The town’s walls

The rig follows the watchman wherever he goes, and past the town’s walls too. At his start the view shows 64 units of Background west of the west wall, since the rig centres him and the wall is only 64 from him:

A Confiner is a rectangle the view is kept inside: a room. The rig still moves the camera; the Confiner then holds the picture so no part of it shows past the rectangle.

The Bounds are centred on the node: 2624 across from 1280 is x -32 to 2592, the west wall’s outer face to the east wall’s; 368 down from -56 is y -240, 240 above the street, to 128, the underside of the street’s stone. The view never shows past the walls, nor more than 240 above the street.

At his start the camera’s node is where the rig wants it, at his framed point; the room holds the camera 64 units east of it. Left to the rig alone, the view would show 123 units past the wall’s outer face once he stood against it. At the street’s east end he never reaches the wall: the castle door, in the gatehouse passage, takes him into the castle first. The view reaches the east wall’s outer face all the same, a view’s width before him.

The castle has had no camera. It gets the town’s, rig and all, by copy and paste: the clipboard keeps what you copy when you open another scene.

The castle’s rooms are smaller than the street, so the reference is too: 256 by 144, which a 1280 by 720 window shows at 5 screen pixels to an art pixel. The Camera, its rig and its framing are the town’s, but for the Subject; its Scaling, Extent and Width are the town’s Letterbox, unread under the Pixel Perfect as before.

Nothing holds the view yet. It is drawn from where the camera rests, at his framed point, so it shows 64 units of Background west of the hall’s wall and a strip under its floor, as chapter 4’s did west of the town’s wall. The rooms below hold it.

One scene can hold several rooms, each a Confiner. The camera is held by the room whose Bounds hold its node, and when its node walks into the next room the view slides into it over that room’s Blend. The castle has five rooms, in a row, each a rectangle of its own; they meet at their walls and never overlap.

Each room is the rectangle of its hall, walls and floor and ceiling, and each Confiner sits at its rectangle’s middle with the rectangle’s size for its Bounds. The great hall is first in the Hierarchy, so it holds the camera as the castle opens, the view stopped with the hall’s west wall at its left edge. The stair shaft is 192 wide, narrower than the view’s 256: in it the view is centred across the shaft, and shows 32 units beyond it either side.

The hall holds the view with its east edge on the hall’s, x 640, until the camera’s node crosses into the shaft. Then the view slides from the hall’s hold, its middle at x 512, to the shaft’s, at 736, over the shaft’s Blend: 0.6 seconds, 36 steps, easing in and out, the watchman in the picture all the way. The cistern’s slide and the lens room’s take 0.4 seconds, 24 steps; the beacon room’s takes 1 second, the slow one, for the room at the top.

The castle’s view is 144 tall, not the town’s 180, and the shaft is climbed by leaps. Through four leaps the view stays centred across the shaft and inside it, his feet are never within 40 of the picture’s bottom, and his body never leaves it. At the start of each leap the top of his head goes over the picture’s top, by 5.3 units at the worst: the lead is what holds the rest of him in. With no lead it goes 16.3 over.

A room is chosen by where the camera’s node is, not by where he is, and the node is where the rig has got to: a little behind him or ahead of him. Going up, it is behind: when he leaps, the node follows a moment later. So no line between two rooms could hold him in the picture across a leap up through a floor, and every way between the castle’s rooms crosses a wall instead. The way into the beacon room is a door at floor level, at the end of the lens room’s gallery, 300 units long: long enough for him to get up to his walk, so the camera is ahead of him at the wall. Crossing that door, the tightest the picture gets on his body is 6 units from its east edge.

A view that is tilted is kept inside its room by all four of its corners, not only its middle. Chapter 8 tilts the camera; to see it now, tilt the castle’s, and put it back.

The Inspector shows Rotation in degrees, two decimals. The tilted view needs more room than a level one, so the camera rests further from the wall than it would level, and the corner nearest the wall touches it.

The rig, every step, as in chapter 4: it moves the camera’s node after the watchman. Then the room holding the node holds the picture: it clamps where the camera rests so no corner of the view shows past the room, and slides it from one room to the next over a Blend. The rig follows him; the rooms only hold the picture. The Camera tab shows both: Node at, where the rig has put the node, and At rest, where the room holds the camera.