Retouching vs compositing

Retouching fixes a frame you actually shot: dust and sensor spots, blemishes, stray hairs, a crooked label, uneven skin (frequency separation), dodge and burn, color and contrast. The photograph stays a photograph. Compositing builds one image out of several sources: a sky from a second exposure, a cut-out model dropped into a new location, a packshot assembled from a focus stack plus separate highlight, reflection, and shadow passes.

Retouch when the shoot mostly worked. Composite when the picture is physically impossible or absurdly expensive: the product on a clifftop, the whole team who were never in one room, a car parked where no car can park. Plenty of glossy hero images and reflective-product shots are assembled from several exposures rather than one frame, because glass, chrome, and liquid each want their own lighting pass.

Gotchas, in order of how much they cost. Compositing has to be planned on the shoot day: same lens, same camera height, same light direction for every element, plus a clean empty plate of the background. Composites usually read as fake because of mismatched light direction, perspective, edge quality, and grain, though sources that disagree on resolution or level of detail will give the game away too. And retouching a person has a line: cleaning up a spot is normal, reshaping a body is a claim about a human being, and some markets now require disclosure when you cross it.