photostream

In their letter Zeiss said of their modern 50mm Sonnar:

Because of the above mentioned classical characteristic of the lens the best focus position in the object space can not be kept exactly constant for all f-stop settings.

I wondered quite how bad this could be, so I devised a simple test. This is the test target:

A marked rule placed at an angle up a wall. Easy to focus on and easy to see, on a digital file, exactly where focus changes as the f-stop changes.

Note, though, that this isn’t a resolution test (none of these files have been sharpened, or processed in anyway to alter their characteristics). It isn’t even a focussing accuracy test, as I can’t guarantee that my focus by eye is exact. However, as the only thing changed between exposures is the f-stop (exposure on auto) then it will show up a focus drift if one exists. Distance from camera to point of focus was about a metre and a half. The ‘40′ tick was the focus target.

I repeated this test a number of times with identical results, so I’m happy that there is no user error involved. The standard pattern for the test was focus, and then expose at f1.5, f2, f2.8 and so on up in full stops. The camera was on a very rigid tripod.

This is the f1.5. Focus is so far in front of the target (in the low 50’s on the rule) that it looks like rangefinder or user error. Note this doesn’t mean 10 plus centimetres in front of target focus because of the slope on the rule.

However, as the following thumbnails show, focus marches up the rule as the f-stop in changed. In this test, focus seems to have been actually at about the 39 tick mark. By f8 depth of field makes it very difficult to be sure where focus is, but I don’t think focus drifted any further after f5.6. It may even be the case that focus is stable at f4.

Click on the thumbnails to see the files. At the end of the thumbnails there is a file that I’ve converted into monochrome. It is one of the f1.5 extracts. It shows the focus in front of the subject. It also shows an overall softness (i.e. lack of definition – but remember no sharpening here). However, to my eye, it shows a beautiful handling of the out of focus areas. Above f4 the lens works as a general purpose 50mm lens, and I’ve a number of sharp contrasty pictures to prove it. Below f4 it becomes a lens of greatly different character. It might take a lot of experience to get predictable results (focus on the eyes to get the nose in focus at f1.5), but the effort might be worth if for some subjects.

This is f2:

This is f2.8:

This is f4:

This is f5.6:

This is f8:

And this is another f1.5 extract:

Some related posts chosen by the software:

• Focus drift (Sonnar 50mm f1.5)
• Zeiss M mount lenses
• Zeiss 50mm C-Sonnar f1.5 ZM (c. 2006)