Conwy Castle is not the easiest of subjects. You probably have the only viable position here for an overall shot and maybe a tighter crop will be the answer.

Along with waiting for the light of course.

For this technique to work you need to use a manual focus lens but set the camera on autofocus.

A curious analogy. As far as I know the only thing you can do with the ingredients mentioned is bake a cake. The only things you can do with a cake are store them until hard and disgusting (my grandmother useed to do this witrh Date and Walnut cake. I detest Date and Walnut cake) or eat them.

Not quite sure how this relates to photography!

You can select the time displayed in your profile preferences.

You could change to JPEG and the Bright setting and also make sure you shoot in sRGB.

All these things should brighten things up.

Wide angle lenses are more difficult to focus accurately anyway, either by eye or by AF systems. However, DOF helps us here, especially when most wide angle shots are at small apertures.

All this will be most critical for f1.2 lenses used wide open at short distances....

Here's a long shot.

Point the camera at a blank piece of sky or similar and then adjust the viewfinder dioptre until the guide lines are as sharp as possible. It doesn't matter where you focus the lens for this, you are focusing your eye on the focusing screen itself.

Having set that, go back to the problem and try the manual focusing again.

This will eliminate any mis-adjustment of the camera optics for your eyesight.

If there is still a problem, calling in to a friendly dealer and trying your lens on another camera body might be a way forward.

For something 15-20 feet away and of that size focus should just snap in and be spot on. You don't need a tiny spot for that.

The first picture is very soft and even at f1.4 and even on a monitor it looks as though it should be much sharper than that. f1.4 should be perfectly useable for images, just not the ultra-sharp result that say f5.6 would offer.

The only answer may be to try another lens and compare results.

Would that then mean they would be criticised for cherry-picking the best for testing, perhaps even a specially prepared lens? And could that then result in unreasinable expectations when everything performed well above average?

Imperfect though it is I think I would still prefer a random sample. There's not really time to test more than one sample, although the recent AP test of the Pentax and Schneider versions of the 12-24mm and 50-200mm lenses gave a rare opportunity for this to happen.

Everyone can judge the sharpness of the images for themselves, but don't forget that one sample of a lens may or may not be indicative of the general performance. This is one of the problems with magazine tests - they only have one sample.

Having said that, all four 18-55mm lenses we have had have all performed exactly the same.