Month: September 2022

Character 2 Rigging

Character 2 poses to have the most interesting facial rig due to the wooden beam that functions as an eyebrow, however I feel the prior weeks rigging techniques will also work well here. The initial step was adding all the skeletal joints, and being particrualy in the eyebrow placement joints as they will make to be inline so that one is not further forward or back causing mesh collisions.

After this process I once again binded the skin using closest distance in the joint heirachry and got similar results in the skin paint distribution to the first two times I have rigged the characters.

After flooding all the correct mesh to the correct corresponding joints, I was able to achieve the result I wanted that went quite smoothly in process due to last weeks experience.

While these legs were shorter than that of character ones, they are also a lot wider which gave me a lot of leeway into the skin weight distribution and therefore I was able to get much cleaner knee bend on both legs.

The interesting part proved to be the weight painting of the upper ‘beam’ or eyebrow as there needed to be a well blended by even distribution influence on each joint that has a careful gradiaent In the centre so that it will all move uniformly but allow for additional eyebrow expression.

This i feel was particularly successful, and shall we very interesting to utilising during the animation process later on in the pipeline.

and early issue I had was that due to the fact the eyebrows were not 100% Flooded to their corresponding joints, I has some initial issues with the influence on the eye joint due to the proximity during the binding process.

Character 5 Rigging

Once again, reaptitng the same process for all the rigs prior to this one, I was able to successfully skin weight paint the model in the same process and also have even more leeway with the leg Iks due to the slight length of them.

The method of using separate joints for the eyes has been particularly successful, so I have once again uiltised this effect. SInce they eyes are more similar to that of Character 2, I also found this to be quite simply as they only really need to move on a Y Axis.

During the initial Skin bind, there were deformatives in the bottom fo the mesh, due to the high number of stones and mesh collision in that area. This was quite a simple fix by adding the influence, to the central hip joint.

When rigging the door, I thought of technicalities surrounding the pivot point, and alinged the nurbs controller with this pivot point so once it was parented it would follow the same rotational axis and work in the same way. This controller would then be added heriarchally to the hip joint so that it will follow the rest of the rig without moving separately from the rest of the mesh.

Character 6 Rigging

I followed exactly the same process with the rig for character 6, however, instead of rigging one door, I rigged two. The legs on this one proved to be a lot more difficult to weight paint, but applying the same time and concepts to the rig I was able to create an effect that works for exactly what I need it to do going forward.

Next Weeks Goals:

• Complete all character rigging
• Start Environment Modelling

Character 3 Rig

I wanted to initially rig character three as he is the only one with a planned walk cycle, and I wanted to allow time to properly weight paint the legs so they do not deform in unintended ways.

Starting the rigging process of these house models, I immediately checked that all the pivot points for the doors were correct and in functioning order, as well as the translate and rotate limitations.

The initial sept is placing the ‘skeleton’, which has just two IK legs and a central spine. One of the most important aspects of skeletal placement I have learned during my time at LCC is not to ruin the rotations by placing the joins in a messy, disorganised way, and specifically rotating each joint hierarchically and neatly so they are on a correct axis.

From here, I used a joint hierarchy bind to attach the mesh to the skeleton, which also selected the closest in the hierarchy so that the mesh influence would attach itself to the nearest joint. Since The house will not require very complex and naturalistic skin weighting, this option should suit the job quite well.

The initial binding, as displayed below, had several issues in terms of mesh deformities. Due to the default influence of the mesh, all the vertices are corresponding to several joints at once, which works well for a more humanoid and natural rig but will not work in this instance.

There also prove to be many issues with the IK application in the leg, as the knees seem to deform and shrink and the upper leg when manipulated. While the overall effect of the character intends to have an almost ‘rubber hose’ style, there still needs to be no deformities.

Looking into ‘rubber hose’ rigging techniques on youtube to assess whether this is the best course of action going forward, I found that giving the legs that much individual manipulation, it may complicate the animation in a way that is not necessary for this film. However, the tutorial may be very useful later on for a different project.

Another technique I was looking at the exemplify squash and stretch was to make a ‘stretchy’ spine so that the houses could bend and pull in a way that, again, would not massively deform the mesh. I feel this would be useful however, stylistically, I do not think it will essentially work in selling that the characters are stiff like houses.

The most effective way to bind the skin, similar to my skeleton project, the ‘Danse Macabre’ Dance-off, was to use 100% influence of certain mesh to certain joints so that they did not deform upon movement.

As shown, the current mesh, as it stands, warps and deforms massively due to multiple joint influences.

By going through all the different parts of the mesh and giving them 100% Influence over the central joint that will be the main hip control, the mesh was able to successfully follow in a way that makes sense.

Despite this, there are still several issues regarding its movement. One of the predominant issues that stand out is the leg deformities mentioned earlier.

due to the nature of the character, I would not be able to get the leg and the knee to be exactly in line with the joints, as there would be no space to bend. However, I successfully managed to get t the knee and upper leg to work in a way that minimises drastic deformities.

In order to 0 all the translations on the leg to avoid the rig going ‘crazy’ I created individual hierarchical groups for each joint and froze all of the transformations.

Interestingly, one of the issues I had was the mesh collapsed in on itself after trying to set influence to the main body of mesh, so due to this file corruption, I had to close and re-open my prior saved file.

I found that in order to get individual eye control, I had to make different joints for each individual eye and eyebrow and set them all the 100% to the corresponding joint in order to get them to move independently.

Hat constrain

Since I wish to add secondary animation with the hat, I decided to research different hat constraints that would allow it to move with the main mesh but work independently from it.

In doing the hat, I made it individually controlled but parented it to the highest joint in the ‘spine’ so that it was able to sit in its original position and could move independently from the body. Could then turn on and off the influence or ‘attachment’ in the transform settings.

Character 1 Rig

Going back to character one, I essentially used the same process in order to get it functioning, as both characters 1 and 3 have very similar features and functioning eyes.

The main struggle with this character was getting the weight paints right due to the very short legs, as to was very difficult to avoid massive deformities due to the lack fo space to work with.

After playing around with this tool for a while, I was eventually able to create an effect where the information was much less noticeable upon bending and settled for its appearance of it. This is an issue I expect to reoccur since most of my characters have shorter legs.

Next Week’s Goals:

• Continue Character Rigging

Since I have completed the entire character modelling process for my short film, I have started thinking about poster concepts in which all the characters can be placed together to best advertise the film. This also gave me a good indication that I will need to rescale some models later on as some proved to be much bigger and much smaller than their counterparts, and some of them are incorporated in the same scene. This is important to do before the rigging process, as scaling characters after the skin-binding process can effectively mess up the entire rig.

Music Tests

While a majority of the film is based around multiple dialogues conversations, there is an introductory section which I feel needs some musical addition to help set the tonality and theme. Exploring this sound, due to the child-like, stylised intended aesthetic of the film, I decided to play around with sounds that are associated with children’s television shows. I wanted to emphasise this by using instruments such as a glockenspiel and piano, which add a light and simplistic feel to the melody. This ‘happy’ introductory theme and the toy-like children’s aesthetic feels purposely contradictory to the spoken content in a way I feel works as an animated short.

Previsulatiation

As stated in prior blog posts, previsualization is a key and essential step in the filmmaking process to place everything into the correct spacing and see if it works. Since I am behind In the environmental modelling section of my film at the moment, this process provided me with a perfect base to start this process scale-wise and gave me an indication of object placement and overall setting depth. This also made several things very clear to me, as a few ideas did not translate well from the 2D animatic to the 3D Previs. One of these is the idea of having a humanoid character have a silhouette in the light of the window. While this conceptually works quite well and is more visually interesting for viewers, I predict I will not be able to include it going forward as I feel adding a human into the setting while also having a house within a house throws off the world balance. I also think this effect will prove to be very difficult without modelling a character that would make the right kind of silhouette to match the overall aesthetic.

During this process, I also found that using the shutters to make character 5 blink actually appeared slightly strange and uncanny due to its sideways motion of it.

The previs also indicated some issues that I may have regarding the Ncloth attached to character 3, and it may not follow the rig in the way I am expecting and key massively between keyframes. While this is something I would want to research in a more extensive way, I will add it as an additional feature If I have time in order to stay on schedule.

Next week’s Goals:

• Character rigging (priority)
• Environmental modelling

The next character model is a very proud Italian friend of mine who got very unfortunately caught up in the early issues in northern Italy during the pandemic. In order to pay effective tribute to her location and nationality during this time, we both agreed we should look into classic Italian architecture for this model.

Character Model 5

While she is from Valley d’Aosta and resides within the Italian side of the alps, that area has a very distinctive type of architecture due to the cross-cultural impact of France and Switzerland; a lot of the classic architecture can be associated with any of these areas. In order to communicate with an audience very clearly the Italian setting, I began looking at Tuscan houses, which are very distinct in style. These also potentially match the most stereotypical international views other than Rome and Venice.

Beginning the concept art, I really wanted to ensure that the shutters and the roof tiling were in the most visually-referential way I could with the imagery I had selected. I also felt incorporating the shutters in the ‘eye’ windows could effectively work as eyelids later on in the animation process.

First and foremost, I applied techniques I have learned from previous weeks’ modelling to create the roof tiling in a neat and orderly fashion. In contrast to my first model, I feel this friend of mine is a very neat and particular person, and I wanted to portray this visually in the best way I could.

Trying to get this flat, serious expression within this character model was also an essential point for me as a lot of the spoken content from this character is very serious and self-reflecting, and I wanted her expression, even within the limitations of architectural structure, to look intentionally flat and stone.

In order to rig the shutters later on, I wanted to endure the pivot transforms were correct and also had rotational limits on them so that during the animation process, they could be prevented from mesh collision, which I also something I intend to add to every door on the models I will begin to rig.

Experimenting with brick placement, I realised that due to the type of texture these buildings are traditionally made of, this method would prove ineffective, as it would not reflect the style of the building I was aiming to go for. In a lot of the Tuscan houses I had looked at, there was more stonework than brick.

Starting the base layer of stone, I wanted to see if I could make the bricks appear like a stone on the rest of the mesh; however, personally, I feel this made it look messy and imbalanced, which went against my initial goals for characterisation.

After managing to greater three neat rows of stone all the way around the model and using the same guttering technique used on model 3, I feel I was able to create a (somewhat) authentic italian-style house with a clean and simple appearance. I feel that minimally trying to add subtle location-specific details to my models has proved a slight challenge. However, I think it works quite well here, and I hope to continue to adapt this further with the next model.

Character Model 6

Since this character is based on a man who lives/ and has lived/ in Birmingham his whole life, I wanted to really try to get the essential parts of a classic Birmingham home into my modelling. A very classic trait of housing in Birmingham is this crowded attached terrace style housing which has this distinct’ red brick’ look to it. I also particularly like the layered circular detailing on the roofs, which will interesting to incorporate into my model.

Translating this into a character design, I think trying to play on the attached housing will work really well in a movement design aspect, as I can utilise both doors during different speeches to make a more interesting and dynamic piece of animation. Another element that I think will really help in creating expression is using the unique circular brickwork above the windows to use as eyebrows.

Considering this character is the only one who will be ‘indoors’ so to speak, I also wanted to consider the modelling for the environment early on and consider what the ‘house’ that they will be situated in will look like, as I intend to use this character model as a base.

Beginning the modelling process, I felt it appropriate once again everything very neat as this house will be very symmetrical in its visual approach. The tiling for this house also needs to have smaller, more close-knit lines due to the nature of its source material.

After completing all the individual lines of tiles around the whole building, I was very happy with the result and decided to use a ‘boolean’ technique to make the detailing impression for the edges of the roof. This created a nice effect that I was happy with aesthetically in wireframe.

However, when rendering this effect, it did not show up in a way that appeared there dimensional and lost a lot of its detailing and effect overall, looking like two flat planes on top of each other.

Trying to get this idea to correspond correctly within the Ai Toon shader, I created three layers of this curvy line of mesh to create a similar impression. By moving these slightly off-centre from one another, I was able to create a clean layered effect which appears to have a ‘softer’ and more friendly-looking appearance as it feels reminiscent of tablecloths typically associated with grandmothers. Having had northern grandparents myself, this is something I feel I would have typically seen in their house growing up and it helps set the age of the interviewee and their English background.

This effect also worked much more successfully in the Arnold render view, and I actually feel it achieves the aesthetic preference in a cleaner and more visually simplistic way.

Trying to get this central brickwork detailing and also incorporate it as the eyebrow design proved to be quite difficult as the movement would either have to be restricted (therefore reducing eye movement massively) or would have to move as an entire line which would look incorrect, or far too jarring.

Cropping this down to just the circular areas above the windows seems to work and also really helps add an expression that makes the character appear more friendly without losing its Birmingham characteristic.

For this model, I also wanted to add the appropriate door handling, which gives it a little more detail when rendered with the toon shader that separates it from the other door designs.

Below is the finished house model, which I feel encapsulates its location very well and indicates something quintessential from a northern British attached house. I feel quite enthusiastic to animate this character later on as it provides a lot of opportunity for expression with more carefully designed eyebrows and doors. I may potentially have to turn down the specualar on some textures, such as the tiles later on, as they make the model look very clean or wet, which is not the overall intention. Now that all of the main character models are complete, I now aim to start the rigging process with all of the characters and figure out exactly how they will move.

Next Week’s Goals:

• Finish previsualtaion
• Render tests (If time)
• Start Character Rigging (If Time )

The 4th interviewee clearly stated how they were in Florida, and with contextual input of them working in a cafe, I decided to research beach cafes that can be typically seen in that area. The main theme I noticed is a lot of primary blues and light-toned buildings with ‘flatter’ beaches and large umbrellas.

Architecturally, I felt very inspired by the image below as it has a lot of interesting shapes and features, which I believe would translate well into 3D space, such as the railing and simplistic beach cafe sign. I also really liked the use of the tarpaulin roofing that abuts the building and window ledges, which could prove interesting later on in terms of simulation and also texture exploration.

Expanding on from this, I drew a design that accentuates this more square shape and adapts it to fit an anthropomorphized character that requires a ‘mouth’ to speak. Trying to thematically engage with the weather of Florida locations, I wanted to accentuate this by using sunglasses. I also feel that later on in the rendering process, I could add some interesting reflections to them. The key thing I wanted to achieve from this design is how the shape will differ in a silhouette sense from the other characterised buildings and also to create a different dynamic door idea that fits into a more modern aesthetic due to its context.

Beginning with the base model, I felt that the addition of the railing was particularly effective here, and it gives the impression of additional ‘layers’ or floors within the building that give it a different scale and also allow me to play around with things that can go on the rooftop. The inward extrusion of the base, it also adds some surface detailing that is not present in the earlier models that I feel will look good in the final render.

When exploring what would be the most effective way to create the tarpaulin roof, I decided to try and use Ncloth to create a more believable cloth sitting that could be manipulated by wind effects due to its setting.

The first process of this was creating the building as a passive collider and then attaching the end vertex of the mesh using a transform constraint to make them fall over the roughly modelled window.

I feel this made the cloth sit more naturalistically and gave me a good starting approach for how this will later be applied when modelled into the correct shape and length.

The next key step was modelling the facial features, which will be the most essential for the animation process. Since my drawing tries to effeminate this character, I added some ‘eyeliner’ to the mesh to give it a bit more personality and differentially it slightly from the other models without string too far from the general aesthetic.

When beginning to model the sunglasses, I was having a lot of issues trying to start the process, originally using too large transparent squares to try and create the effect. I realised later on that it was proving difficult to try and create the circular effect without reference, and I found a pair online in which I could begin to shape the base around.

Using the reference, I found it much easier to get the right shape and focus on the details, such as the concave shape of the glasses, as this will help with helping distort reflections naturally and also work better generally in three dimensions.

After modelling the glasses, I really wanted to focus on a more stylised form of highlight so that it could match the heavy ‘toonshader’ linework of the renderer. In the concept work, I have the white ‘squares’ that implicate highlights. Trying to translate that into 3d, as visible in the rendered version, the contrast of the specular glass and the stylised highlight do not work well together. And turning down the specular also looks odd because it would block the visibility of the eyes, therefore restricting expression.

The most important aspect of the model in order for successful rigging was the door. I wanted to make an automatic door to add variety and also play around with the differing logistics of door styles and how they would work as mouths, so to speak. Since the main aspect of this motion is both of the doors moving swiftly and cleanly to the side without any mesh collisions, it felt important to make sure I get the transforms correct early on so the rigging process goes as smoothly as possible.

For the font for the front sign, I wanted to play with audience associations again to build a stronger sense of the environment and tried to replicate the font of the famous ‘sandals’ beach resort that was popular in the 80s and could potentially resonate subconsciously with older audiences. This also creates a more ‘beachy’ aesthetic for the whole model.

Now that a good portion of the character modelling was complete, I wanted to start the simulation process again with more accurately modelled cloth and see how this would work with the existing model. As seen in the clip below, the first attempt proved more successful than the prior attempt. However, the cloth is still too long and too thin to effectively replicate the cloth of tarpaulin/ cotton that is typically used for gazebo-type covers.

After shaping the cloth more effectively and using the transform constraints and more points to ensure mesh connection, the cloth sat much more accurately than before and achieved the overall effect I wanted going forward. However, I still felt the cloth deforms could be pushed even more accurately with additional underlying mesh.

Adding the metal bars that would potentially be used under the cloth to help hold and shape them over windows and cafe porches, I created a cylindrical bar and added a passive collider so that the NCloth simulation would sit on top of it.

The final result with the added bar collision, which I think adds a slight bounce and a cloth deform in the centre, would be more accurate to real life.

Applying a similar technique to the umbrella, I initially tried to create the edges of the mesh to have a Ncloth effect. However, this did not work as the Ncloth simulation appears to need to affect an entire piece of mesh to work as effective simulation.

Since the umbrella stick would not, from prior simulation experience, effectively hold the entire weight and length of the Ncloth mesh, I created an ‘invisible’ smaller piece of mesh underneath that it could collide with in a way that could make it appear realistic.

This idea application, with added collision weight and thickness, created exactly the kind of effect I wanted for this simulation. This could also be extremely useful when adding an additional ‘wind’ simulation as it could help just add that slight, subtle animation that helps sell a shot environmentally.

When finishing up the model, I wanted to add some additional detailing that would make the cafe look more full and life-like, such as cactus pots. I also added a ‘sun’ logo to the sign that feels reminiscent of the earlier reference image but acts as its own design. Given more time constraints on this, I would make the sign work as a ‘neon’ light, as beachside cafes/bars generally make use of it due to how dark beaches get at night, and I could play around with the time of day between the differing shots also.

Due to the nature of the type of building it is, I also wanted to add restaurant ‘seating’. Initially, I had thoughts of making the chairs big enough for ‘building’ characters to fit into as that fits the overall theme of the whole world I am building. However, I settled for making ‘human-sized seating as I felt this was a perfect flourish for the model detailing and also implies the existence of smaller living beings in the world without making a direct implication it is human.

Below showcases the final model details, with the textured Ncloth materials, which colour match the building base and also will, later on, match the towels I plan to place on the beach to help fill the environmental scene and make it more believable. Given the additional time, I would add a lot more careful details to the building, such as brick detailing, guttering, tables and little assets such as cups and mugs etc.

Next Week’s Goals:

• Design and Model Character 5
• Design and Model Character 6