A hybrid pipeline is the compromise most small teams end up at. Pure 2D gives the best drawn quality and costs the most per image. Pure 3D gives consistency and costs the most up front. Putting a rough 3D base underneath and painting over it takes some of each, and the interesting question is not whether to do it but where to put the boundary.

The boundary is a production decision rather than a taste one. It should be set once, written down, and applied to every shot, because a pipeline that changes per shot is not a pipeline, it is a series of separate problems that happen to share a folder structure.

What belongs in 3D

Anything that must be consistent across many images. Sets, first of all, because a built room removes continuity work permanently, as covered in the piece on set reuse. Vehicles and machines next, since they are hard perspective problems that punish freehand drawing badly. Repeated props after that.

Complicated camera moves also belong in 3D. A push through a room while the world rotates correctly is very expensive to draw and nearly free to render, and the result is usually better than a drawn approximation because the perspective stays honest through the whole move rather than only at the extremes.

What belongs in paint

Faces, almost always. A drawn face can cheat proportion per angle in ways a model cannot, and audiences read faces with far more scrutiny than anything else in a frame. Hands next, for the same reason: hands carry gesture, and modelled hands tend toward a generic correctness that reads as lifeless.

Hair usually joins them. Drawn hair is a design, not a physical simulation, and treating it as geometry produces the noisy silhouettes discussed in the notes on cloth and hair. Expression, contact, and anything where the emotional weight sits should be painted regardless of what is easy.

Where the seam shows

Line quality is the first tell. A rendered outline has consistent weight and a painted one does not, so a painted face on a rendered body produces a visible change in confidence across the neck. The fix is either to add weight variation to the render, or to paint over the outline in the transition zone.

Shadow shape is the second. Rendered shadows follow geometry exactly; painted shadows are designed. Put them next to each other and the rendered ones look fussy. Most hybrid work solves this by simplifying the render's shadows heavily before any paint happens, so the two vocabularies match before they meet.

Keeping a hybrid pipeline manageable

Write the split as a one-page document: this is modelled, this is painted, this is either. Then resist changing it. Most hybrid pipelines fail not because the technique is wrong but because the boundary drifts, and after twenty shots nobody can say what any given asset is supposed to be or who owns it.

Version control matters more here than in a single-discipline pipeline, because a painted frame is downstream of a render and a re-render invalidates the paint. Lock renders before painting starts, exactly as lighting should be locked after blocking in the blocking piece.

Render passes are the practical connective tissue. Instead of handing painters a finished image, hand them the depth pass, the line pass and a flat ambient occlusion pass as separate layers. Painters can then use the geometry as a guide without inheriting its look, masking off what they want and ignoring the rest. This is the difference between a base that helps and a base that has to be fought.

It also makes revisions survivable. When a camera angle changes, a painter working over separated passes can often re-use most of the paint, whereas a painter working over a flattened render usually starts again. That single workflow choice decides whether a hybrid pipeline can absorb notes or grinds to a halt on the second round of them.

Choosing a split that suits your team

The right boundary depends on who you have. A team of strong draughtspeople should model less and paint more; a team with one technical artist and no painters should do the reverse. Copying another studio's split without their staff is how a small team ends up with a pipeline that nobody on it is good at operating.

Whatever the split, the characters stay fictional and unmistakably adult, and the adult read has to survive both halves of the process. Generation and account controls remain on OC maker, and our working standards are set out in the editorial principles.