Most people treat slicer software as a necessary step between their model and their print — something to configure and mostly ignore. This is a mistake. The slicer is arguably the most important piece of software in the entire 3D printing pipeline. The same physical printer, with the same filament, printing the same object, can produce results ranging from brittle failures to mechanically excellent parts — depending entirely on how the slicer is configured.
This article goes beneath the surface of slicer settings to explain what the software is actually calculating, and why those calculations matter for your prints.
The Slicing Problem: Discretizing Continuous Geometry
A 3D model exists in continuous mathematical space. A sphere is perfectly smooth in your CAD software. A 3D printer, however, can only work in discrete steps — it moves a fixed distance, extrudes a fixed amount, then moves again. The slicer's fundamental job is to translate the continuous into the discrete: to take the smooth sphere and approximate it with a finite number of horizontal layers and extrusion paths.
This discretization introduces staircase error — the stepped approximation of curves that appears as visible layer lines on the side of prints. Thinner layers reduce this error but increase print time proportionally. This trade-off — layer resolution versus time — is the first and most fundamental decision in any slicing operation.
Wall Counting and Perimeter Sequencing
For any cross-section of a model, the slicer must decide how to construct the outer shell. This is controlled by the wall count (also called perimeters or shells): the number of consecutive outlines printed before the infill begins.
A single wall is thin and weak. Three walls create a solid shell approximately 1.2 mm thick (with a 0.4 mm nozzle) that provides significant structural rigidity even with low infill inside. Beyond three walls, the returns diminish unless the part experiences significant lateral forces.
The sequencing of walls matters too. A slicer that prints the outermost wall first (outer-before-inner) produces better external surface quality because the outer wall is laid down without interference from an already-printed inner wall pushing the nozzle. Inner-before-outer sequencing can provide better dimensional accuracy in some geometric configurations.
Infill Patterns: More Than Aesthetics
Infill is the internal lattice structure that gives a printed part volume without requiring it to be solid. Different patterns optimize for different properties:
- Grid / Lines (2D patterns): Fast to print, acceptable for decorative parts. These patterns have directional weaknesses — they resist compression well in the vertical axis but can be weak under forces applied in the plane of the pattern.
- Honeycomb: A classic engineering pattern. The hexagonal geometry distributes compressive loads efficiently in multiple directions simultaneously.
- Gyroid: A triply periodic minimal surface — a mathematically continuous curve that twists through space with no flat sections. The gyroid is isotropic: it has equal strength in all directions. It also has excellent support from any angle, making it the best general-purpose infill for mechanical parts.
- Lightning: A minimal branching structure designed only to support the top surface, not to add mechanical strength. It uses the least material of any infill type and is ideal for purely visual prints where internal strength is irrelevant.
The density of infill (expressed as a percentage) determines how much of the available space inside the model is actually filled. 15% gyroid infill is the practical starting point for most functional parts; increasing beyond 40% rarely improves strength significantly compared to adding more walls.
Support Generation: The Geometry Problem
FDM printers cannot print in empty space — every line of plastic must be deposited on top of something. When a model has geometry that overhangs empty space at an angle steeper than approximately 45–50 degrees from vertical, the slicer must generate temporary support structures underneath.
Modern slicers offer two fundamental support strategies:
- Normal supports: Generated in a grid or line pattern everywhere that overhanging geometry is detected. Simple and reliable, but can be difficult to remove from complex geometries and leaves surface marks on supported faces.
- Tree supports: An organic, branching structure that grows upward from the build plate and reaches out to contact only the specific points of the model that need support. Uses significantly less material, is much easier to snap off cleanly, and causes less surface damage. The algorithm to calculate tree support paths is computationally intensive and takes longer to generate, but the quality improvement for complex organic models is substantial.
The Seam: Where Every Layer Starts and Ends
For each layer outline (perimeter), the print head must start somewhere and return to that point. The place where a path starts and stops is called the seam. At this junction, a tiny amount of extra plastic often accumulates (called a blob), and retraction (pulling the filament back slightly to prevent oozing) creates a slight gap. Both effects leave a visible mark on the surface.
The slicer's seam placement algorithm determines where this mark appears:
- Rear / Fixed: Places all seams at the same location on every layer, creating a visible vertical line running up the back of the part. Consistent and easy to hide if orientated correctly.
- Nearest: Starts each layer at the closest point to where the previous layer ended, minimizing travel time but scattering seam points across the surface.
- Aligned / Sharp corners: Places seams at corners and recesses in the model geometry, where the geometry change helps disguise the mark. The best general option for aesthetic prints.
- Random: Scatters seam locations around the perimeter pseudorandomly. Eliminates the vertical seam line at the cost of small blemishes distributed across the surface — a trade-off worth making for objects viewed from all angles.
Pressure Advance and Linear Advance
One of the most significant improvements in modern firmware and slicer interaction is Pressure Advance (Klipper/Bambu) or Linear Advance (Marlin). This feature compensates for a fundamental problem in FDM: filament compression.
Inside a hotend, the filament is being pushed against viscous molten plastic at high pressure. When the print head slows to round a corner, the pressure does not drop instantly — the compressed filament continues to push plastic out for a brief moment, causing overextrusion at corners. Conversely, when speed increases, there is a lag before pressure builds, causing underextrusion at the start of lines.
Pressure Advance solves this by calculating the expected pressure inside the hotend and pre-emptively reducing or increasing the feed rate to maintain consistent extrusion at changing speeds. The result is sharper corners, more consistent line widths, and dramatically better results at high print speeds.
Choosing a Slicer
The three dominant slicers in the current ecosystem are:
- Orca Slicer: A community fork building on Bambu Studio and PrusaSlicer. Offers the most sophisticated tuning capabilities of any open-source slicer, including built-in calibration wizards for flow rate, pressure advance, and resonance compensation. The current choice for users who want maximum control.
- PrusaSlicer / SuperSlicer: Open-source, widely supported, excellent profile library for a broad range of printers. SuperSlicer is a community fork with additional experimental features.
- Bambu Studio: The default slicer for Bambu Lab printers. Highly optimized for the Bambu ecosystem with automatic calibration and one-click printing. Less flexibility for non-Bambu hardware.
The Takeaway
Your slicer is not a black box. Every setting it exposes corresponds to a real physical decision about how molten plastic is deposited. Understanding the reasoning behind wall count, infill geometry, support strategy, and seam placement transforms slicer configuration from guesswork into informed engineering. The prints that seem inexplicably good or bad are almost always a direct reflection of slicer decisions — visible in the G-code if you know where to look.