In this module, we had to work to create a character. This involved sculpting, reptopoglogy, UV and textures. In the end, we also got to play around with Unreal Engine 5 to create a scene with our character. This was a fun process as it took me through the whole pipeline and I learnt a lot of different tricks to help speed up my process.
To start, I had to come up with a concept for the character. For this character I decided to create a dragon. This allowed me freedom to create a model which I have complete control over.
I started looking at some references for dragons and decided to go off a wyvern design. A wyvern is different from a typical dragon as it has 2 legs instead of 4. examples of references are included below.
To begin, I created a basic sketch of the forms of a wyvern to act as a reference for creating my character block. This is shown below.
This of the forms allowed me to create the building blocks of the dragon. This was a relatively easy process as I just had to trace the forms I had drawn. From this, I moved onto sculpting the forms more. I did this to make the forms smoother and more lifelike, to take the blockiness away from it. I used a combination of the move tool, snake hook tool and the smooth tool in blenders sculpting window. I also found this process easy as I made use of the mirror modifier to create the forms with ease. I decided on adding a lot of spikes around the head of the dragon to create an intimidating nature to the dragon. On top of that, I used a plane to create the wings of my dragon as it made them much easier to form.
From this point, I moved onto the retopology of the model. I followed the process that I learnt from last year by using the snapping tool and shrink-wrap modifier within blender. From this, I manually went around the model creating an outline of the topology first. I then went back to fill in the gaps making a lot of use of the grid fill tool. This tool allowed me to create fill the outline of the model much quicker than manually doing every quad and manually reconnecting vertices. Again, I made use of the mirror modifier for my retopology to make the process smoother. A picture of my retopologised model is included below.
As you can see above, for the most part the process went smoothly. During the process of UV mapping, I ran into the issue with a lot of overlapping UVs. The initial issue was with the wings. Due to the thin nature of the wings, the edges managed to break through either side of the wings. Due to the crammed topology, I decided to completely scrap the wing topology and restart it. This did take some time to resolve, but it went relatively smoothly and I was happy with the results. The updated topology is shown below.
As you can see, the wings look a whole lot smoother now and it allowed the process of UV Mapping to proceed on smoothly.
When starting my UVs, I initially struggled a lot with trying to figure out where to put my seams. Because of this it took some time to get my head around it. After about 2 hours I had managed to make some progress. The UV maps started to look good and I had managed to figure out where to place seams. I had used a checker texture to ensure the UVs weren’t stretching, this was extremely helpful for me figuring out where to cut. The other main struggle I had within this section was with the wings. I just couldn’t figure out where to place the seams. I went to Henri for some advice and he explained it very simply. It was a whole lot simpler than I imagined. After this advice and seeing his examples, it made the rest of the process a breeze. I decided to meticulously hunt through my UV map to ensure all the overlaps were gone. After all those issues, I ended up with a UV map that I am quite proud of. This is shown below.
After my UV maps, I started to do a more detailed sculpt. To do this, I made use of the multi resolution modifier. This allowed me to subdivide without it affecting the mesh itself. This ensured I kept a relatively low poly count to reduce the overall file size and power load on my computer. For this sculpt I went into more detail around the horns and the eyes. I created four eyelids to create a more interesting design. This sculpting was a lot more fun. It was incredibly enjoyable to create a lot more forms and experiment with the designs and tools. After these smaller details around the horns, eyes, claws and wings, I went on to create a VDM brush. This brush would allow me to paint on scales with volume a lot faster than doing each individual one. This was still a time consuming process as I went had to go around the whole model. To create a brush I sculpted a scale on a plane and used the plugin to create the VDM brush. The scale design is shown below.
Again, I made use of the mirror modifier to speed up this process. Although during this process, I had to go back and manually fix parts of the sculpt as it due to how thin parts of my mesh were, the brush was pulling material away from the side I was sculpting on. To fix this I used the inflate brush and the smooth brush on both the laplacian and the surface setting. This helped me smooth out the sculpt and I eventually ended with a finished product I liked. This finished product is shown below.
I was very happy about this final product. I enjoyed the uniqueness I had created within this.
After this I moved onto the texturing. I baked the normal maps within blender to enable me to export the fbx and the files at a low poly count to reduce the load it would take on my computer.
I used substance painter to create my textures. I used a selection of basic materials and modified them to my liking. To texture the majority of the model, I made use of the masking feature within substance. This made it a whole lot easier to create my textures as it sped up the process instead of having to manually paint all the textures, which would have resulted in a messy looking texture. The texturing process went smoothly as I had a much better UV maps to work with than my previous experience with substance last year. This allowed me to create much nicer and smoother textures than last year. These textures are shown below.
This texturing process was also really enjoyable. It was fun to play around with the different materials available and to modify them to my liking.
After this I finally moved onto using Unreal Engine 5. It took a while to install and in all honesty I still really don’t have the hang of it. It was fun to experiment but it was also extremely frustrating getting everything to work. This is where I honestly had the most struggles and I was the most pressed for time. The first main hurdle was actually navigating the camera controls. It was just different from all that I was used to it took some time to get around. I even went in to change some of the keystrokes to make the process easier for me.
The first thing I did was create the scene. I created all the initial atmosphere and lights using the environment light mixer. From this I then moved onto creating a floor. I went into landscape mode to create this. I decided to start sculpting the mountains and hills to create this landscape. This was actually relatively easy and intuitive and was probably the easiest part of using Unreal Engine. This allowed me to then set up my camera to a nice view and then start with the assets and textures for the landscape.
I used the Quixel Bridge feature within Unreal engine to gather my assets. This provided a lot of free photo scanned 3D assets and textures for me to choose from. I had decided to settle on a desert/canyon theme with reds oranges and browns to match the design of the dragon. I had the idea to create a landscape that the dragon could camouflage to hunt its prey. I settled on a few rock formations and plants from Megascans to create my scene. These assets are shown below.
I used a combination of these within the background environment to create a barren landscape within the cameras view. This final view is shown below.
For the sky and lighting, I went with a green tint to show the toxic, dead nature of the atmosphere of this landscape. This helps to add to my context that little to nothing can survive here. This ultimately helped me finalise the context behind the dragon. It helped me decide it is a mutated lifeforms born from a barren world in poison.
After this I added my mesh. This again is where I ran into the majority of my issues with unreal engine. The first issue after importing was the bones. The bones I had created in blender did not translate well to UE. They had managed to be distorted and in different placements than I had placed them in blender. I could not find a way of resolving this and ultimately had to settle on a much more simplified pose than I had created within blender. This really frustrated me as I was not able to fix it at all. This pose is shown below.
I had initially planned on having the claws out in a hunting like dive but due to these constraints I just couldn’t fix it at all.
The other issue I had was with the texture quality within UE. I had exported all textures at a 4K resolution and passed them into UE. For some reason, the textures always seemed a bit blurred, like the camera focus was off. I couldn’t work out what was going on. I searched online and tried creating a static mesh and using the Nanite feature. Even with this I couldn’t fully fix the textures. This really started to annoy me and I couldn’t get it working at all. I even went back and tried exporting in 8K and even went to check the rendering settings within the project settings, but I just couldn’t fix it.
Ultimately, I ended up with this as my final image from UE. I’m not the most pleased with how it came out but considering the time I had, it was much better than I anticipated.
In reflection, I had a lot of fun doing this project. I got to explore new programs and to develop my previously learnt skills. From the beginning, I had a lot of issues with mental health and this put a real big time constraint on this project near the end. I really tried the best I could near the end to try and create a polished product. This time constraint really showed when using UE as I had not a lot of time to troubleshoot and no time to look for answers and advice from my lecturers. I definitely could have done a much better job near the end and I plan on putting a whole lot more work in next semester. This is a real goal I am looking forward to working towards so I can start to enjoy this course a whole lot more. Overall, I really did enjoy many aspects of this assignment and I believe I have improved a lot on the skills I had learnt during last year’s sculpting.