HOME / TIPS / Photo Prep
Photo Prep

Dithering for Laser Engraving — Which Algorithm to Use

2 MIN READ
VERIFIED OPERATOR GUIDE

Dithering converts a grayscale photo into a binary (black/white) pattern that simulates continuous tone when engraved. The algorithm you choose dramatically affects the result — especially on different materials. Reticle Red's Photo Prep tool includes 15 algorithms across four families — error diffusion, halftone screens, ordered (Bayer) grids, and plain threshold. The five below are the workhorses to learn first.

Floyd-Steinberg The standard. Distributes quantization error to neighboring pixels, producing a natural grain pattern. Works well on: wood, leather, MDF. Avoid on: acrylic (error diffusion shows as blotchy).

*Best for: portraits on wood, general photo engraving.*

Stucki Similar to Floyd-Steinberg but spreads error over a larger neighborhood, producing a smoother result with finer grain. Slightly more processing time.

*Best for: fine-detail portraits, high-DPI engravings on maple or cherry.*

Jarvis-Judice-Ninke The largest error-diffusion kernel shown here. Produces the smoothest gradients but can lose fine edge detail. Good on large-format pieces where you're viewing from a distance.

*Best for: large wooden signs, slate tiles.*

Bayer (Ordered Dithering) A grid-based pattern that creates a regular halftone look. No error diffusion — deterministic output. Produces a "screen-printed" aesthetic rather than photorealistic.

*Best for: graphic/pop-art style portraits, logos with gradient fills, anodized aluminum.*

Atkinson Developed for early Mac printing. Spreads only 75% of the error, causing it to preserve highlights and drop shadows — creating a high-contrast, punchy look.

*Best for: dark materials (slate, dark walnut), images with strong contrast, where shadow detail is less important than bright areas.*

Quick reference: - Natural portrait on wood → Floyd-Steinberg at 254 DPI - Sharp detail, light wood → Stucki at 300 DPI - Graphic/stylized look → Bayer at 200–254 DPI - Dark slate or stone → Atkinson at 254 DPI - Large format sign → Jarvis at 150–200 DPI

Always run a 50×50mm test patch before committing a full piece.

Frequently Asked Questions

What is the best dithering algorithm for laser engraving photos on wood?

Floyd-Steinberg is the best general-purpose dithering algorithm for photo engraving on wood. It distributes quantization error to neighboring pixels, producing a natural grain pattern that works well for portraits on basswood, maple, and MDF. For fine-detail portraits on light hardwoods, Stucki produces slightly smoother results with finer grain. For dark woods like walnut or slate, Atkinson dithering preserves highlights better by only spreading 75% of the error, creating a high-contrast, punchy look.

What is Floyd-Steinberg dithering and when should I use it for laser engraving?

Floyd-Steinberg is an error-diffusion dithering algorithm that converts grayscale photos to binary black-and-white by spreading color rounding errors to adjacent pixels. The result simulates continuous tone gradients using a dot pattern that reads as smooth when engraved. Use Floyd-Steinberg for: portraits on wood at 254 DPI, general photo engraving on most natural materials, and any project where photorealistic rendering is more important than graphic style. Avoid it on acrylic — the error diffusion can appear blotchy on hard, reflective materials.

Should I use Bayer or Atkinson dithering for dark materials like slate?

Use Atkinson dithering for dark materials like slate, dark walnut, or dark stone. Atkinson only spreads 75% of the quantization error (instead of 100% like Floyd-Steinberg), which causes it to preserve bright highlight areas and drop shadow detail — giving a high-contrast look that reads well on materials where the engraved area is lighter than the base. Bayer (ordered dithering) produces a regular halftone grid pattern — use it for graphic or pop-art style portraits, logos with gradient fills, and anodized aluminum where the structured pattern is aesthetically desirable.

Privacy Policy| Terms and Refund Policy| Contact and Support