Clone Mode Reference

The street-generated-clones component places 3D models along the Z-axis (length) of a street segment. The segment extends from +length/2 to -length/2. Four placement modes are available, each designed for different use cases.

`mode: "fixed"` — Fill Every Grid Slot

Places a clone at every slot in an evenly-spaced grid along the segment. The number of clones is determined automatically: numClones = floor(length / spacing). The cycleOffset property (0–1, default 0.5) shifts the starting position as a fraction of spacing.

Models are placed at their centerpoint regardless of model dimensions. When modelsArray contains multiple entries, each clone gets a randomly selected model (using seeded RNG) — so despite the name "fixed," only the positions are fixed. Model selection is random, the same as random mode.

fixed vs random: These two modes share the same underlying grid. fixed fills every grid slot. random fills only a subset — it shuffles the same grid and takes the first count positions. Both randomly select which model to place from modelsArray. See How fixed and random relate for details.

spacing = 15m, cycleOffset = 0.5
Segment (length = 60m):
|                                                            |
+------------------------------------------------------------+
|       X             X             X             X          |
|       ^             ^             ^             ^          |
|    7.5m from     22.5m         37.5m         52.5m        |
|    start                                                   |
+------------------------------------------------------------+

X = model centerpoint
Slot size is always `spacing`, regardless of model size.
Models may overlap if model width > spacing.
Models may have gaps if model width < spacing.

Example with mixed model sizes (same fixed spacing):
+------------------------------------------------------------+
|    [=====]       [==]        [=========]       [===]       |
|       X           X              X               X         |
+------------------------------------------------------------+
     ←15m→       ←15m→         ←15m→           ←15m→
     Note: models are centered on X, gaps/overlaps vary

Properties

Property	Type	Default	Description
spacing	number	-	Distance between model centerpoints in meters
cycleOffset	number	0.5	Starting offset as a fraction of `spacing` (0–1)

Use case

Regularly spaced objects like street trees, lamp posts, bollards — objects where uniform centerpoint distance matters more than edge-to-edge tiling.

Example

{
  mode: "fixed",
  modelsArray: "tree3",
  spacing: 15
}

`mode: "random"` — Fill Some Grid Slots

Creates the same evenly-spaced grid as fixed mode, then shuffles the grid using a seeded Fisher-Yates shuffle and takes the first count positions. The result is a sparse, irregular distribution with a guaranteed minimum separation between clones.

Positions are randomly selected from the grid, so minimum distance is always >= spacing
Model dimensions are not considered — placement is by centerpoint only
Model selection from modelsArray is random per clone (same as fixed mode)

spacing = 10m, count = 3, length = 60m
Grid positions (7 slots):
|  -25   -15    -5     5    15    25    35  |
+--------------------------------------------+
|   o     o     o     o     o     o     o   |  o = grid slot
+--------------------------------------------+

After shuffle, pick first 3:
+--------------------------------------------+
|         X                 X          X     |  X = placed clone
+--------------------------------------------+
        -15                 5         25

Properties

Property	Type	Default	Description
spacing	number	-	Grid interval / minimum distance between centerpoints
count	number	-	How many clones to place from the grid
seed	number	-	Seed for deterministic random results

Use case

Scattered objects like parked cars, pedestrians, random vegetation — things that should appear irregularly distributed with guaranteed minimum separation.

Example

{
  mode: "random",
  modelsArray: "sedan-rig, box-truck-rig, self-driving-waymo-car, suv-rig, motorbike",
  spacing: 7.3,
  count: 4
}

`mode: "single"` — One Clone

Places exactly one model. Position is controlled by the justify property.

justify = "start":
+------------------------------------------------------------+
| X←padding                                                  |
+------------------------------------------------------------+

justify = "middle" (default):
+------------------------------------------------------------+
|                            X                               |
+------------------------------------------------------------+

justify = "end":
+------------------------------------------------------------+
|                                              padding→X     |
+------------------------------------------------------------+

Properties

Property	Type	Default	Description
justify	string	"middle"	Position along the segment: `start`, `middle`, or `end`
padding	number	4	Distance from the edge in meters (used with `start` and `end`)

Use case

Placing a single landmark, sign, or feature element on a segment.

Example

{
  mode: "single",
  modelsArray: "bus-stop",
  justify: "start",
  padding: 4
}

`mode: "fit"` — Dimension-Aware Tiling

The only mode that uses model dimensions. Uses a lookup table of known widths (Z-axis extent) and depths (X-axis extent) for supported model IDs. Models are tiled end-to-end along the segment with no overlaps or irregular gaps:

Start at +length/2
Look up the next model's width from the table (default 10m if unknown)
Check if it fits in the remaining space
Place it so its edge (not center) aligns with the previous model's edge + spacing gap
Advance by modelWidth + spacing
Cycle through modelsArray sequentially (index % length)

Unlike fixed and random, model selection is sequential — models cycle through modelsArray in order rather than being randomly picked.

The justifyWidth property aligns models along the X-axis (cross-street direction) using their depth:

left: model's right edge aligns with segment's left edge
right: model's left edge aligns with segment's right edge
center: uses positionX as-is

modelsArray = [A (8m wide), B (12m wide), C (6m wide)], spacing = 1m

+------------------------------------------------------------+
|[==A==] [====B====] [=C=] [==A==] [====B====] [=C=]        |
+------------------------------------------------------------+
 ← 8m →1← 12m    →1←6m→1← 8m →1← 12m    →1←6m→
         ↑ spacing gaps

Edge-to-edge tiling: each model's edges touch the gap,
not centered on uniform grid points. No overlap, no uneven gaps.

justifyWidth effect (cross-section, X-axis):
               Segment Width
         ←─────────────────────→
left:    |[Model]               |  model hugs left edge
center:  |      [Model]        |  model centered
right:   |               [Model]|  model hugs right edge

Properties

Property	Type	Default	Description
modelsArray	string	-	Comma-separated model names, cycled sequentially
spacing	number	-	Gap between model edges in meters (not centerpoint distance)
justifyWidth	string	"center"	Cross-street alignment: `left`, `center`, or `right`

Use case

Buildings, walls, fences, seawalls — anything that should tile seamlessly along a street edge with no overlaps or irregular gaps.

Examples

// Buildings tiled along a street edge
{
  mode: "fit",
  modelsArray: "SM3D_Bld_Mixed_4fl, SM3D_Bld_Mixed_5fl, SM3D_Bld_Mixed_4fl_2",
  spacing: 0,
  justifyWidth: "left",
  facing: 270
}

// Fence sections tiled end-to-end
{
  mode: "fit",
  modelsArray: "fence",
  spacing: 0
}

Mode Comparison

	Dimension-Aware?	Position Method	Model Selection	Spacing Meaning
fixed	No	Regular grid	Random*	Centerpoint distance
random	No	Shuffled grid	Random*	Min centerpoint distance
single	No	`justify` prop	First	N/A
fit	Yes	Edge-to-edge	Sequential	Gap between edges

* When modelsArray has multiple entries and RNG is active.

Common Properties (All Modes)

These properties are available across all clone modes:

Property	Type	Default	Description
mode	string	-	Clone placement mode: `fixed`, `random`, `single`, or `fit`
modelsArray	string	-	Comma-separated list of model short-names
facing	number	0	Y-axis rotation in degrees
randomFacing	boolean	false	Apply random rotation to each clone
positionX	number	0	X-axis offset from segment center
positionY	number	0	Y-axis (vertical) offset

How `fixed` and `random` Relate

These two modes are closely related — they share the same underlying grid mechanism:

	`fixed`	`random`
Slots filled	All of them	A subset (`count` out of total)
Position order	Sequential (slot 0, 1, 2...)	Shuffled (random grid positions)
Starting offset	Controlled by `cycleOffset`	Starts at `-length/2 + spacing/2` (no `cycleOffset`)
Model selection	Random from `modelsArray`	Random from `modelsArray`

If you set count equal to the total number of grid slots, random would place the same number of clones as fixed — just in shuffled positions and without cycleOffset support.

A simple way to think about it:

fixed = fill every grid slot
random = fill some grid slots (randomly chosen)

Both modes randomly select which model goes in each slot from modelsArray. Neither mode is aware of model dimensions.

Choosing the Right Mode

Street trees, lamp posts, bollards at regular intervals: use fixed
Vehicles, pedestrians, scattered objects with irregular placement: use random
A single sign, landmark, or bus stop: use single
Buildings, walls, fences, seawalls that need to tile without gaps: use fit

mode: "fixed" — Fill Every Grid Slot​

Properties​

Use case​

Example​

mode: "random" — Fill Some Grid Slots​

Properties​

Use case​

Example​

mode: "single" — One Clone​

Properties​

Use case​

Example​

mode: "fit" — Dimension-Aware Tiling​

Properties​

Use case​

Examples​

Mode Comparison​

Common Properties (All Modes)​

How fixed and random Relate​

Choosing the Right Mode​

`mode: "fixed"` — Fill Every Grid Slot

Properties

Use case

Example

`mode: "random"` — Fill Some Grid Slots

Properties

Use case

Example

`mode: "single"` — One Clone

Properties

Use case

Example

`mode: "fit"` — Dimension-Aware Tiling

Properties

Use case

Examples

Mode Comparison

Common Properties (All Modes)

How `fixed` and `random` Relate

Choosing the Right Mode