True Shape Nesting vs. Rectangular Nesting: What's the Real Difference in Material Savings?
A practical comparison of true-shape and rectangular nesting with real utilization benchmarks. When each approach makes sense for laser, plasma, and CNC cutting shops.
When I was designing Lapas’s nesting engine, one of the first decisions was whether to support true-shape nesting or stick with rectangular (bounding-box) nesting. Rectangular nesting is simpler to implement and faster to compute. True-shape nesting is harder to build but recovers significantly more material on irregular parts.
For shops cutting simple squares and rectangles, the answer doesn’t matter much. For shops cutting brackets, gussets, flanges, gaskets, or any part with an irregular outline, it matters a lot. This post covers what the difference is, what the utilization gap looks like in practice, and when rectangular nesting is actually sufficient.
What each approach does
Rectangular nesting treats every part as its bounding box — the smallest axis-aligned rectangle that fully contains the part. The optimizer packs these bounding boxes on the sheet. Two parts that both fit within a 100×50 mm rectangle are treated as identical rectangles regardless of their actual shape.
True-shape nesting uses the actual part outline. An L-shaped bracket is packed as an L, not as the rectangle that surrounds it. The optimizer can slide parts into the concavities of other parts, rotate them freely, and find arrangements that rectangular nesting structurally cannot achieve.
The difference is most visible on parts with significant concavities, irregular outlines, or high aspect ratios. It’s least visible on parts that are already close to rectangular.
Utilization comparison by part type
| Part type | Rectangular nesting | True-shape nesting | Typical gap |
|---|---|---|---|
| Rectangles, squares | 85–92% | 87–93% | 2–5% |
| Simple L-shapes, angles | 72–80% | 84–90% | 10–12% |
| Complex brackets, gussets | 58–68% | 80–88% | 18–24% |
| Organic / freeform shapes | 45–60% | 74–85% | 25–30% |
| Mixed jobs (variety of shapes) | 65–74% | 82–91% | 15–20% |
These ranges are consistent with published research on 2D bin packing (Bennell & Oliveira, 2008, European Journal of Operational Research) and with real-world data from nesting software vendors.
For a shop cutting purely rectangular stock (flat bars, standard plate sections), rectangular nesting is adequate. For any shop cutting irregular parts — which is most laser and plasma shops — true-shape nesting recovers material that rectangular nesting structurally leaves on the table.
A concrete example
Consider an L-shaped bracket: 120 mm wide, 100 mm tall, with a 60×50 mm notch cut from one corner. The bounding box is 120×100 mm.
With rectangular nesting, each bracket occupies 12,000 mm² of sheet area. The actual part area is 7,800 mm² (12,000 minus the 4,200 mm² notch). The nesting efficiency per part is 65% — 35% of the rectangle each part occupies is empty space.
With true-shape nesting, the optimizer can interlock two brackets — rotating one 180° so its notch faces the notch of its neighbour. Two brackets that interlock this way occupy roughly 14,800 mm² instead of 24,000 mm² (two bounding boxes). Efficiency per part pair: 87%.
That gap — 65% vs 87% — is 22 percentage points on a single part type. Across a full job with 100 brackets, it’s the difference between needing 3 sheets and needing 2.4 sheets.
When rectangular nesting is sufficient
There are real scenarios where rectangular nesting is adequate or preferable:
Parts with no significant concavities. If your parts are mostly rectangles, trapezoids, or convex shapes with minor chamfers, the gap between rectangular and true-shape nesting is small. The added computation time of true-shape nesting may not be worth it.
Grain direction constraints. On anisotropic materials (grain-direction-critical aluminium, directional rolled steel), parts often can’t be freely rotated. If all parts must run in the same orientation, the interlocking advantage of true-shape nesting diminishes. Rectangular nesting with fixed orientation may be within a few percentage points of true-shape nesting.
Very simple parts at high volume. A shop cutting nothing but 200×100 mm rectangles all day doesn’t need true-shape nesting. Rectangular packing gets to 90%+ on uniform parts.
Prototyping and one-off jobs. If you’re cutting a single prototype part and wasting the rest of the sheet regardless, the nesting method doesn’t matter.
Outside these cases, true-shape nesting is the right choice for any job where material cost is a meaningful factor.
The rotation question
True-shape nesting’s advantage over rectangular nesting depends substantially on rotation freedom. More rotation options mean more packing configurations, which means higher utilization.
Free rotation (any angle) produces the best utilization. For irregular organic shapes, the difference between 4-angle rotation and free rotation can be 5–10 percentage points.
Fixed angles (0°, 90°, 180°, 270°) is the most common practical setting. It respects grain direction and material properties while still allowing the interlocking that rectangular nesting can’t achieve.
No rotation eliminates most of the interlocking advantage. True-shape nesting without rotation still avoids the bounding-box penalty, but the utilization gains are smaller.
For most laser and plasma jobs with mild steel or acrylic, free rotation is the right default. For structural steel plate with directional mechanical properties, fixed-angle true-shape nesting is a reasonable compromise.
How to check if true-shape nesting would help your specific job
Run the same job twice — once with rectangular nesting (if your tool supports it as an option) and once with true-shape nesting. Compare the sheet count and utilization percentage.
If you don’t have a tool that supports both modes for comparison, the quick estimate: look at your parts and calculate the ratio of actual part area to bounding box area. If that ratio is below 75% (your parts are less than 75% of their bounding boxes), true-shape nesting will materially help. If it’s above 90%, you’re probably fine with rectangular.
In Lapas, you can see the utilization percentage after every nest. If you’re consistently at 85%+ on irregular parts, the nesting is working well. If you’re below 80% on parts you’d expect to pack tightly, check whether rotation is enabled — restricted rotation is the most common cause of underperformance in true-shape nesting.
FAQ
Does true-shape nesting take longer?
Yes — finding optimal placements for arbitrary part shapes is computationally harder than packing rectangles. A fast heuristic true-shape nest typically takes 5–30 seconds for most jobs. A deep metaheuristic optimizer may run for 1–3 minutes on complex jobs. Rectangular nesting finishes in under a second.
For production jobs on expensive material, the extra seconds or minutes are worth it. For quick quoting or checking whether a job fits on stock, the fast engine is adequate.
Can true-shape nesting hurt utilization in any case?
Rarely, but it can happen when parts have features that prevent interlocking — for example, parts with large convex curves that can’t nest efficiently regardless of rotation. In practice, true-shape nesting never performs significantly worse than rectangular nesting; the worst case is roughly equivalent utilization with more computation time.
What about nesting parts inside the holes of other parts?
Part-in-part nesting — placing small parts inside the cutouts of larger ones — is an extension of true-shape nesting that some tools support. If you have a large plate with a 150 mm circular hole, a small disc that fits inside that hole can be nested there instead of consuming separate sheet area. This can recover meaningful material on jobs with large cutouts.
I use LightBurn — does it do true-shape nesting?
LightBurn’s built-in nesting is rectangular (bounding-box). For true-shape nesting with LightBurn output, the typical workflow is: nest in a dedicated tool like Lapas, export the nested layout as DXF, import into LightBurn to apply toolpaths and send to the machine.