Halftone Dithering FX

Portraits usually fail when the grid is either too fine to read as halftone or too coarse to hold eyes, lips, and hair edges. Start around 16-24 pixels on typical web images, then move in small steps while watching the split preview. A tighter grid keeps more likeness; a wider grid gives a stronger poster or screen-print result.

For close-up faces, protect the eye line first. If the pupils and brow shape collapse into the same dot cluster, reduce grid size or lower contrast before changing algorithms. For full-body or environmental portraits, you can often push the grid larger because the overall silhouette carries more of the image.

Jarvis-Judice-Ninke and Stucki spread error farther than Floyd-Steinberg, so gradients on cheeks and backgrounds often look calmer. They are useful when you want the portrait to remain flattering while still showing a visible dither texture. Burkes can feel sharper and punchier, especially on high-contrast black-and-white images.

Ordered dithering is better when the goal is a controlled print pattern rather than photographic smoothness. It gives a more regular structure that works well for comic panels, zines, album art, and posters. If ordered dither makes skin look too mechanical, reduce contrast or increase smoothing slightly.

Contrast should define the portrait, not crush it. Increase contrast until the main features are readable, then adjust gamma to recover midtones. A small amount of smoothing can remove noisy skin artifacts, but too much smoothing can erase freckles, eyelashes, and fabric texture.

Before exporting, compare the processed output against the original with the split line across the most important feature. Export PNG for web posts, SVG when you need scalable dots for print or vector editing, and save the settings as a local preset if you expect to process a series with the same lighting.