Deep Siege Dev Blog

ONE LAST PUSH!

As we finish out this quarter, we’ve been scrambling to get as much into our final build as possible! Highlights include:

A completely new helm model

Some quick steering animations

Button press animations

Implemented explosion fx when the monster is hit

Fixed several sound/networking bugs

Additional Crew Quarters props

Pretty early in development, we ran into a tricky hurdle... we only had one character model for a 3-player game! This meant that in old builds, there would be three identical sailors running around. When more than one of them was on screen, it would instantly break immersion and just look generally silly.

However, given our time, resources, and manpower, making additional character models simply wasn’t going to be feasible. So we had to come up with something a bit more creative...

Music is one of the easiest ways to make or break the tone of a game. For Deep Siege, we needed a soundtrack that could be directly influenced by the phases of gameplay.

To do this, I wrote four looping stems that can fade in and out in-game. 

(Listen to what it ends up sounding like here!)

The stems are basically separate layers of the same track... they’re all exactly the same length and loop seamlessly. Once I had those, I hooked them all into a Sound Cue in UE4.

Just plugging the stems into a Mixer node would make them all play at the same time, but that’s only part of what we needed. The Modulator nodes are what hold custom parameters that can control their volumes in Blueprint.

Here you can see where the parameters are set up. I named them and locked their values between 0 and 1.

(NOTE: I later realized that setting the Modulator to 0 is actually optimized to stop that loop, which ruins the sync effect I wanted. Instead, I set the value to a number like 0.00001, so the loop would keep playing but remain inaudible.)

In Blueprint, a Music Intensity variable set which stems should play at any given moment. Ths number is determined by a combination of sub damage and the monster’s remaining HP.

Notice that this is only setting “Target” volumes. The actual volume amounts are interpolated elsewhere for a smooth fade between levels.

The end result is a much more compelling track that responds to players’ progress and builds toward the climax of the game.

One of the biggest design changes we’ve implemented so far is the removal of the scanning mechanic. Today we’ll talk about what it was, why we got rid of it, and the impact that has had on the game...

So what was it?

Originally, the scanner was the sole means of locating the monster. Players would climb inside it and scan back and forth for the light indication light. In practice, it looked something like this:

See the little SW up there? Both the scanner and the helm had compass bearings, so once the player in the scanner found the monster, they could tell the player at the helm where to steer.

We hoped this would enforce Deep Siege’s core of teamwork and communication, and it was also our sneaky way of not having to model and animate the actual monster. But in testing, it fell flat.

What went wrong?

What we thought was a layer of complexity turned out to just be a layer of confusion. Players struggled when first using the scanner, and even once they got the hang of it, it just wasn’t that fun. It also made life very boring for the player at the helm: they couldn’t really see anything, so they essentially were just holding down a button when someone else told them to.

Since no one wanted to do this, what many testers did was assign one person to jump back and forth between the scanner and the helm, which was tedious, awkward, and created a lot of guesswork...

So we got rid of it.

All navigation is now done at the helm. We opened up the front window and added a moving ground plane and an actual monster for players to look at (more on that later!).

It’s now much more intuitive for new players to use, and feedback is much more obvious!

Ripple Effects

While Deep Siege was now much less confusing, it was also disgustingly easy to win. The three players now only had three core tasks instead of four (steering, shooting, repairing), so there was no longer any incentive to swap roles mid-game.

Here’s some gameplay footage of our latest build with commentary!

Over the next few days, We’ll be rolling out some posts breaking down key features, how we got to this point, and where we still plan to take it.