additional Environmental Modelling

Asset Modelling- Tree Style 1

An essential part of the modelling process will be the variety of trees that make up all the different environments and set them apart aesthetically. For the first environment, I wanted to model some pine trees with a contradicting length to the shorter character to fill up the three-dimensional space behind the character.

Finding an interesting tutorial online, I decided to follow it in order to achieve a more low-poly effect to help reduce render times later on and match aesthetically to the less detailed character models. The initial step was to create a flat plane which would act as the base of the layers of ‘leaves’.

Using the multi-cut tool, I created some rough shapes around the edges to give the impression of a group of trees that will eventually be layered similarly to that of a skirt.

Using a wrap deformer, I was able to get the mesh to form cylindrically so that it would be easier to place around the trunk of the tree and also to manipulate shape-wise.

Using the soft select tool, I adjusted the faces so that they look less uniform and more natural in build to assist in selling the environmental shot.

Repeating this process several times, I was able to build a layered tree which I think has the simplistic aesthetic alignment I need for my project but does not sacrifice the impression of detail, which is exactly what I wanted to achieve. While it is not the neatest of models, it will work perfectly well for my project and its time constraints. In future, I would like to play around with the text textures with transparent backgrounds that I layer to give the appearance of individual leaves.

X-Gen Grass Test

Since Grass is an essential part of some of my shots design, I wanted to learn how to use x-gen simulations to see if I could add this to the shots for more dynamic passive animation. I feel passive animation is a really important aspect of animation, in particular when it regards the environment, as otherwise, the shot can appear quite stale and lacking in life.

After creating a plane, forming the X-Gen description and adjusting settings such as the tapering and bend parameters, I saw that all of the strands of grass moved uniformly, which does not appear naturalistic at all. In order to fix this, on the bend parameters, there is a slight line of code that will randomise the bend direction of each blade of grass. By going into the expression editor and typing ” rand (-1,1), it adds a random disparity of direction between the values of 1 to -1 so that while all the blades are facing different directions, they do not collide with the mesh.

The effect of this looked much more natural and realistic, and I was happy to move forward to the overall textured look from here.

Since in the tutorial, they use an Ai standard hair surface to texture the planes, I thought it would be useful to test how the Ai toon shader would look applied to this simulation as I am trying to keep a consistent aesthetic within my film.

As shown above, the overall look, once the edge contour colour was changed to green, looks quite fitting however doesn’t have the strength of the black toon shader line, which could cause some visual discrepancies when rendering later on. While it does not perfectly match the time constraints I have at present to complete this 3-minute film, I feel the look will work just fine.

Placing this grass into the environment to check how this looks the character model and the trees again, I think that due to the fact it does not have a heavy outline, it actually draws your attention more to the central aspects of the screen rather than the ground. This will aid me cinematically, I feel, as it does not distract viewers from the actual animation but still creates more atmospheric density to the scene that will enhance the overall setting. I will apply the same technique to the other scenes that will require grass.

Room Modelling and Light Testing

Shot 9 and 13 have quite an extensive environment, leading onto an entire Birmingham-inspired street; I wanted to start modelling this and testing the lighting set-up early on. Using the base of my previous model. Due to the attached nature of the street in question, I wanted to try and fit all the houses together in a way that sells it as a street that is on a slight incline so that the height differences between the roofs match that of the reference imagery, and separated the houses from one another, so it does not just look like one big, long house.

The most difficult part of the process here will be correctly lighting the interior part of the building model. As learned from a tutorial from last term here at LCC, I know that for a successful lighting set-up to work, there needs to ideally be three sources of light hitting a model to make it very clear what is happening in the scene and reduce noise in the render. For this room, in particular, there will be the key lighting from the window, a ‘directional’ light that I will use in the place of a ceiling light, and a mesh light, which will emit from a lamp I intend to model that will brighten the corner.

In order to start getting effective light emission tests, I began modelling the lamp that will be a key feature in the room. For this, I just used references of standard floor lamp images while not being too extravagant to make this home seem as ordinarily British as possible.

For texture and colour inspiration, I drew reference from these very classic green and gold desk lamps I particularly like the contrast and feel the colour scheme matches the brown of the character.

The next most important part of the modelling process is the ‘sofa’ or, more accurately, an armchair for the character to sit in. When modelling this asset, I decided to use the character rig to get I direct size reference and see exactly how they would fit size-wise into it.

In order to decorate the walls of the interior room, If I had the time, I wanted to create different family photos, including this house with potentially a smaller house, giving implications of a child/grandchild or a wedding picture that sell the ‘family photograph’ aesthetic. However, due to timing constraints, I decided to use different renders from the film as that is already in the works currently and makes direct reference to the other characters that are relevant within the film.

Below showcases a more finished idea of the ‘street’ outside the window, and I feel the houses working in descending order draw your eye to the central house (with eyes) and also the car asset, which helps decorate and add realism to the scene. I used the car that drives past in shot 2 also, giving reference that potential it is this character’s car and that they reside in the same world.

Below is the first render test I did for this street. However, there is an overwhelming amount of light as there are 5 different skydomes all attached to the individual referenced assets. This also highlighted a transmission error in the window that I need to create a pan shot out of.

In order to fix this, I changed the material type from Ai toon to an Ai standard surface, as there are more options to choose from and aspects such as transmission and emission can be adjusted, which aided in achieving this see-through look so that the character in the shot does not instantly vanish and allows for continuity.

Next Week’s Goals:

• Finalise and Adjust Rigs