[stoppingdown]

Ok, sunny day so I ran a test with the proper procedure. There is clearly a problem, even though is less severe than it seemed after my first quick test. Below a link to a comparison image with the borders a 23mm ƒ/8. Top row: test ran 9 years ago when I bought the lens; bottom row: test ran today. Same post-processing. Beyond the 16 vs 24 megapixels there is a clear difference in contrast due to the light (the old test was run in a crystal clear day of January, today it's hazy; there could be a slight impact of warm air blur, but not so much (20°C). I tried to compensate a bit the colour temperature to minimise the difference.

While the right border might be at the same level (I think most of what appears better in the old test is contrast, which is due to the different light), the left border has clearly a problem. It's less severe than I thought a few days ago because at first I concentrated to the dome only — which is terrible —, while the problem seems to happen abruptly precisely in the area of the dome (compare the two red stairs in the buildings behind the church: the right stair is considerably better). I must add that 23mm looks like it's the worst focal, so it might be impacted by a slight problem of focusing; but the sudden drop clearly is not related to a focusing problem. It is visible from 50mm down to wide.

Thoughts?

https://freeimage.host/i/Hgt8oRn

---

Here a second comparison screenshot including the center (the first image of each row) and where I added 30% more contrast to the new photos to compensate a bit the difference in light.

https://freeimage.host/i/HgtsQmN