[Steinar1]

I recommend you read this article published today by DPR. It is a synthesis of earlier articles he has written and research done in his lens rental company.



[url="http://www.dpreview.com/articles/7333489584/variation-facts-and-fallacies"]http://www.dpreview....s-and-fallacies[/url]

[Klaus]

Sorry, but this ...

"Really bad, soft, out-of-acceptable range lenses do occur. They are fairly rare though and easy to detect."

... is bizarre.



If not it would the conclusion would be that all out test cameras are out-of-alignment.

[wim]

[quote name='Klaus' timestamp='1322259213' post='13244']

Sorry, but this ...

"[color="#ffffff"]Really bad, soft, out-of-acceptable range lenses do occur. They are fairly rare though and easy to detect."[/color]

... is bizarre.



If not it would the conclusion would be that all out test cameras are out-of-alignment.

[/quote]

That actually is the correct conclusion. All cameras are, unless they are individually aligned to perfection, and no camera maker does this, other than maybe Alpa.



However, it doesn't matter too much if the size of the alignment difference is relatively small, and the number of MP isn't too large (it becomes more noticeable with higher MP counts). This is actually one of the reasons why some landscape photographers using MF backs have gone to aligning their backs individually for optimum alignment with the camera mount, or have it done (if they are separate backs it isn't too difficult to do). And the same is true for the lens mount on dslrs - some people do have it adjusted to align as perfectly as possible with the sensor, or even do it themselves, although this requires some equipment if one doesn't want to let it depend on lots of trials and errors.



Underneath most camera lens mounts there are shims with uneven thickness over their radius in order to correct for this. But it isn't perfect, as those shims are standard shims. The same is true for sensor mounts. Considering the precision required to get it really right, it is amazing that things work as well as they do <img src='http://forum.photozone.de/public/style_emoticons/<#EMO_DIR#>/biggrin.gif' class='bbc_emoticon' alt='' />.



BTW, due to the thickness of film emulsion, this generally wasn't a problem with slrs, other than making sure film was flat enough. And alignment was easier with film slrs, as one only had to align the film guides to the lens mount.



Kind regards, Wim

[Klaus]

[quote name='wim' timestamp='1322265561' post='13249']

That actually is the correct conclusion. All cameras are, unless they are individually aligned to perfection, and no camera maker does this, other than maybe Alpa.



However, it doesn't matter too much if the size of the alignment difference is relatively small, and the number of MP isn't too large (it becomes more noticeable with higher MP counts). This is actually one of the reasons why some landscape photographers using MF backs have gone to aligning their backs individually for optimum alignment with the camera mount, or have it done (if they are separate backs it isn't too difficult to do). And the same is true for the lens mount on dslrs - some people do have it adjusted to align as perfectly as possible with the sensor, or even do it themselves, although this requires some equipment if one doesn't want to let it depend on lots of trials and errors.



Underneath most camera lens mounts there are shims with uneven thickness over their radius in order to correct for this. But it isn't perfect, as those shims are standard shims. The same is true for sensor mounts. Considering the precision required to get it really right, it is amazing that things work as well as they do <img src='http://forum.photozone.de/public/style_emoticons/<#EMO_DIR#>/biggrin.gif' class='bbc_emoticon' alt='' />.



BTW, due to the thickness of film emulsion, this generally wasn't a problem with slrs, other than making sure film was flat enough. And alignment was easier with film slrs, as one only had to align the film guides to the lens mount.



Kind regards, Wim

[/quote]



There is, of course, no perfection but if 80% of the results are fine and 20% are suspicious it is also a valid conclusion that the camera is not the problem. And I wouldn't say that 20% is "rare".

The flaw in this article is the assumption that the center sharpness defines the production quality (as shown in their graphs). If they would have a look at the 4 border regions (which is, of course, a substantially higher effort) they would not have choose "rare" to describe the situation.

[Steinar1]

[quote name='Klaus' timestamp='1322259213' post='13244']

Sorry, but this ...

"[color="#ffffff"]Really bad, soft, out-of-acceptable range lenses do occur. They are fairly rare though and easy to detect."[/color]

... is bizarre.



If not it would the conclusion would be that all out test cameras are out-of-alignment.

[/quote]



That is in my opinion not his conclusions, but rather that all cameras are built within certain tolerances, as are lenses and if the deviation on both camera and lens go in the same direction, you will have more "misalignment" than the ideal specs. This is a very well known phenomenon within statistical process control.



The obvious conclusion is that one must have ones body aligned with each of the lenses in order to obtain the best results and it does not mean that either lens or body is "bad", just within tolerances.