People also ask
Why do seamless patterns create moire, flicker, or banding?
Seamless patterns create moire, flicker, or banding when their repeated details approach the sampling limit of a screen, camera, printer, or resized export. The repeat itself may be mathematically correct, yet fine lines and regular spacing interfere with the destination pixel grid or halftone screen. The practical cure is to test at final size, identify the stage that introduces the artifact, and change scale, detail, contrast, filtering, or output method there rather than damaging a clean master tile.
If the seam is correct, what exactly is producing the moving pattern?
A visible seam is a discontinuity at a tile boundary; moire is an interference pattern created when two regular structures are sampled together. One structure is the artwork: stripes, dots, woven marks, grids, or repeated edges. The other may be a display pixel matrix, a camera sensor, a printer halftone, or a second pattern layered in the design. Their frequencies do not align perfectly, so the viewer sees a slower false wave, rosette, or dark band that was never drawn. Because the artifact depends on scale, it may vanish when you zoom and return at another percentage.
Flicker is the time-based version of the same problem. During scrolling, animation, or camera movement, fine repeated features land on different pixels from frame to frame. A line may be represented strongly in one frame and weakly in the next, creating shimmer. Banding can also describe broad false lanes caused by a highly regular motif rhythm, so first distinguish optical sampling from composition. Export a still at final dimensions and inspect it at one hundred percent. If the waves change dramatically with zoom, sampling is the leading suspect; if fixed lanes remain, examine motif placement too.
How can you prove whether the artifact comes from the artwork or the preview?
Create a small diagnostic set instead of repeatedly changing the original. Export the same crop at final dimensions as PNG and high-quality JPEG, then view both at one hundred percent in two independent applications. Also print a modest proof when print is the destination. If the master looks clean but one application shimmers at a fractional zoom such as 33.3 percent, its preview resampling is responsible. If every final-size export contains the same waves, the conflict is already baked into the rendered pixels and must be addressed before delivery.
Keep the comparison controlled. Use the same color profile, dimensions, sharpening state, and crop for each version. Disable browser scaling and operating-system magnification where possible, because a perfect 1200-pixel image displayed in an 817-pixel box is being resampled again. Take a screenshot only as evidence of what the interface shows, not as the production asset. Record the exact zoom, viewport, device-pixel ratio, export size, and application. Those notes turn an apparently mysterious visual defect into a reproducible sampling case that another designer or developer can verify.
Once sampling is confirmed, which feature is too fine for the destination?
Temporarily separate the pattern into value, color, and texture components. Convert a duplicate to grayscale to reveal whether alternating light and dark edges drive the effect. Blur another duplicate slightly; if the moire collapses, high-frequency detail is responsible. Inspect individual motif families and background textures on separate layers. Very thin outlines, one-pixel gaps, tightly spaced dots, and uniform hatch marks are common triggers. The overall repeat size may be generous while a small internal texture still exceeds what the output can represent reliably.
Measure rather than guess. At the planned output size, count how many destination pixels describe the narrowest line and the gap beside it. Features that receive only one or two pixels are fragile, especially after compression or movement. In print, translate the physical feature size through the intended resolution and ask the printer about screening and substrate spread. A fabric weave, uncoated paper, or soft product can merge details that survive on a monitor. The limiting feature is therefore not always the tile boundary or main motif; it is often the smallest repeated contrast transition inside them.
Should you change the repeat scale or simplify the internal detail first?
Choose the adjustment that preserves the design idea. If the motif is meant to read as a bold stripe, increasing its physical scale may be correct. If the motif size is already right but a grain overlay causes waves, simplify or randomize only that texture. Scale the complete tile proportionally; cropping or stretching one edge can break continuity. Prepare a few documented alternatives, such as 85, 100, and 120 percent repeat scale, and compare them at final output size rather than judging enlarged editor views.
Simplification does not mean making the image bland. Widen critical lines, soften extreme contrast, reduce perfectly uniform spacing, or replace a mechanical micro-grid with controlled irregularity. Preserve enough variation to describe the material or style while removing information the destination cannot carry. Make these changes in a derived production version and retain the full-detail master. One master can then support a print version, a web version, and a motion version without forcing the strictest technical limit onto every use. Label each derivative with its intended dimensions and environment.
What changes when the pattern will move, scroll, or appear in video?
Motion makes static inspection insufficient. Render the pattern at the exact delivery resolution, frame rate, camera path, and compression settings. Fine details that look stable in a still may shimmer during a slow pan because their subpixel phase changes continuously. Test several motion speeds, including the slowest movement, which can expose crawling edges. Avoid applying aggressive sharpening after resizing; sharpening restores local contrast precisely where temporal stability needs gentler transitions. If the rendering system offers appropriate mipmaps or texture filtering, verify that they remain enabled in the final pipeline.
For responsive websites, the same image may be resampled into many CSS sizes and device-pixel ratios. Provide sensible source dimensions, avoid transformations that animate scale continuously, and test common breakpoints on real screens. A CSS background set to an arbitrary percentage can create unstable fractional tile dimensions even when the source is clean. Prefer repeat sizes that resolve predictably, and allow a simpler small-screen asset when necessary. Motion safety and visual comfort matter too: a high-contrast pattern that flickers technically may also distract or trigger discomfort, so reduced-motion modes should not merely stop movement while leaving a harsh vibrating field.
How does printing introduce a different kind of interference?
Printers convert continuous tones and colors into device-specific dots, passes, or screens. A regular pattern can interact with that structure and produce rosettes or bands even when the digital file is clean. The result depends on print process, ink, material, resolution, screen angle, and color separation. A desktop proof may therefore fail to predict offset, textile, sublimation, or large-format output. Ask the production vendor for its preferred file preparation and submit a representative crop containing the finest detail and several complete repeats.
Do not attempt to outsmart an unknown press by adding random blur to the only master. Let the vendor test a production derivative and identify whether the artwork frequency conflicts with its screening. Possible responses include changing physical scale, simplifying fine detail, adjusting a texture, or using a different screening strategy available to the printer. Inspect the proof at realistic viewing distance as well as close range. A microscopic rosette may be normal and invisible in use, while a broad false band visible across the product is unacceptable. Approval criteria should reflect the finished object, not only a magnified loupe view.
When do compression and color processing make the problem look worse?
JPEG compression divides and approximates image information, which can disturb repeated edges and smooth gradients. Chroma subsampling may soften colored detail differently from luminance detail. Heavy video compression adds block boundaries and temporal prediction, so a regular texture can pulse even when the uncompressed render is stable. Compare an uncompressed or lossless reference with the delivery codec. If only the compressed version fails, increase quality, reduce fragile detail, or allocate more bitrate rather than rebuilding the repeat geometry without evidence.
Color conversions and sharpening can change the same thresholds. Converting into a smaller gamut may collapse two nearby colors into one in some regions while leaving them distinct elsewhere. Noise reduction may smear alternating marks; export sharpening may make them snap back too strongly. Apply color management once in a documented pipeline and judge after the conversion that the audience will receive. Keep a lossless intermediate so each test does not recompress the previous file. This isolates whether the false pattern originates in design, scaling, color, sharpening, or codec instead of mixing every stage together.
What final checklist prevents moire from returning in another format?
Archive the clean tile, the approved production derivative, and a repeated test field. Document repeat dimensions, target physical size, pixel dimensions, color profile, export settings, codec or print process, and the applications used for approval. Include stills at one hundred percent and a short motion proof when relevant. Test the narrowest lines, the highest-contrast texture, and any area where two regular structures overlap. A single attractive mockup is not enough because it may hide the scale that fails. The record should explain why a derivative differs from the master so future editors do not restore the removed detail accidentally.
Finally, repeat the test whenever the destination changes. A version approved for a product label is not automatically safe for a large LED wall, social video, or textile press. Start with the archived master, derive once for the new output, and follow the same sequence: identify final size, inspect at native scale, compare lossless and delivered versions, test motion or print, then approve in context. Seamlessness answers whether edges join. This checklist answers the separate question of whether the repeated information survives sampling without inventing waves, shimmer, or bands.