Name: Jonah Holden-Maillard

P Number: P13193617

Email: p13193617@myemail.dmu.ac.uk

Project Title: Apocalyptic Weapon

Overview: This is a weapon based on a concept by Mike Kelly and adapted for use in a game such as Bioshock or Dead Rising. The weapon was modelled and unwrapped using 3DS Max and then textured in substance painter. The handed-in, game-ready model uses a 1024 x 1024 texture and 10,000 tris, however the model displayed here uses an extra 500 tris and a 2048 x 2048 texture.

Design: A starting concept was provided along with the brief, but it was also stated that we should iterate and improve upon it. After several iterations, I settled on a variation that improved the real world believeability and readablilty of the firing mechanism.

Unwrap: Unwrapping was done in 3DS Max and then packed using iPackThat.

Textures: The model was textured using substance painter. The in-game weapon uses (from left to right) 1x 1024 x 1024 Albedo, Roughness, Metalness, AO and Normal.

Final Renders: Rendered in Marmoset Toolbag 2 and Photoshop.

Conclusion: This has been another largely successful project. I'm at least satisfied with every aspect of this asset. My time management was excellent this time around and I feel I spent the correct amount of time on each stage of the process. The final stages of the process like texturing can sometimes end up getting neglected, but I left myself plenty of breathing room to really push the textures and subsequently get some nice renders. The modelling is efficient and clean, but some areas are optimized more aggressively than others. This is due to butting up against the tri limit at the end of the modelling phase as my concept added a second major element with the propane tank. Aside from this there are some other minor improvements I'd like to make, but for now I'm happy to call this finished and proud of what I've produced over the past 3 weeks.

Just a quick update before my hand-in post on Friday. I seem to be mostly on schedule, although of course I would've liked to be ahead of schedule by this point.

I finished modelling last week with the final triangle count hitting nearly 12,000. I optimized it down to around 10,500 without too much trouble. Ideally, I would get it under the 10,000 limit but to do so I would need start removing important loops and it would start to become noticeably lower poly.

​I may reduce it down for the final hand-in but currently I am reluctant to do so as it’s within a 10% margin and my design is significantly more complex than the original concept. At this point I’m satisfied with the model and ready to move forward with it.

In previous hard surface projects, I’ve been very cautious about sharing UV elements and optimizing texture space because I can’t (sorry) unwrap my head around how this affects hard edges in the normal map. I haven’t been able to get any definitive answers online, mostly because my specific query is very difficult for me to put into words, so I’m going to have to find out through my own experimentation.

​With this project the texture budget is extremely tight so I decided this would be the one and threw caution to the wind. I mirrored most of the…mirror-able shells, detached and offset them. They’re highlighted below in green.

And so far so good it seems. Aside from some standard mirroring issues with text on the propane canister, the normal surface transitions seem to hold up fine as long as they are offset during baking.
​
Once again it seems my caution was a result of chronic overthinking and I could’ve been overlapping UV’s all this time. On the upside though, future projects will feature much more efficient use of texture space and this problem will no longer keep me up at night because now I definitively know – yes, you can share UV islands with hard edges.

It’s already the start of the final project before the final project. Hopefully this one will be a good one! I’ve chosen to do the apocalyptic weapon brief. I made a weapon during the summer, so I’m familiar with the workflow and have some insight into the sticking points of a project like this. By far the most time consuming and troublesome parts have always been during the high-poly modelling phase. To get around this and try and complete this project quickly, I’ll be experimenting with a slightly different workflow, but more on that next week.

This week, before I started modelling, I spent a few hours concepting the weapon. Included in the brief are 6 concepts by our lecturer Mike Kelly. The brief suggests we select one and spend a short amount of time improving upon it, before creating it in 3D for a semi-stylised game such as Dead Rising or Bioshock Infinite. So I began by breaking down the design I was most drawn to, highlighting its positive and negative aspects. 

​I selected the ‘blowtorch sword’ because I thought it was a really unique and fun design. I loved the light weight aesthetic and elegance compared to the other designs. But where I thought it fell down was in its believability, particularly in the firing mechanism and the weapons centre of gravity. So I set about improving it, while trying to retain the visual elements that caused me to choose it in the first place.

After several iterations, I settled on a final design, however this will continue to evolve through-out the project. During the modelling phase, I realised I was going to really struggle with the 10,000 triangle limit if I wanted to retain such a complex silhouette. I toyed with the idea of cutting the red blowtorch but Mike pointed out to me that without it, the design becomes really boring. I definitely agree with this so I will probably exceed the triangle limit to create a ‘portfolio version’ and then hand in a chopped down version that meets the briefs requirements.
​
So other than that, I’ve just been modelling away and day dreaming about my FMP. In the space of about a week, I’ve gone from about 20 vague and pretty ridiculous ideas, to 2 solid, achievable ideas. In a couple of weeks it’ll be time to present them to tutors and the rest of the year, so by then my ideas should be fleshed out and viable and everything will be OK...right?