I already had the problem that apps with small windows did supply HighDPI screenshots. That resulted in the screenshots being shown with a larger-than-reality size which I want to avoid. I don't think enforcing one specific scaling factor is a scalable (sorry) solution.

I think that something like a "pixel density descriptor" as in HTML (e.g. scale="1.5") would be a good solution. Not necessarily to supply multiple scaling but just to get the output of the image right.

This would probably apply to videos as well but not sure.