Brief Introduction & Research
I chose to centre my dungeon project around a ship's Lower deck / Orlop. The project should be contained within an 8 by 8 meter cube and it needs to be modelled with a character's proportions in mind to keep sizes consistent.
Dungeon located at the very bottom of boat, specifically a Historical Naval Warship from the late 18th century to early 19th century (1700 - 1800)
because it offers cool wall structures & it is where the loot / prisoners are kept. Treasure, creatures from around the world in cages
A dungeon where they keep fantastical beasts?
SOUNDS: creaking wood, echoing footsteps, rolling barrels, squeaking oil lamps as they sway , rats squeeking
Light sources: Deck Prism, oil lanterns, Staircases / ladders going up to higher decks
H.M.S Victory & Cutty Sark interior
ANIMATIONS: Static vs Dynamic animation = swaying items, animals, rodents
Personal items to break up the generic space so it doesn't look like an ikea / show room
Large warship / pirate ship : HMS Victory
Analyse an In-Game Environment
a simmilar location to my idea (boat interior). (I had actually forgotten there were ships in genshin)
I know of instanced assets in scenes but not of entire sets being reused over and over so often and effectively as this. The game artists managed to get away with duplicating entire sets of props all over this place because they package the scenes using different lighting, foliage, rotation etc to make it seem as if they are different items entirely. I managed to find at least 8 of these scattered around the ship but each encounter felt very different. Their use of composition and focal points distract players from looking at these insignificant collections too hard (often filling up darker areas of the environment) and draws their attention to the puzzle mechanics and story instead, which is where most of the game developers' time is spent and where the unique assets are seen.
The chains created in the boat environment was done so well! I had to distort the camera and look at them really close to bypass the illusion but even then, once I stepped back, the illusion snapped right back and I had a hard time believing it was made up of only two alphas! Definitely something I should consider for my project because I am thinking of having lots of chains and cages in my dungeon scene.
The chains created in the boat environment was done so well! I had to distort the camera and look at them really close to bypass the illusion but even then, once I stepped back, the illusion snapped right back and I had a hard time believing it was made up of only two alphas! Definitely something I should consider for my project because I am thinking of having lots of chains and cages in my dungeon scene.
I was also quite interested in the netting which hung from the ceilings - this helped break up the environment with variety. I was wondering how I would add alphas effectively to depict ropes for my own project (because I am definitely not going to sculpt them to use up so much of my budget, time, and sanity on topology & unwrapping - *flashbacks to the braid on my supervillain*) and so finding a really simple and effective way here was incredibly insightful!! They used alphas for these thin pieces of netting (exactly how I was planning to do so) but they also added simple cuboid tubes that run along the alphas' boarders to hide the flat side. This made it really difficult to bypass the illusion. Adding a barrier so that players can stand on the netting also fed the believability of them being real netting.
Underwater ruins - Institute of Natural Philosophy
I could spend hours just looking at the graphics of this game and I would be happy so they are clearly doing something right and I believe it will be useful to analyse their environments. The colours are highlighting the assets which have been instanced.
Tiling textures? Decals? Repeated asstets?
How is the player guided through the environment? Assets? Lighting? sound? Narrative effectively being conveyed / showcased?
How do I plan to implement these aspects into my own project?
This was the first attempt at blocking out my dungeon. I dragged in an 8m cube and made it transparent (Alt + X) to give myself a limitation to feed inspiration. To speed up the process and to make use of instancing, I populated my scene by duplicating items.
This compositions is sort of interesting? but I thought the space felt a bit too predictable. When creating this scene I imagined how a chracter would travel from higher decks down to this one. Perhaps they cover the trap door with items so that others are less inclined to snoop. The focus oint of the monster's escape is hardly the centre of attention or interest point so this is a poor composition.
In my second composition I decided to shift things around and make the space less predictable but still keeping in mind the importance of composing a scene so that it directs the audiences' attention to the interesting focus point - which should be the creature's escape. I was happier with this composition so I wanted to briefly check it's lighting potential to see if cast shadows would aid my narrative.
Rough block-out in Unreal to judge the composition, cohesiveness of it's layout, and see if the arrangement speaks 'ship'. The curvature of the wall and the stereotypical barrels, ladders, and hanging lanterns help to push the visual identity further. I was sort of sceptical about the wall infront which might be too thick and obscure too much of the environment but if I dress it up with purpose it might work to give the space a cramped and eerie atmosphere. It also helps to counter the leading lines that point away from my desired focal point which is the broken floorboards in the back.
I really liked how this composition was looking but the balancing of focus was lacking. I had pushed the point of interest (caged trap door) all the way to the back and left a massive empty space at the front so it felt unbalanced. I liked the positioning of the staircase however, because it isn't something that requires immense focus that it should be near the front of the composition but is visible enough to give the scene more spice.
To improve upon the design on the left, I moved the trapdoor to the front. This worked really well. I also tested out whether the wall would be best if I dropped the wall that slightly hides the baskets right at the front but I'm still sort of in favour of the wall th the front because it gives more layers to explore int the space. The floating slab of wood feels unnatural and it doesn't give any plase for the hanging ropes to connect so they are just floating.
I combined these points above and made this final blockout. I think I made the most of modularity and managed to create an environment that will easily tell my narrative visually.
Some shots of the dungeon in Unreal.
Advanced Modelling Techniques
More composition attempts:
Played around with the latice modifier (which we were introduced to in this lesson) to quickly build the cages. This modifier is very useful and will be really helpfull for this brief. Despite this, this composition turned out quite poor. everything is hidden beneath the upper floor.
I tried moving the pillars a little to the centre of the composition and this was quite effective. I also tipped one of the baskets (which will be filled with fish or some kind of dead feeding resource) so that it would show the mess the small kraken has made in it's path.
Paradise Tanger Bird / Chicken
Kraken Slime / blue blood?
Advanced Modelling Techniques
I talked with my lecturer how I could go about making use of a tiling texture whilst also having a natural transition from the environment to the negative space. He suggested I complete the tiling texture first and then use that as a guideline to where I should cut out the loose boards.
Started on the pillars in 3ds Max
Interesting texture and imprinted details which I could take inspiration from >
In ZBrush to add carved details and wear and tear.
Whilst using the Scribe Chisel it started to affect the inversion of the strokes. On the left side you can see the strokes are protruding from the pillar when I wanted it to look like the right; Chiselling inwards. I tried to use Backface Mask and also resetting the move gizmo which always seemed to work in the past but it didn't work this time...
I didn't want to symmetry the imperfections away because I had already begun to add asymmetries like scratches and wear and tear but I guess this is a lesson to myself to always start on symmetrical details before adding asymmetries to my sculps to avoid problems like these above.
The centre of the pillar is also effected... so I separated the pillar into three sections; top, middle, and bottom.
Added wreckage to the pillar using brushes Deco 1, Rake
Such a beautiful relic (found on a boat) that matches the vibe of my environment.
The mesh was being buggy and you can see sharp artifacts sticking out of the corners. I got rid of these by using Rect lasso
Symmetrised the back frame
I didn't use this chair in my final environment because I was already over budget. I also didn't like the outcome. It didn't look right and I feel like a stool would have been better.
Having problems with the bend modifier?? It's not bending it it's only squishing it
3ds Max view with smoothing groups.
Realising ZBrush doesn't care about the smoothing groups and shows how boxy the mesh is.
When I brought my barrel to ZBrush, The faces would become triangulated. This prevented me from using 'Creases' in ZModler properly. Instead of creating a sharp crease around the circumference of the barrel, it would follow the zig zagged lines and go all over the place. My lecturer said this was because ZBrush can sometimes dislike imported meshes that aren't object file formats. To overcome this, Just export the mesh using an .obj file format
Exporting an Obj file format helped solve the issue I was having but it also showed me the N-Gons I left unsolved. I left them because I thought it was unnecessary to clean them up just yet to bring into ZBrush but I am now aware that this does matter when trying to create high polys.
Removed all N-Gons back in 3ds Max.
I used ScribeStandard and DamStandard brushes to carve details into the barrel, and used Flatten and TrimDynamic to make corners look more realistic.
Moving onto the half barrel bucket; I was having issues with the symmetry tool. To exaggerate the issue to ease visibility here: You can see that some sides were cutting into the sculpt whilst others were extruding out of the sculpt. I'm still not sure why this issue occurs despite having back face mask on.
Here I explored different methods of modelling hatch gratings.
In my first attempt, I used a Boolean modifier to subtract holes from the mesh with a cube. Then, I applied an Array modifier to duplicate these holes evenly across the surface. While this approach worked, the result is overly simplistic, and I realised I would need to spend additional time cleaning up the topology.
For my second attempt, I noticed that the underside of these grates are slightly offset. To replicate this, I used the array modifier to duplicate the cuboid, creating the grid and offsetting the intersecting lines.
This second attempt was much more effective as it added more depth and produced cleaner topology that won't require additional work.
High Polly version created in ZBrush. This is it's wireframe after it has been decimated.
Using a long cuboid to create one bar, I then used the Array modifier to multiply it evenly across the hatching to create bars.
I then broke these bars apart by using soft select and the Bend modifier
To break up the hatching underneath, I brought in a cylinder and used it's many faces to bend it into this strange crunchy shape. I then used Boolean to subtract it from the high poly hatching.
The staircase was difficult to retop but I found a pattern in the way the steps joined at the sides and created loops which combined the front edges to the back edges. These had to be tilted loops because the mesh is tilted.
I reference an antique oil lamp for my boat.
This is the first time I used the Boolean Modifier but I'm not sure how to duplicate these cylinder sticks to the other two sides. I also had a hard time fiddling around with this modifier since I don't fully understand all of it's uses yet. It turned out that it was just a bug with the software and not me. You just have to move the camera a bit to get the modifier working properly.
These bottles are meant to be bottles of Whale Oil because they are used to fuel oil lamps in boats. I had to cut these out because I was over budget.
Really simple, just a cylinder with it's top face inset and pushed downwards. Chamfered the edges and made it extra round.
I used a ZSphere to make a long tapered body. I then added a sphere for the eyes, pupil, and jaw. I used a cylinder for the thin fins (because it has a nice flat side that I utilised) and I used Bend Curve (transpose) tool to quickly mould the cylinder into fun shapes by squeezing and bending different parts of the transpose line. Back in 3ds Max I used Bend to make them look organic and flimsy in the baskets.
To make the most out of instancing, I should utilize symmetry. Depending on how much of the budget I have left, I could make three variants, bent in different ways.
ALPHAS : CHAINS & ROPE NETTING
I started with a cube, then I inset two faces on opposite ends and removed the inner square. Then I chamfered the edges to make them extra round. Used Array modifier to duplicate and offset the chain rings. Brought it into ZBrush to smoothen the faces and to add bumps and scratches.
One cylinder (make it extra tall). add loops to make sure there isn't faceting and sharp corners. Array modifier to duplicate it radially, the the twist modifier to twine them together. Repeat that process (Array radially, and twist). But this time, the bend needs to go the opposite way so that it counters the stretched geometry from the previous twisting. The rope still looked short so I used array again (this time on the x axis rather than radial) but there were still seams showing. I didn't think much of them because I thought I could smoothen them out in ZBrush but it turns out they don't smoothen out very nicely but it doesn't really matter since the alpha will be so small anyways
Prep for export & Baking;
Colour each asset by picking a colour, going to; ZPlugin > Subtool Master > Fill > Colour
(will colour everything currently visible)
everything is merged down (PLAIN SEPERATE!) (have duplicates as backup)
everything being exported in the same polygroup (Ctrl + W)
I now realise the importance of making each asset a different colour in ZBrush before exporting it out because this made it a bit more difficult to utilise masking
You can see a thin white line
The base colour is now coloured. This is the first time I have seen the 'bleeding' of the background colours my lecturers have talked about before.
To retop the rope I used a cylindar and added loops to it and conformer it into place. This was quick and it did it's job despite it loosing a bit of realism in the final silhouette.
Cages are going to be extremely expensive to have. I should have thought about my budget when deciding on the theme for my dungeon. I tried using the lattice modifier to quickly create the bars from a simple cylinder but there where issues around the verts. Attempts at cleaning them up were really messy (seen above)
I tried other methods like using symmetry modifiers or building cages myself using inset etc..
The floor is made up of a plain with many edges going across it. I used soft select to bend the downstairs flooring like above. To add the holes for the drainage doors, I selected the necessary amount of edges to connect them, and then deleted the faces within.
Created steps by holding shift on the edges and moving them in place, and combined both floors together. Then used the shell modifier
Breakdown Existing Textures
I was initially planning to sculpt the whicker basket texture in Zbrush so I watched this video linked on the right >
However, after being introduced to Substance designer - a software that excels at tilling textures - I decided that creating the texture using nodes would be far more effective and less of a hassel than if I were to sculpt it. It would also be reusable for any asset I wish.
Introduction to Substance Designer
Playing around with blending together the grunge noise and the brick generator to see how close I could get to a semi-stylized wood texture. I couldn't get the warp to work so I'm going to have to look into that.
Created a brushstroke using a shape and then using noise to slop blend. Then tiled it so that it would have a semi-stylised look.
I'm confused with how them manage to make the tiles stick out like that. I feel like I don't fully understand the normals and how I am using them.
I now need to add colour & value variation, height difference, and chipping to the edges of the planks.
This turned out very dark.
I kept getting LOTS of errors in designer. There would be stripes that appeared in random places despite it never showing up in any nodes prior.
Height gradients would also act glitchy and would show up in streaks.
I was unsatisfied with this material because it doesn't seem like the kind of flooring you would find in a broken and rotting ship.
For some reason, I had issues with the Flood fill to Gradient node because it was creating these strange barcode-like lines on some areas. The map was the correct size and the prior nodes had no errors in them either so I wasn't sure why this was happening.
These final nodes are needed to pack the maps into the three we need: Base Colour, Normal, & AORM (Ambient Occlusion, Roughness, Metallic). If you don't pack them using these, AORM is exported separately which is unnecessary.
What is happening??? all of my materials turned purple..
To edit the tiling frequency of a material, add a Texture Coordinate Node. The number used is multiplied to get the final result so instead of using negative numbers, use decimals (0.5) to make the tiling more frequent (and the texture smaller).
The editing window disappeared at this stage but you can get it back by going: Window > Views > Properties - Paint
LIGHT COMPLEXITY CONSIDIRATIONS
I lit my block out in unreal to get a feel for my environment. On the left you can see the sources of light - glowing eels & lamps - and on the right you can see the use of an extra light that helps to reduce the shadow intensity / increace indirect light. This helps to reveal the details of the environment whilst keeping it mysterious.
When importing my final assets into unreal, I had to move the lighting around because it is larger than my block out.
I had thought I lost the floating assets on the right of the wall when I was unwrapping but it turns out that it remained there, its just that it was hidden in the unwrap. to make it appear again, click on the 'Eye' Icon twice.
it's really fun to see the plains become alphas in an instant.
Ctrl + i = inverse selection
I needed the lamps to glow but only the glass areas. To do this I need to set the faces with different material IDs. In Unreal I need to duplicate the Texture so that one texture is the unlit version whilst the other is lit.
I was trying to get the lights to look the same way it did in my blockout because it worked really well. This version has too much contrast and the colours of the lights are unnaturally saturated so I changed the rect light for a point light which improved the strange light scattering
The puddle was looking more like paint than water so I had to go back and forth until I settled on a deep blue colour. It was still too opaque to look like a puddle so I brought the decal up a bit so that the transparency was increaced.
Alphas: Chains, Rope, Feathers
I used this tutorial to animate one of the oil lamp's swaying.
This is the blueprint for looping the Oil Lamp's swaying.
Screenshots & Post Mortem
I hadn't really set any goals at the beginning of this brief, only to learn new things such as the new software: Substance designer, and work on improving my knowledge on the software I have used before. I knew I wanted to portray the narrative of a beast who has escaped the depths of the boat and is now wrecking havoc on the rest of the ship. I think this was quite an ambitious narrative to create with the time and budget we were given and although I don't believe I managed to display this narrative, or the mood I was going for, I am satisfied with the final screenshots of the environment that I have created DESPITE how far it is from my initial idea.
In my last environment piece back in Block 2 last year, I felt dissatisfied with how I hadn't given myself time to dress the environment with props. This time I really wanted to focus on props because I thought it would make the space feel interesting, but this is where I think I slipped up yet again. I think I focused on the props a bit too much and I ended up leaving the walls and architecture as an after thought. Now I have finally realised that I need to find a good balance between both. I need to think about the architecture and the props simultaneously. They are meant to work in unison after all.
Yet again, every aspect of my project is pulling its weight APART from the textures. I really do struggle with colors in particular and I personally prefer my project before the textures where added. The narrative isn't being displayed either and it is left looking quite generic. Then again, I didn't so myself any favours when I chose such a prop-reliant environment. I wanted to sculpt more creatures such as rats (makes you think of boats), and fantastical creatures to put in the cages but that would have been far too expensive. I do see the influence Genshin's environment had on my final outcome - from the glowing particals to the colourful lighting set up - but this wasn't my plan to make my environment look so whimsical and colourful.
I am so happy to have been able to recreate the chains I had seen in game for my own work. The outcome looks rather whimsical and inviting which was NOT what I was going for. The warm pinks, purples and blues were what worked the best with the wood colours but I had wanted to use a more neutral colour pallet to make this space feel less like a toy set.
I found this 3D environment piece by Margaux Duboc and I instantly recognised it as being inspired by the concept artwork by Andy Walsh (Below) . Now THIS is what I had hoped my dungeon to have turned out. it looks ominous, cold, and intriguing with it's use of lighting, staging, composition, textures, and props. The biggest contrast this environment has from my own is the saturation in lighting, and textures. The lighting compliments and enhances the textures and mood of the environment making the space feel cold, damp, and rotting. The props also feel more impactful and meaningful than my own. Although it is an extreme fire hazard to have so many candles within a ship, especially on the ground where it would be so easy to trip over, it proves that whoever owns this space is clearly not afraid of taking risks. The broken floor boards and the growing barnacles also suggest that this ship is rotting and probably not being used as a ship but rather a space more like a stationary mansion. Due to time constraints I didn't even look at references of broken floor boards and so my own recreation looks very unbelievable in contrast to this artist's work where the floorboards look amazing. I tend to do this a lot when a project grows nearer to it's deadline: I forget to look at references and I forget Pinterest exists. This is probably why my textures often come out looking basic and unsatisfactory. This is something I should definitely keep in mind for the future. I should collate a reference board right at the beginning that I can easily access throughout the process. I should also remind myself that looking for references isn't a waste of time when I am nearing the end of the project.