From the Darkroom to the Algorithm: A Critical Review of Creative Concepts in Post-photographic Image Engineering

Abstract

Photographic practice has been reorganised twice within a single generation. The first reorganisation replaced chemical development with electronic sensing and file-based editing; the second replaced single-exposure capture with multi-frame estimation, learned reconstruction and, most recently, with generative synthesis that requires no optical encounter at all. Commentary on these shifts remains divided between humanities scholarship concerned with indexicality, authorship and evidential authority, and engineering scholarship concerned with reconstruction quality, perceptual metrics and detection performance. The two bodies of work rarely cite one another, and the resulting picture of what has actually changed in creative practice is fragmented. This critical narrative review examines how creative concepts inherited from darkroom and early digital practice have been displaced, preserved or redefined by computational and generative image engineering. Literature was identified through open scholarly indexes and citation searching, appraised for methodological adequacy and conceptual contribution, and synthesised thematically rather than catalogued study by study. Four arguments are developed. Manipulation is a constitutive rather than a deviant feature of photographic culture, so accounts that treat generative imagery as a rupture from an honest analogue past misdescribe the historical record. Computational capture has already relocated many creative decisions from the moment of exposure to statistical priors embedded in processing pipelines, which weakens the assumption that generative synthesis introduced an entirely new condition. Claims that generative systems are creative agents remain poorly specified, and the strongest available evidence concerns changes in the distribution of creative labour rather than machine creativity itself. Perceptual, forensic and provenance approaches to authenticity each fail in complementary ways, and no single approach supports the evidential expectations placed on photographs. Priorities for future work include longitudinal studies of practitioner decision-making, cross-domain evaluation of detection systems, and empirical assessment of provenance infrastructures under realistic distribution conditions.