Hi everyone,
Last time, I walked through the work happening on the Fallen Earth servers. For this development blog, I want to shift over to the other side of the project and show you what we’ve been building in the new Unity client.
A lot of the early client work isn’t immediately visible in screenshots. Before we could worry about how the game looked, we needed a reliable way to read the original content, reconstruct characters, load world areas, and get all of those pieces working together.
We’ve now reached the point where those systems are beginning to form a playable slice of Fallen Earth.
Importing the wasteland
Fallen Earth stores its content across more than 30 custom file formats. Together, they contain more than 30,000 textures, thousands of models and animations, approximately 1,000 world areas, and the data used for terrain, lighting, collision, and physics.
Unity doesn’t understand any of those formats on its own, so the team built a collection of importers to read and convert them automatically.
That process gives us a repeatable way to load Fallen Earth content directly into the new client. If something changes in the importer, we can process the content again rather than having an artist manually repair thousands of individual assets.
One of the first major complications was that the original engine and Unity use different coordinate systems. They don’t agree about which direction is forward or how objects should be oriented in 3D space.
If we applied the original data without converting it, models appeared mirrored, characters faced the wrong direction, and animations moved incorrectly.
The correction has to remain consistent across every part of an asset. A character model may look fine while the skeleton controlling it is still flipped, which produces some fairly spectacular results once an animation starts playing.
Solving that across models, animations, terrain, collision, and character skeletons was a major step in making the imported world behave correctly.
Reconstructing characters
Character appearance presented a different challenge.
The client has to take the stored appearance information for an existing character and rebuild the correct combination of head, body, hair, facial features, and customization choices.
Some of that information was stored in compact formats that required a fair amount of investigation before we could interpret it correctly. Small errors could select the wrong feature, place something incorrectly, or cause the assembled character to look nothing like the original.
The new client can now reconstruct a character from that existing appearance data and display it during character selection and inside the world.
There is still visual work to do, but seeing a familiar character appear correctly in the new client was an important milestone.
Loading into the world
The separate client systems are now working together far enough for us to complete the basic entry flow.
We can log in, display character selection, choose a character, load an imported area, and move through the environment.
Getting into the world requires more than displaying terrain. The client also needs to place the character correctly, maintain movement and camera behavior, load nearby content, and process the information it receives while the player is connected.
That flow is now functioning as the base for the gameplay work ahead.
Building the sky
We’ve also been developing the new environmental systems used above the imported terrain.
The new client has a procedural sky with a moving sun and moon, thousands of visible stars, cloud coverage, and changing weather. These elements move and change through the day instead of being presented as a static background.
Fallen Earth has always relied heavily on its environment to establish the mood of the wasteland, so getting the sky, lighting, and terrain working together has been an important visual target for the team.
The first gameplay slice
Inside the world, player movement, camera controls, targeting, and the first portion of combat are now operating together.
This gives us an early vertical slice that can be used to test the full path from selecting a character to interacting with a target inside the game.
Combat remains an active area of development. The current implementation gives us the structure needed to begin connecting and testing more of its individual behaviors.
Each additional system makes it easier to identify problems that only appear when everything is running together.
Rebuilding the interface
Fallen Earth contains roughly 50 interface screens, including inventory, maps, character information, chat, missions, trading, and guild features.
Instead of building each screen as a completely separate piece, the team created a shared interface toolkit. Common elements such as windows, buttons, tabs, lists, and item slots can be assembled and reused throughout the client.
This keeps the interface consistent and allows improvements to a shared component to carry across multiple screens.
A large portion of that framework is now in place, and individual screens are being connected as their related gameplay systems come online.
We are also testing methods for remastering the original icon artwork so it remains clear on modern high-resolution displays. I’ll save the details of that work for a future development blog.
Current progress
The new client can now complete the basic flow from login to entering an imported world area. Existing character appearance data is being reconstructed, the environmental systems are running, and the first movement, targeting, combat, and interface components are operating together.
The next phase is focused on expanding that playable slice by connecting more gameplay systems, completing additional interface screens, and continuing to improve the imported content.
There is still plenty ahead, but this is the point where the work begins becoming much easier to show. We’ll continue sharing focused development blogs as individual systems reach a state where we can properly walk through them.
Thanks,
Matt