I composed sheets that had the fully rendered version of the weapon, separating the sword and the shotgun, along with a turnaround of the Tidal Captain himself. I wanted the lighting to be very clear so I picked a white light to make sure everything in the image was clear.
After placing all the objects together in Blender, I exported everything into Marmoset Toolbag 4 for rendering.
Working on Marmoset is fairly straight forward, so it was generally a smooth process. However, due to my PC’s specifications, it started to get slower as I advanced, so I had to resort to viewing the model in wireframe in order to make sure my process was more smooth. This did impede on the speed of my process but it did make it more possible.
I gathered several references for the lighting that I wanted to achieve. I wanted to make sure the lighting was suitable for the themes and environment the character was in. Since my genre was dark fantasy, gothic, and mostly taking place on the sea, I wanted the lighting to be very dark and cool, with lighting that looked as if it was coming from the moon. I also wanted to use the emissives and glows from the lamp and the sword to give it a more interesting and eerie composition and with this in mind, I made a couple of initial renders.
I added a solid shape within the lantern and gave it an emissive effect as I noticed it looked quite empty without it. After observing these renders and having some feedback, I decided to make the lighting more in theme, and looked into the present backgrounds/skies provided by Marmoset. Along with this, I managed to find a suitable solution to make the lantern look the way I wanted, as the solid block within it was not to my satisfaction. I did not like how it looked, so I used a preset glass material I found within the library, and the results were much more satisfactory.
The process for rendering turned out to be very tedious due to my own computer specs. I found that I had to go back and forth between the wireframe and highest quality version of the viewport as my own system would run very slowly on the highest quality setting. Eventually, I managed to get a lighting setup that I was satisfied with for my project, with many processes of trial and error.
Eventually, I settled for this lighting setup:
As I thought it was the clearest, as well as having the same atmosphere that I wanted to achieve. I also experimented with using fog in the scene, as I like how the lighting from the lantern looked, as well as making it look like he is walking somewhere quite misty or haunting, which was fitting for the scene.
After applying the various maps on the visor, I could see that it looked up to my satisfaction, I advanced into making the weapons for my character. I deemed this a crucial element that was necessary step in regards to the whole project as it effectively showed my character as someone who is capable of combat, as well as portraying the narrative as combat-heavy, and the genre of the media if it were a game as I intended.
Using the various methods and techniques, I began to model the weapons out on ZBrush. I chose to model the weapons on ZBrush first so that I could make the higher poly details with more ease. I made sure not to make the same mistakes when approaching making the weapons, such as making sure there are no errors with the mesh. When making the sword and the gun, I wanted to make sure there was a good level of accuracy to them, and so, as I modelled the sword after rapiers and thinner swords, I made sure to compile a mood-board with a handful of references to aid my overall accuracy when making the sword.
When it came to making the gun, I wanted to make sure the guns were period accurate, as my some of inspirations for the aesthetics were around the 1700s and 1800s and less modern eras, and thus, I wanted to stick with flintlock-operated firearms. Like for the rapier, I compiled a selection of references, including ones that I had taken for primary research in the Royal Armouries, as I noticed that flintlocks have very distinct shapes and intricacies and I wanted to replicate that in my design. I also looked to my major inspiration, Bloodborne as they have gone a similar route in terms of their firearms. I used my previous concept as a foundation and built off of it, figuring out how the sword would lock onto the top of the gun and making sure it would look functional. I figured this would be easier to solve in a 3D setting and I proceeded to model.
I repeated the patterns and motifs that I sculpted onto the armour on certain parts of the shotgun, as I wanted the design to be consistent. When it came to the actually flintlock mechanism, I wanted it to make sense, so I made sure to look at the references and diagrams for how the contraption operated, as well as looking into a video that highlighted how it operated in the YouTube video ‘How Flintlock Firearms Work’ by MouseGunner.
I repeated the same patterns and designs for the rapier, but keeping the details more minor on it as I did not want to overcrowd the design and end up making it look too noisy. Due to my developed skillset, I was able to get both the sword and gun done fairly quickly. I made sure to already size my models appropriately so that they would match the high polygon version, making the baking more smooth. I reused the emblem I had previously made for the cape, and put it on certain sections of the weapons with height information so that it would make everything look more cohesive
I made sure to add various layers of dirt and grime to convey that these weapons have been through many battles- however, for the sword, I put slightly less blemishes on it, as narrative-wise, it holds something akin to holy power, and I wanted to convey this. I also added an emissive light to the sword to make it have a dull glow in the final render as it would further convey this idea of a holy blade. For the shotgun, I also decided to add written engravings on the side, on one side having the text: “Lucet Omnibus” meaning “the Sun/Light shines for all” and “Ad Solem” meaning “towards the Sun”, which is a derivative from “Arduus Ad Solem” found in Virgil’s Aenid.
When finishing the textures, I moved onto rigging my model. Though I could have manually rigged my model, I knew that there were many industry-standard and widely used programs that handle auto rigging efficiently. I was advised that I could use Character Creator 4, which automatically rigs models using only pinpoints. Due to the efficiency and ease of this program, I chose it to proceed with the next step of my project. I took out my character’s capelet during the rigging process, as I wanted to alter it after I had finished posing, as I wanted the capelet to look more dynamic in the pose, which I would be able to sculpt and use cloth simulation for. Using a brief guide by Reallusion under the title ‘Auto Rig Tool for Creating Skeleton Characters: Advanced AccuRIG | Character Creator 4 Tutorial’, that I had found online regarding setting up the character rig, I managed to proceed with ease. I felt that this process saved me a lot of time and provided me with a very efficient and suitable rig for my character, but I saw that there were several errors that I could have prevented such as how the arms bent due to how I initially modelled him.
As I saw there were some slight defects with the rig, I imported the model and armature back into Blender and adjusted the weight painting. Looking back, I should have made the mesh and topology of the shirt more accustomed for rigging, as I saw many polygons stretched out, which in turn would make the textures look warped. I had to slightly manipulate the pose in order to make this less obvious, given that I had already made the textures beforehand. After adjusting the weight painting and getting the overall pose of my character in position, I looked at multiple other references deriving from the pose to position everything in a correct and realistic fashion. My main goal with the pose was to give the impression that he is stepping into or out of combat, and had the weapons in mind. As I knew I wanted to make a hybrid weapon, inspired by Bloodborne’s trick weapons, I made his hands go into respective positions, looking at references to see how each weapon was held to ensure accuracy. After this I imported the capelet to ZBrush, using the tools and cloth simulation to make the cape flow how I wanted it to. I was initially stuck on how the logistics of the cape within the character rig, as I knew it would cause clipping issues if I were to pose my character. After some feedback, I was told to already pose my character, so I could simply sculpt the direction and position of the cape, therefore meaning that I would not have to worry about any potential clipping issues in my project.
After this I imported the capelet to Zbrush, using the tools and cloth simulation to make the cape flow how I wanted it to. I didn’t want the pose to feel flat, so I gave the capelet a more dynamic position, keeping in mind how the larger cape was positioned. I noticed some errors in the mesh due to flipped normals, but could not find a way to fix them. Due to the erroneous section being one that was harder to see, I decided to leave it, but made sure that I would prevent errors like this moving forward. I found that I had to adjusts the cape’s texture slightly when re-importing it, which only required minor fixes. I baked the folds and high poly details into the cape, which brought out the folds more. Though, I realised I could have been slightly more bold with the folds and creases, as they ended up being quite subtle on the cape.
For this project, I used Marmoset Toolbag 4 for my renders. As I have not worked with such high quality textures and armour before, I tested how the materials looked when applied to certain subtools. I also wanted to see if my system was suitable for running my render. After applying the various maps on the visor, I could see that it looked up to my satisfaction, I advanced into making the weapons for my character.
I only decided to create high polygon versions of objects that needed specific patterns or clothing that needed obvious folds such as the shirt and trousers as this would be hard to detail in Adobe Substance Painter. I decided not to sculpt in noise or other minor scratches on the other subtools that I would not end up baking because Substance Painter is capable of creating such an effect through height information and various other texture overlays which are provided in preset materials. When moving onto the baking stage, I had imported a high polygon version of each subtool that I made a high quality version of. I overlooked the sizing my project, and saw that the high polygon versions were not the same dimensions as my mesh, so I manually fit each one in Blender. Moving forward, I would make sure my meshes are the exact same dimensions by not manipulating the height when importing to Blender. After resizing the components, I moved back to Adobe Substance Painter and successfully baked all the needed meshes. There were slight artefacts that I did not know how to fix, but I decided that they were minor enough that they would not interfere with the fluidity of the whole project. However, looking back, I should have learned how to fix this issue, not only to enhance my skillset, but to make sure that I would also know how to fix future projects such an incident happens again.
Since my system found it difficult to run constant 4K textures, I moved to 2k textures which had little to no difference when it came to appearance. The base metal materials from Adobe Substance Painter were not to my standard when it came to how I wanted my model to look. There was far too much shine and polish to them considering the nature of my character’s origins and occupation. Though I could have fine tuned the base textures and gotten a satisfactory product, I thought it would be more efficient to search for the materials online or find presents that suited the aesthetic I needed. I then managed to get access to suitable materials, such as various medieval inspired metals and cloth types. This ended up saving a lot of time and being very efficient as I could reuse all the materials in the weapons and future projects.
For the capelet of my character, I decided not to give it folds via sculpting before beginning to texture as I thought it would make it look more accurate when I put cloth simulation later. Looking back, this was not a good decision, as cloth simulation and further alterations to the cape will stretch polygons and alter the way the texture looks. I made sure to reuse the same colours and repeat materials throughout to make sure the model looked consistent throughout. Due to Adobe Substance Painter’s easy project reconfiguration, I was able to easily alter my model if I saw something wrong with the UV or was missing an object. For certain faulty UVs that I did not manage to tweak during the texture test, which I saw through stretches in quality, such as on the bottom of the greaves, I was able to easily alter and configure this and end up with a better result. The flow of Substance Painter’s design allowed me to go back and forth between programs with ease. Before I advanced further with the texturing, I decided to pursue making the hood. As I was unfamiliar with Marvelous Designer, and was informed that it would be useful to learn the program for my overall skillset and is an industry-used tool, I decided to use it to create the hood of my character.
I initially had a lot of trouble with Marvelous Designer due to my lack of experience but I eventually found that the interface and mechanics were very intuitive and realistic to the point where it was quite easy to adapt to how it worked. Eventually, I managed to export a hood that I liked.
I went back to Adobe Substance Painter 3D and applied textures to all the materials.
I changed the colour of the armour’s metal to fit more with the concept art I made, along with the colours that I had conceptualised for Primeval Ilythium, a metal-like material present in my narrative that is made of a primordial ocean deity’s body, making it appear that mild teal colour. The colour aligned with the oceanic theme, as well as made a unique appearance to the armour. I added the lighter trims by using layers and black masks, as this was present in the concept art. It translated well into 3D as I could see how much the contrast brought out the intricate patterns that I baked on and therefore was successful. By doing this, I achieved that tide-like effect that I had intended when I created the original design.
When it came to the gold trims on the hat and cape, I used alphas that I was also given access to, and utilised them in order to make a consistent pattern along the edge of the respective subtool. To give them height, I simply edited the material so that height was enabled, along with adjusting the height to the appropriate setting so that it achieved a more seamless and natural trimmed look. I found that this really elevated the gold which I did want to be clear as it was important to show his past status and glory.
It was more efficient to use alphas in order to get the desired ripped effect on certain clothed parts of my character, being the reason I made them a single plain, such as the hood and the cloth around the waist. Using the alpha blending shader mode in Adobe Substance Painter, and after looking up various videos, I learned how to make the fabric appear ripped and dishevelled- as I wanted to sell the fact that this character has been through combat. To pull off this effect, I followed a guide by Abe Leal 3D in the video ‘Painting Opacity Maps in Substance Painter’. I looked to multiple references of torn fabric in order to achieve the desired effect, and this turned out to be very useful, and made the project look more convincing.
When approaching the larger cape at the back of my character, I made sure it had two sides and was not a flat plane, as I did not want both sides of the cape to be similar. Though these cause complications, I compromised by checking if the appearance was consistent. I could not look for many references of burnt capes, so used the Soul of Cinder from Dark Souls III as a reference for what I wanted to achieve. As I couldn’t find burnt capes, I resorted to pictures of burnt cloth, which ended up helping me with the visuals of the effect. I also utilised emissives in order to get the appearance of glowing embers on the end of his cape, which ended up enhancing the design much more. I made an emblem for my character that conveyed his past affiliation, mixing a spear with a sun in a coherent composition. After some feedback and some testing on the character, I saw that the pattern was not clear enough, so I thickened some main aspects in order for it to stand out. I found that giving it height made it look more embroidered and regal, and with the tattered, burnt appearance, it gave the impression of a fallen, age old hero, which I was aiming for. I was also intending for this symbol to be the emblem for the soldiers who worked for the Kingdom of Sol.
I moved on to giving him the final assets in Blender, which, in hindsight, I should have added from the very start so that I would not need to model more. I did not intend for my character to have a face, as I wanted him to exude an elusive energy that would translate into his personality. To give this effect, I simply added a flat plane behind his mask that I would later intend to make black. I then added the planes for the glass on the lantern, and the feather that adorns his hat. I configured the project, easily adding the new meshes in. I found that I had to make new materials for them, which I quickly did before proceeding with finishing off my textures. To make the feather seem more full, I added an additional layer with more height to not make it look completely flat. Another thing I could have done, looking back, is using more planes to add those layers and make it look less flat.
For the lantern, I simply looked to Bloodborne’s lanterns and noticed that there was this fogged appearance to them that I wished to replicate. Along with this, I made sure to give the boots and greaves a more muddy, dirty look, as this character’s line of work and environment is mostly on wet grounds such as the coast and bodies of water, so I wanted to give the impression that the character is constantly walking within these environments.
Overall the texturing brought me many new insights, and I was able to learn many new skills as well as enhancing them, as I wanted to slowly gain more approved industry practices and integrate them into my own.
As I wanted the armour to have more intricate detail, I watched various techniques that would be used for making the ornate patterns on the plating. I found that using masks, inverting and using deformation to inflate it was a very effective and precise way to add those details as instructed by Seth Thompson in the YouTube video, ‘ZBrush – Sculpting Ornamental Designs’ . With this technique, I finished adding texture to the visor and the chestplate with these techniques. When composing the pattern, I wanted to blend the narrative themes into it. I wanted a way to convey the ocean and tides into the design so I drew some curved patterns that were reminiscent of the idea, along with this I added a sun motif into the patterns, overall becoming a harmony of the ocean and sun, making it another method to show my character’s affiliations. Though, I found that this method produced slightly flattened results and resorted to using a scribe brush instead. When using the subtract feature of the brush, it created an ideal protrusion for the patterns and I repeated the same tidal motif on the gauntlets, the armour on the legs and the armour on the hip. To create more detail, and the impression that this character is combat oriented, I added several dents, small and larger scratches and imperfections on the armour like chipped edges. I made sure the designs flowed together to create a coherent look, and in the end, I was satisfied with how it was composed.
As I have UV unwrapped previous projects before, the UV unwrapping process for this project was fairly simple. Though, I seemed to overlook the thickness of some of the subtools of my model. Namely most of the cloth such as the capelet and the shirt- due to the thickness, it was easy to make mistakes when selecting vertices to create into a seam. Looking back, I should have kept the capelet and the shirt as as single plain thickness. I also observe that the face orientation was convoluted due to the thickness of these materials, and looking back, I should have adjusted them before I started texturing. I also feel that I have much to learn in regards to UV maps and unwrapping them in the most efficient method.
I also should have thought of more efficient ways to minimise the amount of subtools that I had. Despite the lack of complexity in my model, there were still over 85 subtools within the project, and moving forward I would have to look at how to reduce that.
After unwrapping every object , I moved onto testing how the textures would look on my model to try and weed out any errors.
I imported my model into Adobe Substance Painter, as I was most familiar with this program when it came to texturing my model. Using textures from the program itself, I applied them onto my model for the sole purpose of seeing if the UVs were to my liking and to search for defects in them. However, most likely due to the quality of the textures and the cleanliness of them, I do not think I was able to see most of the errors in my UVs, as I was testing with only 1k resolution. After this test, I was satisfied enough for this model to go through higher quality texturing.
Before I could fully move into exporting the model into Blender for unwrapping and finishing the design, I sought verification that the design was cohesive and all the elements worked to each other’s favour. After receiving feedback about the helmet being too crowded, and should have more function and be more seamless with the hat, I adjusted the design so that the side plating of the helmet had a hinge that would allow the visor to be sat atop the helmet or in front of the main part of the helmet easily, as I thought about the efficiency of such a thing in battle. I also looked to several types of helmet models in order to gauge what would make sense and found that the design I ended up with made logical sense compared to my previous design which may have been too crowded. I wanted to add intricacies into the design, like various patterns, but I decided that would be more suited to do directly on the 3D model, so that the 2D art would be completely accurate to the 3D model (as there could be some details I could struggle to translate from 2D to 3D). After some final adjustments to the 2D design, I was advised that it would be better to figure out the logic and mechanics behind the structures in the 3D process.
With the advice given to me, I adjusted the render of my concept art.
Along with this, I adjusted the helmet within ZBrush and thought about how the helmet would latch onto the tricorn, so rather than a three layered design, I settled with a two layered one.
I then finalised the model by adding the bags on his leg.
I then imported the version of my model with the lowest subdivision levels into Blender. The polygon count was to my satisfaction, ending up at slightly over 200,000. Though this was still a lot, it was a vast improvement considering my skillset. I decided not to make the larger cape, the feather and the cloth around his waist on my model as of yet, along with the glass panels on the lantern, as I thought it would be easier to make in Blender using poly-modelling, as well as flat planes with no thickness being easier to manipulate on Blender. I began to make the planes on Blender, but realised that the cloth simulation in Zbrush could be utilised well for the cloth around his waist. I found that this got the desired effect for the way the cloth fell. I did the same for the cape, this time, manually creating the folds with several of the cloth brushes that Zbrush provided. I did this on Zbrush as Blender does not have not have brushes with such a specific utility, and I felt moving to Zbrush benefitted me more.
When approaching the design I used the masking tool in ZBrush in order to get accurate and clean shapes for my character’s armour. Using the deformation function ‘polish’, I managed to clean the topology of the chestplate, along with using the ZRemesher tool in order to decrease the polygon count into something more optimised and functional for the mesh, which I had not done for previous 3D modelling projects and found that my polygon count was too high for what my models were. Utilising this mistake as a stepping stone, I made sure to integrate a habit into my workflow that would make sure that my polygon count was as optimised as could be. I also learned through several tutorials that mesh topology could be further optimised by using the ZRemodeler brush, which I used to correct, deduce and add polygons that were needed.
It was my first experience doing hard surface modelling on a character, so I decided to look up many tutorials regarding the practice in order to gain habits that were necessary for efficient modelling. I mainly used the ‘How to Create Clean Hardsurface in ZBrush’ guide by Abe Leal 3D, which ended up teaching me many basic and necessary practices for my project. Rather than using the extrude tool to give thickness to any of the objects I used the Subdivision levels tool, where thickness can easily be adjusted and adapted to the shape of the object . Using the subdivide tool, I gave every object the appropriate level of thickness, thinking about the weight and necessary amount of protection. I used various polygon modelling techniques like insetting faces, extrusion, bevelling and inserting faces in order to give the desired subtools borders like armour trims. Looking back on my design, I decided to use the trims/borders I was making in order to achieve the extruding patterns that I conceptualised and that I would add the more curved areas in later when I do the high-poly details so that I could bake them on later as I have not baked much into my previous projects, and thus decided that I would incorporate that progress into my workflow. Constantly looking back to my references, I realised that the plating on the hip should have a clear connecting point with the main chestplate. So, as is present in many old models of armour, I added buckles- suggesting protection as well as full mobility in the suit.
After coming back to the project, I found that my chestplate subtool had corrupted, which resulted in this:
I did not know the error that could have led to this corruption, but I surmise it was due to too many insertions near a poly-loop, which would result in empty ‘gaps’ between a line of vertices though it would appear flat. Due to this error, I had to remake the entire chestplate, but made sure to avoid mistakes like the potential one I made.
This new chestplate was indeed more stable than the other one and the topology, especially for the neck hole area was actually much better than the last one.