Announcement trailer journey: expectations and reality

Wow, time’s flying… Almost 2.5 months passed since the last devlog. And if you remember, I planned to make a Steam page with an announcement trailer till AdventureX conf. Well, I was never so naive 😅

So, the plan was to make multiple scenes showcase, uncovering the world and the main story

And the first thing to prepare was the production-ready view of the main character - Zoti.

Redesigning Zoti

New art style, new animations, new everything. The old design just wasn’t cutting it for what I wanted to show in the trailer.

To make the character feel more alive, I also added an emotions controller, based on Spine animations mixin.

If you’re not familiar with Spine, here’s the simple version.

Imagine you have one animation of Zoti walking, and another one where she just blinks. Typically, if you wanted her to blink while walking, you’d have to create a whole new animation from scratch. Same problem with emotions: walking + angry, walking + sad, walking + surprised… the number of combinations explodes very quickly.

Spine’s mixin system solves this elegantly. It lets you layer animations on top of each other. So the walk cycle drives her body and legs, while blinking or facial expressions play independently on her face. In practice, this means Zoti can walk normally while looking angry, confused, tired, or surprised — without me having to redraw every possible combination.

I like to think of it as animation Lego blocks: small, reusable pieces that snap together in different ways depending on what the scene needs.

As a result, the emotion system became much more flexible. Now Zoti can walk, run, interact with objects, and express emotions simultaneously, dynamically, rather than being baked into dozens of separate animations.

Of course, this wasn’t just Spine work - some coding was required too. But thanks to the spine-unity-runtime, hooking everything up turned out to be surprisingly painless.

The Train

The next big thing was making a train chase scene. It took a very long time to make, but I’m kind of proud of the result. Again, Spine provided a lot of help with it. Physics and transform constraints in such cases are great and easy!

Objects bouncing, wheels rolling, and even the visible part of the engine are implemented using those constraints, not just bone transformations. Technically, the engine really rolls the wheels!

The Detour

Then I got… distracted. It all started from the problem of “sticker” effect in the village scene - I made all buildings, but they looked like they were floating over the background. Manual adjustments to the shadows improved things a bit, but buildings were still not fully integrated into the scene. And the main advice I found is to hide the lower edges of the buildings behind the grass/ground.

And from an art point of view, it looked like a real challenge - drawing so much grass manually is very time-consuming.

At this point, I remembered that I am a programmer in the first place, so I thought, “Why not automate this process?”

So, I made a script that allows me to create a path and automatically place a random grass sprite from an atlas. To keep control over the result, grass is generated in the editor, not at runtime, so I can still edit/remove each sprite separately for artistic purposes. And it saves some resources in runtime.

“That’s cool,” I thought, and immediately reused the same script with another atlas to generate background trees for the village!

So far, everything has gone well - this script looks really useful, because it allows covering any edge with any kind of stuff. And the process of working on it was also so satisfying, since most of the time I did what I know best - programming.

All of this led me to the detour mistake. Somehow I came up with the idea that it will be the same or even more helpful to make… a road generator 🤯

Yes, you read it right - procedural road generator for a 2D game with pseudo-perspective support. Really, now I don’t have a clue how that happened, but I spent a couple of days working on it, and even got some results. And to be honest, I might spent much more time on that, but luckily I received a wake-up call relatively fast.

The Wake-Up Call

I already mentioned that I’m listening to and watching podcasts and other useful content on YouTube quite often. And I was fortunate to stumble upon a video from Chris Zukowski accidentally: How to NOT screw up your indie game launch. 30 mistakes in 30 minutes.

All the advice in this video is super helpful. But for me, the groundbreaking one was about trailers: “Indie game trailers should show gameplay, gameplay, and gameplay.” Not just cutscenes or pretty environments - an actual vertical slice of what playing the game feels like.

This was like a cold shower - what am I even doing? Spending dozens of hours on the road generator to skip 2-3 hours of drawing static roads? Crazy…

And this also affected the whole plan of the trailer. At the moment, all I had was kind of a cinematic walk through the scenes, and this is not actually what I want.

Real Gameplay

After a short revision, I figured out how to show gameplay alongside the world reveal. Still, I realized that one of the significant parts of the gameplay is completely missing from the trailer: puzzles. In Sly Crow, the player will not only have inventory and conversations. There will be plenty of puzzles to solve along the way. And none of them are presented in my trailer plan.

So, the next couple of days I spent designing and implementing the first real puzzle: the player needs to untie the rope at the well to use the bucket.

And again, most of the work was done in Spine: each knot and rope has its own animation, and the puzzle’s state is reflected in the way those animations mix.

Where We Are Now and what’s next

There is still plenty of work to finish all the desired scenes, including gameplay inside them. Currently, I’m working on NPCs and UI:

UI

NPCs

And here’s what I’m tackling next:

1. Update the UI - I don’t want to go into the trailer with the default Adventure Creator’s UI
2. Introduce more gameplay - each scene (excluding short cinematic placements) will have a part of actual game actions
3. Add one more puzzle - I think it will be a steam train repair puzzle. Pipes and steam!
4. Actually finish the damn trailer - No more “just one more cool system” detours.

The Steam page is still the goal. Not “someday” - this quarter. I need that page live so I can start collecting wishlists and getting honest feedback from potential players.