Almost nobody ships pre-rendered backgrounds anymore. They were everywhere on the PS1 — Final Fantasy VII through IX, the early Resident Evils — a static 2D backdrop with real-time characters composited on top. The moment hardware could render 3D in real time, the technique was dropped as the stop-gap it had been. The usual epitaph: fixed camera, can’t change anything mid-scene, enormous to store.

Dropping them was a loss — and I’m not the only one who thinks so. There’s a real, lasting appetite for what those backdrops did; people still go back to FF7’s prerendered cities and the painted halls of early Resident Evil for the way they look. Bringing that back was deliberate, and it was part of the vision from the start — not a compromise I settled into.

What I wanted was the beauty of a pre-rendered backdrop — the composed, painterly stillness the PS1 RPGs spent their whole graphics budget on — fused with something I rarely see in tactical RPGs: a world to explore. The genre tends to run mission-to-mission — pick a battle off a map, fight it, pick the next — without a place you actually wander. I wanted the tactics to live inside a real, walkable world of towns and dungeons, with the seamless flow of Chrono Trigger carrying you between exploring and fighting.

This post is about what makes that world renderable — the pre-rendered backgrounds and the tactical grid — and why the fixed camera everyone retired is exactly what lets a painted frame and a tactical board fit together.

The fixed camera was never the weakness

The thing about a fixed camera is that the player looks at the frame. Not for the half-second a moving 3D camera grants before it pans away — for as long as they’re in that room. Resident Evil understood this and leaned into it: a set camera let the directors compose each shot for exactly the tension they wanted. The static frame isn’t a constraint you tolerate, it’s a frame you get to author.

Mine aren’t 3D scenes rendered down to 2D, which is what “pre-rendered” technically meant. They’re hand-painted. That’s more control, not less — there’s no rig, no render, no lighting bake between me and the image. I paint the exact frame I want.

The catch: I’m composing the grid at the same time

Here’s the part no general history of pre-rendered backgrounds gets to, because they’re not about tactics games: I’m not just painting a dramatic angle, I’m painting the board a tactical grid has to play on — at the same time.

I paint against a reference layer for grid placement and angle. So the composition and the playable space are authored together, in one pass: the most dramatic frame and a tactical grid that sits cleanly on it are the same decision. Drama and readability don’t get traded off, because they were never separate steps.

A real-time 3D tactical game structurally can’t do this. Its camera rotates, so the grid has to read from every angle and the scene can’t be hand-tuned to any one of them. Committing to a single frame is exactly what lets me tune both at once.

Faking depth on a flat painting

A pre-rendered backdrop is dead flat — it’s a picture. But walk around one of my scenes and it reads like a space you move into, not just across. That’s a deliberate mechanic — I call it the depth-of-field system (no relation to the camera-lens kind) — and it’s cruder than it looks.

Each scene can carry a perspective zone: a box I lay over the painted ground, its bottom edge “near,” its top edge “far,” aimed at a vanishing point on the painting’s horizon. While you’re inside it, the game measures how deep you are — a straight 0-to-1 ratio from the front of the box to the back — and uses that one number to blend two things between their near and far values:

• Scale — you shrink as you head back.
• Speed — you move slower the deeper you go (the wading-into-the-distance feel).

(The same lerp can drive camera zoom too, but I rarely reach for it.)

Then your path bends toward the reference point — the spot I set for the painting’s perspective to converge on, which isn’t necessarily straight up the screen. Head toward it, deeper into the scene, and a sideways nudge pulls your line of travel in; move away and it fans back out — the exact convergence of walking a road that narrows to a point. Stack shrink + slow + drift-toward-that-point and a static painting becomes a space with a real forward.

It isn’t real perspective math — it’s a linear map on screen height, tuned per scene. That’s the whole trick: I’m not computing a correct projection, I’m dialing the numbers to one specific painting until it feels right. Every scene gets its own box, its own near/far scale and speed, its own vanishing point — and it runs on the whole party, so trailing members shrink and tighten their spacing as the line recedes.

It’s fragile in instructive ways, too. The Godot 3→4 jump silently flipped the sign of the angle function the drift relies on, and suddenly walking up the hill out of the church square shoved the lead character left instead of right — same line of code, opposite result. (It was one of the silent renames I wrote about.)

The hard part: a grid can’t live on that depth

That depth is exactly what a tactical grid can’t tolerate — and it’s the constraint that decides, before I lay down a brushstroke, whether a scene can even have combat.

A tactical grid is a roughly flat plane of cells. For the cells to stay consistent, every point of the playable ground has to sit roughly the same distance from the camera — a high enough, angled enough view that foreshortening across the board is near-uniform. The moment a scene leans hard on perspective — ground receding toward a far vista, a big foreground — that uniformity breaks and the flat grid stops matching the painted ground.

So depth becomes a design lever, not a free choice:

• Laria’s Main Square has only a little — small enough that the ground stays roughly equidistant, so it keeps its grid and supports combat.
• The Dwarven Ruin entrance is the opposite: the depth is the whole scene, the entire frame plunging inward. There’s no honest way to lay a flat grid on that — so I made it a non-combat scene. The drama is the point, and combat would have to give it up.

When you try to force the two together, the failure surfaces at interaction time: a unit snaps to the nearest cell to reach an enemy, and if that move crosses real painted depth, the unit appears to jump across depth to a cell that reads wrong — the math is right, the picture disagrees.

So the rule I paint by: dramatic depth is reserved for set-piece, non-combat scenes; anywhere a grid has to live, the camera height keeps every point roughly equidistant and the board stays a board. Painting around the grid — knowing which scenes get to be dramatic and which have to stay honest — is most of the discipline.