14 min lesson
Coordinate Spaces / Transforms
- Trace a point through model, world, view, clip, and screen spaces.
- Connect matrix transforms to translation, rotation, scale, and projection.
- Spot common transform-order mistakes before they become visual bugs.
Coordinate spaces are named promises
The same point can be local to an object, global to a world, or flattened onto the screen.
1. A point changes meaning by space.
In model space, a vertex belongs to one mesh. In world space, that mesh has a place among every other object. In view space, the camera becomes the origin that the scene is measured from.
2. Matrices are ordered adapters.
Scale, rotate, translate, and project are not interchangeable. The order determines which coordinate system each operation speaks, so swapping two transforms can move an object somewhere surprising.
clip = projection × view × model × vertex
3. Debug one handoff at a time.
When geometry appears mirrored, tiny, or behind the camera, isolate the last known good space. Draw axes, freeze the camera, and verify each handoff before blaming the shader.
AI 코치에게 질문하기
현재 레슨 범위 안에서만 짧고 정확하게 답변해요.
AI 연습 문제
레슨 맥락으로 문제를 만들고, 루브릭 기준으로 바로 채점해요.
문제 만들기를 누르면 현재 레슨에 맞춘 주관식 문제가 생성됩니다.
학습 완료
문제 풀이까지 마쳤다면 레슨을 완료로 저장하세요.