How it works
Wigglegram 3D estimates relative depth and anchors a stable plane in the centre of the composition. Eight sinusoidally spaced views move nearer and farther surfaces horizontally by different amounts; intensity changes that actual travel instead of crossfading an unmoved duplicate over every frame. Each view uses a depth z-buffer so nearer surfaces win overlaps, depth-matched bilinear sampling for clean edges, and farther-side background continuation for the narrow gaps revealed beside foreground objects. Full-strength travel is capped at 2% of image width and fades to zero at the side borders. The unchanged centre view appears twice for a seamless loop, while pixel content and transparency move together in a lossless animated WebP. Very large inputs are proportionally capped to 1.75 megapixels so all eight lossless frames stay within the download transport limit.

























