How Many Lines for 3D-Printed String Art?

There is no universal best line count for 3D-printed string art. In StringArt3D, start with the Max Lines value resolved by your chosen size and detail preset, let the preview finish, then compare one lower and one higher budget while keeping the photo, crop, size, Anchor Count, Min Gap, and Thread Thickness fixed. Keep the lowest complete preview that preserves the subject’s defining features and useful negative space at the intended viewing distance. More lines can deepen tone, but they stop helping when they mostly darken areas that are already dense.

Who this guide is for

Use the built-in photo presets as the baseline

These are StringArt3D’s current resolved photo-mode Max Lines values, not universal portrait targets. Printer fit and selected shape can affect the available physical size, while a manual Advanced override can replace the preset value.

Treat Max Lines as a budget, not a quality score

Max Lines sets how many segments the optimizer may place; it does not guarantee that every additional segment contributes equally useful portrait information.

Computational string art approximates an image with straight segments between boundary anchors. The optimizer evaluates candidate paths against the remaining image demand, chooses a useful connection, updates that demand, and repeats until the line budget is consumed. Early segments often establish the strongest shadows, outline, and facial hierarchy. Later segments work on smaller residual differences and may contribute less obvious recognition value.

That is why a higher number is not a simple resolution setting. Line count interacts with Anchor Count, Min Gap, the processed target, the selected physical size, and the virtual Thread Thickness used by the preview. Adding lines can reveal a cheek boundary or eye socket when the baseline is genuinely sparse. It can also overfill dark hair, close highlights, and make the background compete with the face when the important structure was already present.

The word “Max” matters. It is the requested drawing budget for the generation, not a certification that the source contains that much useful detail or that a printer can reproduce every crossing. Judge the completed output rather than the number beside the control. Two different portraits at the same size can reach their best readable result at different budgets because their crops, tones, backgrounds, and identity cues differ.

Start from the resolved preset

Choose printer, size, and detail first, clear any old Advanced override, and use the newly resolved Max Lines value as the baseline for that exact configuration.

StringArt3D’s built-in photo defaults currently resolve from 700 lines for Small/Low to 2,700 for XL/High. The ordinary Max Lines slider spans 500 to 4,000 in steps of 50, while direct numeric entry accepts a larger value. The larger input ceiling is flexibility, not a recommendation. Values outside the preset range require more evidence because they can increase generation and fabrication work without improving the image.

Preset values are deliberately size-aware. A line budget that looks dense in a 120 mm Small piece is not equivalent to the same number distributed across a 300 mm XL piece, and the available anchors also change with size. Never copy a total from a manual nail-and-thread tutorial or another generator without matching its anchor layout, solver, opacity or line-weight model, and physical output method.

1. Choose the supported printer and size. Physical diameter and printer fit establish the fabrication context before a solver override is considered.
2. Choose Low, Medium, or High detail. The detail preset resolves line budget together with preview resolution, virtual line weight, and related settings; it is not merely a line-count button.
3. Open Advanced Settings and check Max Lines. If an earlier manual value is still active, reset it so the new preset can establish a clean baseline.
4. Generate to completion. Do not compare a partially drawn preview with a finished one; late segments can change local tone even when the subject already appears.
5. Save the exact baseline. Record size, detail, Anchor Count, Max Lines, Min Gap, Thread Thickness, crop, image controls, and a screenshot.

Run a controlled three-preview comparison

Compare the preset baseline with one meaningfully lower and one meaningfully higher Max Lines value while every other input remains unchanged.

The purpose is not to discover a magical number by exhaustive search. It is to find the direction of useful change. If the lower preview preserves identity and opens the background, the preset may be spending lines on density you do not need. If the baseline restores eyes or silhouette that the lower preview loses, the lower budget is too lean. If the higher preview adds no recognizable feature, stop at the baseline.

Use the slider’s 50-line increment for fine adjustment only after the three-way comparison identifies a real boundary. A sequence of tiny changes encourages confirmation bias because neighboring results can look nearly identical. Start with a visible difference, locate the useful interval, and then make at most one or two smaller comparisons inside it.

1. Freeze the source and crop. Use the same pixels, zoom, rotation, inversion, brightness, contrast, and background treatment in all three runs.
2. Freeze the physical contract. Keep printer, shape, size, detail preset, Anchor Count, Min Gap, frame settings, and Thread Thickness unchanged.
3. Choose one lower budget. Make the reduction large enough to see a real density difference, then generate the entire preview.
4. Return to baseline. Regenerate or reopen the recorded preset result rather than relying on memory.
5. Choose one higher budget. Increase in a controlled step, finish generation, and avoid changing contrast to make the denser version look better.
6. Compare at equal display size. Judge all three from both close inspection and the intended wall-viewing distance.

Score identity before darkness

Choose the budget that best preserves the subject’s defining cues and tonal separation; do not reward a preview merely because it looks darker, busier, or more finished up close.

For a person, inspect the eye relationship, mouth and nose placement, hairline, jaw or profile, and the separation between face and background. For a pet, inspect ear shape, muzzle, eye placement, markings, and silhouette. For a building or object, inspect the few edges and openings that make it specific rather than generic. Write those cues down before comparing so the densest version does not win by spectacle alone.

Then inspect negative space. Highlights in the eyes, gap around the face, separation inside dark hair, space between ears, and a quiet background can be as important as printed paths. A higher line budget that fills those regions may reduce recognition even though it lowers image error in some areas. The line network needs both marks and unmarked space to create structure.

View every candidate at the same scale. A full-screen zoom favors thin local detail that may merge on the wall, while a tiny thumbnail can hide fabrication-relevant congestion. First use the intended display distance to choose the recognizable portrait, then zoom in to reject tangles, repetitive crossings, or fragile-looking local concentrations.

Know what a line-count change can and cannot fix

Increase Max Lines only for an incomplete but otherwise valid target; fix the source, crop, anchors, fabrication, or display when the first mismatch belongs to those stages.

More lines can help when the processed target clearly contains an eye, boundary, or tonal transition but the completed line preview remains too sparse to represent it. They cannot restore a feature that is blurred, clipped, hidden, cropped out, or flattened in the source. They also cannot make a busy background become a clean portrait unless the image preparation first gives the subject a stronger hierarchy.

Anchor Count controls the available boundary endpoints, while Max Lines controls how many selected connections are placed among them. If an important curve has too few useful endpoints, adding many connections may repeat the same coarse geometry. Conversely, increasing anchors without enough line budget can create more choices without enough selected paths. Change one control family at a time and record the result.

A good preview followed by a bad physical print is not evidence that the line count was wrong. Missing launches, droop, beads, fused crossings, nozzle contact, or poor backing contrast belong to fabrication or display diagnosis. Keep the digital settings frozen and compare the commanded map with the real network before spending another generation on a different budget.

Respect the detail preset’s coupled controls

Low, Medium, and High alter more than Max Lines, so use Advanced Max Lines when you want a line-only test and use the detail preset when you intentionally want the whole quality profile to change.

The current photo presets change preview resolution and virtual line weight along with the line budget. StringArt3D also adjusts the virtual line weight when a manual requested budget differs from its baseline so extra lines burn the target more slowly and reduced budgets burn it somewhat faster. This helps the optimizer use a changed budget, but it means line totals from unrelated software are especially poor comparison targets.

A revealing example is Medium size: Medium detail resolves to 1,650 lines and High to 1,700, only one slider step apart. Their output can still differ because resolution and line weight change. If you switch from Medium to High and see a better portrait, do not conclude that 50 more segments caused the entire improvement. Compare Max Lines inside one fixed detail preset to isolate that variable.

When changing size, begin again from the new size’s preset. Do not preserve a manual 2,700-line override simply because it worked on an XL portrait. The same budget on Small uses a different physical field and anchor count and may produce needless congestion. Reset, generate, and earn the override again.

Carry the preview decision into fabrication

After choosing the lowest complete preview, export that exact configuration and verify it through the supported printer path; the browser image is evidence for representation, not proof of physical success.

Every added segment becomes more motion, extrusion behavior, crossings, and opportunity for error in the physical artifact. It can also affect generated file size, computation, material use, and print duration. Those relationships are not safely reduced to one universal minutes-per-line or grams-per-line rule because frame moves, travel, printer profile, size, and the resulting path all matter. Use the generated estimate and slicer or printer evidence for the actual job.

Inspect the first real piece in stages: perimeter and anchors, early launches, middle spans, dense crossings, final arrivals, and release from the build plate. Compare the physical network with the saved preview. If the highest-density regions fuse or sag while the rest matches, return to calibration and fabrication constraints before reducing lines solely to accommodate an unresolved printer problem.

For a high-value gift or sale, qualify one representative sample before treating a setting as repeatable. A line budget proven on one clean high-contrast headshot does not certify a different multi-person photo, pet with dark fur, or XL printer configuration. Preserve the full evidence record with each approved design.

Stop at the lowest complete result

Stop increasing Max Lines when the next completed preview adds density without improving a named identity cue, tonal boundary, or normal-distance reading.

A practical keep rule is simple: the chosen preview must preserve the defining features named before the test, retain useful highlights and background separation, and remain recognizable at its intended display distance. Among previews that meet those conditions, prefer the lower budget unless the higher one adds a visible, meaningful improvement. This controls complexity without turning minimal line count into a goal of its own.

Save a short decision note: source filename, crop, printer, shape, size, detail preset, Anchor Count, Max Lines, Min Gap, Thread Thickness, processing controls, generation date, and the reason the lower and higher alternatives were rejected. Screenshots at equal scale make a future reprint or variant easier to audit than a remembered impression.

If none of the three previews is convincing, stop changing Max Lines. Return to the first failed stage in the portrait-troubleshooting workflow: source, crop and processing, line representation, fabrication, or display. A disciplined stop is more useful than a five-digit line total attached to a portrait that never had the right information.

Primary references

These sources provide technical context for the guide and its fabrication recommendations.

• TU Wien: String Art—Towards Computational Fabrication of String Images

  Primary research describing computational approximation of a target image with an optimized connected graph of straight string segments between boundary points.

• Birsak et al. StringArt research repository

  The official code repository accompanying the Eurographics 2018 paper and its discrete optimization approach.

• StringArt3D Generator

  First-party product source for the current size, detail, Max Lines, Anchor Count, Min Gap, and Thread Thickness controls described in this guide.

• StringArt3D Portrait Troubleshooting

  First-party evidence-stage workflow for deciding whether the first mismatch belongs to the source, processed target, line preview, fabrication, or display.

Continue exploring

Use these pages to test StringArt3D, prepare a stronger source image, or compare related fabrication workflows.

• String Art Generator
• Why 3D-Printed String Art Does Not Look Like the Photo
• How Long Does 3D-Printed String Art Take?
• Best Size for 3D-Printed String Art
• Calibrate a 3D Printer for String Art
• Best Anchor Count for 3D-Printed String Art
• Best Min Gap for 3D-Printed String Art
• Best Thread Thickness for 3D-Printed String Art
• Best Photos for Personalized Portrait Gifts
• How Much Filament Does 3D-Printed String Art Use?

Frequently asked questions