Setup

This time, I had the experience I learnt during my last game jam. So regardless of the time-constraints I had a better chance of setting up a good action plan that was realistic and would help me achieve the goals I had in mind before the deadline. I spent the first day setting up the project in Nvidia’s new branch, tweaking the various RTX settings, and above all, prepping a schedule that would help me during every day

One thing was clear to me from the beginning; I didn’t want to spend any more time that I needed doing Asset Development. So, I made use of the vast free options online - some from Megascans, some from Sketchup. This help make sure I was focusing on what mattered.

After setting up most of the characters, some tests with the environment assets I was planning to use, I knew I had a good concept to go forth. There were many details I was going to leave to the very end (post processing, fx, etc). So, I started getting on with Level 1. This stage taught me something invaluable I hadn’t anticipated. Good level design takes A TON OF TIME. Not sure the design for the game play itself (making sure it doesn’t drag, that the loop is efficient, etc) but decorating alone. While I always respected environment artists, going through the motions gave me another layer of respect. It can be so slow, but when done right, the reward is worth it. In total, I made 3 levels, I probably spent a combined total of 18 hours just making the environments look decent. But since I was going for a more realistic look (with RTX) I knew I had to spend the time.

A big concern I had during the development of the environments was performance. Regardless of RTX technology, I expected that having too many assets making up the environment, and adding a ton of different corals for the atmosphere could cause a performance bottleneck both in the CPU and GPU. Here’s where I got to play with one aspect of Unreal I hadn’t before. Foliage! A very effective tool to paint the corals around the maps knowing that they will be efficiently instanced. On top of that, combing all the environment assets to back them and force them to be instanced definitely helped big time.

The thing is, if the environment could look good, I needed to make sure everything else did too, and while for many this would not matter, I really wanted the goldfish the player plays to look good too and in particular FEEL good too. Here’s where I did possibly one of my favourite effects in the game. Using Control Rig, I made it so that regardless of what animation is being played (swimming, idle, etc), the fish’s body would lag behind any turn. Allowing for a more fluid-like movement that felt natural for a fish. I was pretty proud of how this turned out.

I knew from the beginning I was going to be using a fair bit of Niagara too. I did want to make more use of it but lacked the time. The main focus was on the one side, the plankton, which swirled around a particular spot, but also followed the player. Not only I wanted to feel like it moves nicely, but it was going to be big contributor to lighting too as each of those particles was a dynamic light on it’s own. Another good use of Niagara was the fake fog I created which leveraged Unreal’s default smoke sprite which I got playing in an effective way so that it looked like fog around the whole environment, in some parts it looked a bit too much like smoke, but I didn’t have time to address it further. And the final use of Niagara was for the Jellyfish’s lighting FX. Which consisted of beams that sparked out in electric-like shapes.