That pretty well sums it up nicely.If it isillegible it is illegible. So obviously if it produced accurate alphanumerics there was enough data to sharpen and interperet the image correctly. And extremely convincing is different than real.
Even if the subject is an everyday object -> if it has an abnormality that is not recorded by the sensor, then it wont be shown on an AI image either.
In some cases it might even be remowed, even if it is captured by the sensor.
And don't get me wrong I tottaly understand that for most cases AI generated detail is good enough. But there are also plenty of cases where you might want the real detail.
I shoot a lot of commercial stuff for ads and usualy I don't care if those images are AI enhanced or not.
But I'm also working on some editorial stuff and some personal projects. Last week I photographed a 92 year old farmer/hunter in front of his house with his fathers rifle in his hand and some tools and trophies up on the wooden walls of the house. There were plenty of details that are interesting. Some are visible some not. But in this case I would rather choose a higher MP camera to show the hand made engravings on his rifle, the details on his face, engravings on trophies, details on old tools, then have AI make up "extremely convincing" ones.
And about MP needs.... even when the AI generated detail is convincing enough or when you just need to sharpen or to clean the image with the help of AI, there are cases when you might have a need for a high MP camera. Otherwise you might just want to tell the AI to create everything from scratch.
Upvote
0