I’ve got 2 posts worth of writing this week because I finally ‘finished’ the first vehicle. TADA!
It could use some polishing and really needs a base with some extra details, but I’ll get to that. Right now this seems like a good time to evaluate my process and see what I should do differently moving forward.
- This is not a design project. I never intended to do any concepting and it was never written in to my brief. I quickly discovered that trying to shortcut this by combining 2 existing vehicles makes modelling much more difficult and can lead to an unexpected final outcome. For the next 2 vehicles I will stick to existing designs and simply make minor alterations, rather than trying to come up with something completely new.
- I should have the textures planned out once modelling begins so that I can unwrap and split up material ID’s as I go. Seems like an obvious one, but I chose to do this in one chunk with the pickup and it was extremely painful.
- Keep the vehicle topology even with good edge flow and split the body apart into panels after the fact. Again obvious with hindsight, but apparently I had to try out the hard way first.
- Model everything high poly and optimise down. I thought that it would be more time efficient to model everything at the final level of detail because more polys = more time right? Apparently not.
- Splitting texturing in to 2 passes, a ‘base pass’ and a ‘dirt pass’ in Substance Painter, works really well. But I need to be strict about separating the 2, or else I end up wasting time trying to unify the properties of dirt on different components. And that’s even less fun than it sounds.
- Glass created in Substance Painter seems to only render well in Substance Painter. This is another point in favour of doing 2 passes. It allows me to do a few things in between, which are still easier in Photoshop.
- Less GTA, more Dying Light. My brief specified Dying Light as an example game because the vehicles are mainly used as props and they’re not driveable. GTA on the other hand has driveable vehicles with simulated damage, so the models are much more complex. As I was referencing a lot of vehicles from GTA V my pickup model became more complicated than it needed to be. Again an unnecessary waste of time. The next 2 vehicles will be as simple as they can be and I will instead focus my time on making them visually excellent.
- There is no reason for me to include engines for the vehicles in this project. It isn’t written anywhere in my brief and it’s an enormous time sink. Unless I have all the bonnets open, no one is even going to see them. I’m happy that I’ve done one with an engine, but I see no upside to doing another.
- Creating the Pickups interior was very difficult, time consuming and is visually the weakest part of the model. I put this down to 2 reasons; a lack of good ref and trying to cram too many unique texture islands into a small sheet. To solve the texture space issue, I’m going to reuse more texture space and simplify the interiors. They will be less detailed, but higher resolution and should look much better. To solve the ref issue, I have purchased plastic model kits of my next 2 vehicles. It should provide a significant advantage in gaging the interior proportions and aid with modelling in general.
So I’m just over a third of the way through and so far nothing has gone horribly wrong. Since a few weeks ago, I’ve made up lost ground by putting some extra time in and most significantly, trimmed enough fat from my project to put me back ahead of schedule. Let’s hope I keep it up and things continue to not go horribly wrong.