Name: Andrew Simpson
P Number: P13202810
Email: Andrew_Simpson@hotmail.co.uk / P13202810@myemail.dmu.ac.uk
Project Title: FMP Bali Temple
Overview: The idea behind this project was to create a playable demo of a Realistic Balinese temple environment that mixes fantasy elements that are found in Tomb Raider and Uncharted Games.
Research and Design: Research for this project was mainly done by me as Gathered as much knowledge and visual reference from different sources. I began by collating as many images of real life reference of the type of environment i want our level to portray and this was done through a mixture of ways. The first and quickest way was to image search on google and pintrest to expand our visual library. I also used Google map's road view feature to search through the temples and local environment when google image was not enough.
Luckily some prominent temples in indonesia such as mendut, prambanan and Borobudur were part of google's efforts to create a virtual map of historic landmarks. To accompany this I enlisted the help of family members and friends that live there to take as many photos as they can. Further more, it was helpful to note that my experience in that region helped me to give crit and advice on the work of my team members in recreating the feeling of being in Indonesia/Bali.
At the start of each project I do, I drew quick sketches that would frame the overall layout of the environment/level that we are creating, this helps both Mark and Denise have a clearer picture as to how the level will look like when it came to blocking out and concepting.
These are intended to show a rough and crude idea.
Over time the design of the temple changed rapidly as we found certain structures and shape language would not transfer as well into a digital game and further more give the sort of feel we were going for in the playable demo. One example of this would be the fact most temples in the region were situated on flat open land, this is especially so with the temples situated on the Island of java. Another would be the shape and intricate designs of the balinese temples such as pura besakih would be lost to a general noise because the level proportions became rather large and another issue it brought up was time constraints.
Unwraps: Here are a few examples of unwraps on some of the Assets I made. Note: modular walls and temple pieces were unwraped to fit the tileable wall texture and as such would not look great to show. All pieces were unwrapped and manually packed, this was done because in my previous projects automatic packing softwares such as Ipackthat did no produce good results on organic and simple pieces.
Textures: Here are a few examples of the textures I made for my assets, the texturing program I mainly used was Quixel however during the project i have gone through various iterations using different texturing programs for specific parts of the level, such as Bitmap 2 material for quick landscape textures and the learning and use of Substance painter and Designer, of which are very alien to my texturing workflow. It was enjoyable to learn rapidly through this project in how best to improve my texture work and at times been a frustrating struggle to strive for the best quality industry standard textures.
I've learnt a whole host of new methods of producing nice texture results, most of which came through the use of sculpting and my attempts to copy the pipeline of industry professionals. This was through research of their pipelines be it through their portfolio breakdowns, through observation of their art station assets and by examining the quality and standard that is present in games.
Sculpt Renders: These are examples of the sculpts made for the project.
Renders: The following are some renders of assets with textures on them.
What went well:
1 - Level creation
at the start of the project we were rapidly going through iteration after iteration making big changes and seeing big results coming out of them, one such thing was cutting down the size and scope of the level. Originally this environment was planned to have a village area that we would walk through to establish a sense of scale and setting of the environment. This was cut in favour of making a smaller more tight focused environment and i think by doing so allowed us to focus our time and attention to getting a nice looking vista, temple courtyard and cave.
Although we did achieve this outcome and done so well I Personally believe that if we cut out more of the level in favour of just a simple vista benchmark and temple courtyard area we could have created an even better piece of work with a higher density of population and a greater standard and quality in our work.
2- Communication between Colleagues
As with all group projects there will be ups and downs and to be honest there have been a few times when we disagree with each other and it has led to poor performance and morale, however it is the fact that we were able to learn and grow from our difference of opinions that allowed us to commincate more effectively between eah other. The communication between the group has been overall phenominal as we are able to sit together and offer feedback and crit easily between us to improve our work. This was none other better exemplified by Denise whenever either Mark or I was stuck with ideas on how to improve the level or unsure of how best to start creating our assets we could simply ask Denise to draw, sketch or simplify by breaking it down into further detail.
3 - Simplified engine file management system
Dom who has been extremely vital to our project was also able to further enhance our workflow as he set up the perforce source control, this meant that we were able to submit our work to a centralized server in order for us to have a more efficient downloading and uploading of engine files. This also meant that as there are essentially three people working on the level (Myself, Mark and Dom) we would all know who has the level checked out at certain times preventing any work done on one engine file not recognised in other files because it was out dated. Not only that it also means that any work be it something small or large does not need to be labouriously copied onto the main file. This meant fewer files in general, a backup centralized file on a private server and a more efficient environment creation workflow.
What did not went well: I'll try and keep this short.
1 - The size of the level
As said before I mentioned the fact that we had to scale down the size of the original level, this was done as the tutors did not see a way for us to accommodate the time to create a village area, this I felt was also true of the final outcome. The level would have benefited far better from a more compact and tightly focused area. I expressed my concerns about the level being far too big and to an extent we have cut the level size down a lot but in order to justify the gameplay elements we needed to have the level that had progression. However this then produced the inner cave area which does not make much coherant sense, this was a rescue attempt to salvage what we had at a late stage in the project. If we focused more attention to the blockout phase of our level we would have more understanding as to how we would finish the level.
2 - Learning of new software
This project stood as a mile stone in my familiarirty with digital sculpting softwares such as zBrush. In previous projects I only used Zbrush for small things such as a quick sculpt of cloth folds or testing tiling rubble sculpting. Never had I used it in such amount nor in various ways. This proved to be one of a the most time consuming parts of the project. As i progressed through the project using Zbrush more i found ways that would have drastically reduced the amount of time I would have spent previously. One particular point would be the Eagle statue and it's spiky feathers on the body. This was one of the most time consuming parts of my Zbrush sculpting, the fact it had to be individually placed and intersected was annoying and thus slowed down my workflow.
This learning of software was also prevelant in understanding Speed tree, marvelous designer and at one point having to learn world machine which was then scrapped in favour of faking mountains. The bulk of my time in this project was getting to grips with the software which is a good and bad thing. It does mean that in my future usage I will be faster and more familiar with the programs but it has reduced my workflow. Side note: 3D coat was a particular problem and the assets that I have made have had to be manually retopologised as the program kept bugging out, This was also true when I came to baking AO and Normal maps and numerous other assets. Only when utilising Get Baked script was i able to perform consistent Texture bakes.
3 - Time management
My time management has been better than most of my other projects in a way. Working within a group has made me rush some things and become careless in organizing what things need to be done. I feel as if I need to fully set out and write down exactly what must be finished on a week by week basis and to underline basic asset list breakdown in spreadsheets when working as a group as I tend to get confused and get narrow sighted on each asset i do. Its been one of my downfalls where I focus too intently on a particular piece that I lose sight of the big picture.
Improvements for the future:
Blockout more: I think if I were to do this project again I would first spend more time blocking out the overall level more and go into detail the exact player path and layout a strong visual shape language to guide the player through the level using key visual identifiers more akin to uncharted's level design. There was a huge debate during the project about the placement and design of the courtyard of the temple and where the player's eyes are being led. I tried to do so during my initial stages of blocking out and sketching however it has since been lost in favour of getting things done and in. The level's progression was also rather confusing as the player starts out somewhere nice and ends up inside a monster's mouth? Either having more time or if we spent our time better planning the level's story we would have had a more solid outcome.
Use alphas in sculpting more: One thing I neglected to do throughout most of my sculpting phase was the use of alphas to create the big picture sculpts. This would produce a faster result in my sculpts, another thing is to not be so precious with the sculpting phase. The main enemy in a game artist's path is time and when creating assets it should be cracked out as fast as possible whilst trying to maintain a decent standard. Alphas would provide this lee way.
Keep the level small: I've mentioned this through my post mortem but it is something that i should stress. The level is far too big. Having a smaller area does not neccessarily mean a less impresive one. I think that if the level was kept to the vista benchmark area and the temple courtyard and perhaps a small shrine inside the temple the level would have been enough and there would still be a coherant narrative, however things change through out the project and what was meant to be a small cave tunnel that leads to a shrine room became an inner temple cave and huge cave lake with a monster inside it. When i spoke to Casto he mentioned the crit that the shrine temple room was bland and generic and I should think bigger, it led to something of a lord of the rings massive cavern darkly lit with one single lightsource to what is now a boat ride in a cave. I think the cave was built out of pressure due to time and lack of planning as was mentioned before.
Spend more time in lighting and post processing: One thing to note about the level is it's confusing lighting. The sun is at a sunset position however the level is still relatively brightly lit. perhaps changing the lighting situation to have a midday position would make more sense however the vista would be compromised, one other thing to note is the position of the lightsource and the path the player takes. Most of the level the player walks away from the sun light and so the subtle roughness variations and texture quality I spent time doing was not best appreciated as the textures look flat when the light is behind you. coupled with the fact that textures look completely different to what Quixel or Substance painter had made the process of texturing harder than I thought it would be by having complex shaders to get the same result.
 |  |  |
|---|---|---|
| | |
 |  | |
 |  |  |
| | |
|  | |
 |
Conclusion:
To conclude the post mortem, I do believe that we have met our original brief of making a playable demo that shows and demonstrates our own abilities that fit towards our own specific Game art career. The team worked very well together and the result is a very good example of the combination of our individual strengths. It shows we can do foliage, modular assets, good efficient UV unwrapping, set dressing, level design and population and finally a good standard of texturing.
As with all projects I learn more after doing certain things that were wrong and ending up with results that were dead ends as I now am exponentialy better as a game artist after overcoming these pitfalls. There are plenty of improvements that I can think of from various points in the project, workflow wise and design of environments but all in all I truly feel we have done an extremely good job with our Final major project and I am proud to have this final result.