They've basically finished replacing all the street lights here with LED units and I was pondering the change of light pollution (thankfully within a certain radius of observatories they've made sure that LPS lighting is the only lighting used). Given the current trend and ignoring what may happen in the future, and that right now the sky now glows with a bright white light and there's not really any chance of a filter doing anything anymore, how much detail can you actually recover out of absolute rubbish? The picture isn't intended to be artistic so please do not comment on subject, it was purely to see how much detail remained in far far far from optimal conditions (clouds, on the edge of a city facing away, no smog at all). I used a (somewhat decentered, yet actually the best copy of this lens I've ever used) Canon EF 50mm f1.4 on metabones speedbooster on E-M1, this lens is not corrected into the UV range (I could have used a 1A filter, didn't) so there is huge flaring/contamination for many stars. The exposure was 4 seconds at t1.4 (f2.0 on lens) ISO 3200. The lighting in the area was enough that the trees are actually correctly exposed (despite having no direct light on them). I was actually surprised at how much detail remained from a single unstacked exposure, ignore the massive colour shifts which I didn't bother to mask and correct for (I choose this framing out of curiosity to see what the shifts would be given the method chosen for detail recovery). I found that if I pushed any harder than this I started to get lots of banding from the E-M1 phase detect pixels (there's already magenta bands forming), since it's a single short exposure with little information to begin with colour is also lacking.