GLOSSARY
AI Upscaling
Enlarging images or video beyond their original resolution — from faithful 2x sharpening to generative models that invent plausible new detail.
Upscaling spans two philosophies. Interpolation-based upscalers (Real-ESRGAN lineage) enlarge faithfully — edges stay sharp, nothing is invented, the result is always safe. Generative upscalers like Magnific re-imagine the image, adding texture and micro-detail that looks like what a better camera would have captured — spectacular for art and renders, dangerous for evidence.
In this directory upscaling appears everywhere: as a finishing step after generation (every image tool), as the core product (Magnific, Topaz-class), and inside video pipelines where low-res output gets enlarged for delivery. The rule of thumb: faithful models for documents and faces you must trust, generative models for images you must impress.