Marmoset – rendering the assets

So i wanted to have a go at using a PBR rendering software to see how the normal maps would look in real time. I also found that marmoset has a plugin that unreal engine and has the ability to export it as a unity package which is perfect for me since i want to turn this into a game environment. I used a free trial of marmoset but ended up buying it because it was so fun to use, the shaders and lighting provided with the software is fantastic and easy to use. The viewport as well is easily manipulated and has excellent quality. Altogether it is an excellent software and very fun to use.

Below are some renders to show how the normals look on the models.


Sculpting in Zbrush

Zbrush was always a program that i wanted to get into, i’m glad i had the opportunity to use it this year, this project gave me a good opportunity to begin to learn the basic and get a good workflow going for making game assets.

The program itself seems to have a high learning curve but over the space of the project i managed to get the hang of it. It took a while to get the workflow right  for sculpting but now i have a good routine.

I will bring the model in combined if its a complex part, split it in the subtool menu, dynamesh each part with a high enough resolution to get a good amount of detail, then merge them again, autogroup them so i can work on them individually while seeing them all combined, then set a morph target and continue to set new morph targets and i get further through the model. This worked well throughout the project and the models turned out very clean but very high poly. I then messed around with decimated the mesh, it lowers the poly count a lot but it leaves the topology very messy.

For sculpting i had a set routine with what brushes i would use to achieve the hard edged blizzard look that i was aiming for. To get this look I used the trim dynamic brush, it gives a nice sharp rock like surface to the mesh. I also downloaded the orb cracked brush set for zbrush, this allows for some really nice clean sharp cracks in the object that look natural or like sword strikes.

After sculpting everything high poly and being happy with the result this is when i would start baking them down into normal maps.

Imaging and Data Visualisation Reflection

The Imaging and Data Visualisation module was a tough module, it was challenging trying to find the best way to visualise data in a way that presents the data but is also appealing. However through the Floating Rome module we learnt the best (and worst) ways to do this. One thing that I especially like in this module and in this course is that failing and doing the wrong thing is always actually the wrong thing. Making mistakes is a great method of teaching us how to do better. I feel that the amount that we were taught was a great amount and the way it was taught was great for us beginners. It helped us to become fairly adept when it comes to using maya as well as making movies and understanding how to make animations. The variation in the module was great, I think that having one based around presenting data early on in the year allowed those who were confident in maya to do well in that area but it wasn’t necessary for the module as other in the class showed excellent ways of presenting data like Kristina. I think having this first was a good idea and allowed us to develop our skills in maya as we went but put our focus on our presentation skills. The second module, the head module, was completely centred around what we had learned in maya. I liked this model because it allowed us to showcase what we had learned throughout the year, it challenged us but I feel like it was needed to show that we paid attention through out the year. Overall I think the layout of the course works really well. It allows for a broad range of experimentation in maya as well as showing what you have learnt through out the year. The separation between a maya orientated module and a data visualisation orientated module works well by dividing the work equally through out the semester.

Modelling the Defence District

The defence district was designed because of Ancient Rome’s militaristic  ways. For a large part of Rome’s history the Roman state existed as an entity almost solely to support and finance the Roman military. The military’s campaign history stretched over 1300 years and saw Roman armies campaigning as far East as Parthia (modern-day Iran), as far south as Africa (modern-day Tunisa) and Aegyptus  (modern-day Egypt) and as far north as Britannia (modern-day England, south Scotland, and Wales). It changed many times over this period, from an unsalaried civilian militia to a professional force. Its equipment was altered greatly through time as well to keep up with the technological advancements in order to expand Rome’s domain or protect its borders.Due to its power and position throughout history it was important to include this as a district to show case how powerful Rome’s military is and how it has stayed powerful even in this new chapter of Rome.

Since our Rome showcases elements from Rome’s past like its architecture and inventions we did the same for this. A classic part of Rome’s history was the gladiator area. We wanted to  include this in out defence district as a means of training the army

Modelling The Historical District

I really enjoyed modelling this district as it’s purely just an area of classic Roman architecture. which is a lot of fun to create and experiment with and try and condense into a small area. I studied a map of Rome and chose the top classic buildings in Rome and located where they were on the map. Where they were situated within Rome is where they are situated within the Historical District. I drew a plan for the layout of the city,


Unfortunately by the time we got creating I had to narrow down how many buildings we could build, in the end I went with 5 buildings (Pantheon, Spanish Steps, Roman Forum, Palace of Justice, Colosseum). I did some drawings of the buildings I would be building, my main goal was to have them be built with basic shapes so that it would transfer into maya. One thing that really helped with this was looking at the plans for the buildings that are drawn onto a grid, like maya. One thing I did change was the Roman Forum, I only modelled part of it due to its scale.

The overall idea of this district was to store Ancient Rome’s history through one of their major achievements which was their masterful architecture. Through history the whole of Rome’s buildings may not have survived but these buildings showcase some of Rome’s historic architecture and it’s ability to stand the test of time.

Here are the finished models and how they fit into the circular layout.

Head Model

The model was a tough challenge, it took a lot of planning and a few attempts to get it to even look like me. However it was very enjoyable. It was one of my favourite projects, I really love modelling on maya so modelling Rome and starting to be able to design a human head is a lot of fun. I wanted to get right into it, my first step was getting reference images, I have multiple from uni which were useful to look at at the start because it gave a good 360 degree view of my head, one thing I was concerned about the model because of my beard but I wasn’t an issue due to its length. When i started modelling the head I took my own photos and imported them to maya for reference.

After taking the photos I lined them up and drew lines across the images to make sure they lined up correctly so the model would be accurate.


My style of modelling wasn’t to chip away at a cube adding more edge loops but to draw polygons and extrude edges around the eyes, mouth and nose and then connect them. I liked this method as it allowed for a lot more freedom with creation being able get the structure of the face and head as accurate as possible. One major issue I came across in my model was the topology, as I went on with the head model I found tri gons, disconnected edges and incomplete edge loops causing strange deformations within the geometry. I remedied this however at the end once the model was completed model i made the model live and used the quad draw to create a new topology for my model. I used these images as references to correctly get the outward loops of polygons from the eyes, nose, mouth etc to minimise any poles in the geometry as well as any non manifold geometry and other errors.

One issue that i couldn’t fix was connecting the ears to the head correctly. Since I modelled the head and ears separately, the ears had a lot of polygons and so the number of edge loops were more than the edges on the head. I did my best to hide these issues behind the ear as we were told if we are meant to hide poles or other errors there. When I created a new topology I tried my best to keep the polygon count as low as possible. Since in the future I want to go into game development and I want to get into character modelling I think this is a goo start. As Alec has previously said, for games ,like phone games, low poly models are preferable in video games since its all real-time processing and processing a high poly model in real – time takes a lot of computer power. The low poly version of it looks less like me, whenever it is smoothed it looks more like me but I think that’s meant to happen. Here are some renders of the updated model.

Gladiator Arena

So as we begin to model we have split up the sections so we can being to design each of the sections of our city. I have been assigned the historical district and the military/defence district, which I am very happy about since they provide a challenge and I’ll have a lot of fun creating the models. So I started working on a gladiator arena last night and finished it tonight. It was actually fairly straight forward to make with the help of tutorials so making the rest of the main historical monuments and buildings from Rome should be fun to do, but maybe not straightforward. I also had a bit of fun messing about with nCloth and having gravity act on materials as well as playing around with wind pressures to make the flags look in motion.

I’m not going to talk through the steps of how I made the arena but you can find the tutorial here.

Here are screenshots of my model throughout the stages.

This slideshow requires JavaScript.

I think it turned out really well, there isn’t a huge amount of detail however I think it would be hard to put in more detail than this within such an expansive model so I will most likely leave the buildings with this amount of detail until I feel like there should be more. It also depends on how well my computer can render out the model.

We had a lecture today on lighting and rendering so I had a go at trying to render out a few images of the arena. You can find them below.

This slideshow requires JavaScript.