Viewer unconsciously fills missing spatial information through experience — cuts across the line or discontinuous editing still read as coherent. Brain completes the picture.
Our eyes constantly fill in gaps that the camera doesn't show. You frame a character looking left — your brain immediately infers that something must be there, even if the next cut misses the point of focus or has a completely different spatial orientation. This is projective completion: the viewer's ability to piece together fragmented information into a coherent image through experience, spatial logic, and emotional context.
On set, you notice this when you cut across the axis — classically frowned upon, but it often still works because the viewer's brain integrates the spatial inconsistency. Cutting across the axis should be confusing; instead, the viewer fills in the orientation themselves. An actor sits at a table, camera from the left, next shot from the right — geometrically impossible in reality, but the viewer only sees: Two people are talking to each other. The brain works with you.
Practically, this becomes dangerous when you cut away too much. A hand entering the frame, a body part cropped — the viewer automatically completes the missing person off-screen. If this person then suddenly appears from somewhere else entirely, the completion no longer works, and this creates irritation instead of tension. Conversely: cut to a couple in conversation, both framed at the same depth of field, and the viewer doesn't need an establishing wide shot — they already know where everyone is sitting.
The relationship between depth of field, reframing, and editing plays directly into this. An out-of-focus background forces your eye onto the character in focus; your brain still coherently completes the spatial environment. Fast cuts with jumps only work because the viewer completes the movement themselves — the match cut consciously exploits this. You show less so the viewer sees more — mentally.
Don't rely on it blindly. Test with a real audience. Some jumps only work in the final cut; in the rough cut, they seem chaotic. And vice versa: too much information, too many cuts that explain everything — then projective completion no longer works, the viewer becomes passive. Give them space to complete it themselves.