Capture One isn’t perfect but its a great start and is a good comparison – it has the ability to auto correct either vertically, horizontal or both (you choose which is appropriate – I generally choose vertical as a starting point) and of course it allows manual adjustments as well. It certainly doesn’t get it right all the time (I’d say about 80%) though this is certainly a great start. Strangely it struggles a little with seascape horizons, they often need a slight tweak afterwards. So using an auto correct is a quick starting point if the model can determine the dominant verticals (or horizontals) and I find it does a reasonable job.
The ability to run through perspective correction immediately after you upload a new set of photos in one batch makes a big difference. Yes you still need to adjust a few manually but it’s a lot quicker than placing points manually on each image and moving through your ingested batch 1 photo a time.
The four points correction is used less often as, as you say Joachim, its quite a unique use case. You also mention cropping, CA1 at least lets adjust the crop and you can heal/clone into the alpha areas though these tools are a little rudimentary in CA1. I find if I really want to infill I’ll have to take the photo out to something that is a little more intelligent with its infill.
I’ll have a further play with Nitro. Thanks for your thoughts