Digital images scanned from 1960/1970s photographs printed on honeycomb (hexagonal) texture paper

Topaz Photo AI struggles to enhance digital images that have been scanned from 1960/1970s photographs printed on honeycomb (hexagonal) texture paper. In some cases the software enhances the hexagonal pattern making the “enhanced” image worse than the original. The human eye can clearly discern what the enhanced image should look like, so should an “AI” algorithm; at least the hexagonal pattern should be detected and removed.

Someone (compupix?) reported a similar problem in November 2022. You might consider giving each feedback message an unique ID number that can be used for future reference.

They have responded before to recent reports that it is only trained on digital camera noise not scanner/media noise. I would suspect it would be a big effort to train on a century of media formats and decades of scanners!

Dissapointing since they are selling to consumers wanting to restore old photo collections, yet theses are very likely to not be digital cameras. But they recently posted a request how they could further use AI so I think this category of physical media/scanners has merit to investigate.