Now let's proceed to the famous Siemens-stars and what they will tell us with respect to the resolution-power of the Nikon 14-24mm vs. the Sigma 10-20mm.
I shot a 5-target subject on a flat, mirror-covered surface, auto-focus on the middle target, like that:
This means the following:
- If any of the lenses I test has field curvature
, that will not be distinguishable from the "natural" drop in resolution at the corners. I don't shoot focus brackets to find a point where the corners are sharpest.
- If AF fails so be it. As I'm no manual-focus
-man I have to rely on what the camera makes out of the target. That said, all lens-tests so far have been made with the same D80 body. And normally you can immediately identify one bad focus image out of a series. But should micro-focus not be spot on, that's it then. In these cases you should get a quite pronounced increase in centre-sharpness when stopping down
- If the camera is mis-aligned
(not exactly perpendicular to the target) you should see quite different resolution in the 4 corners. But the mis-alignement should hit both lenses the same way at the same focal lengths, as I didn't change the position of the tripod for a single focal length.
- If a lens has a decentering
defect, you should see quite different resolution in the 4 corners. If that sounds like above, you're right. But: normally two lenses don't have the same decentering defect. So when one lens has differing sharpness in different corners and the other lens has not (at the same focal lengths and tripod/camera position), then this lens is decentered.
- Resultant from the above two critical points I use the best of the corner
resolution results to avoid any discussions about "not an equal playing field" and comment on it. You can view all corners of all my Siemens-star shots at flickr in full resolution and make up your own mind if you care.
- These 5 targets are good for resolution testing in that the stars have a grey disk
in the middle, the diameter (and form) of which is directly (reverse) proportional to the resolution of this lens at this spot: I.e. a 30% smaller diameter of the grey disk equals 30% better resolution. All my lens-reviews are comparable as they use the same target, but my measures of the grey disk may differ from yours, as they depend on the pixel-pitch of your monitor. The measures I give are taken from a 21" 1600x1200 monitor!
- I try to avoid camera shake
as I shoot from a tripod using a remote control and flashlight. That are perhaps not the very best conditions but should be pretty stable
- The targets are at a magnification
of around 1:50. That is far away out to give you a good approximation of far-field performance.
- For the pics that I present here I've done a RAW>JPG conversion
with CaptureNX 1.3.1 at standard sharpening
(not altered from what the body was set to). Be aware though that I don't do this conversion with all my Siemens-star shots at Flickr. The other images being standard large jpegs directly off the camera. There are people saying that the RAW-conversions are sharper than the JPGs directly from the camera.((EDIT: After looking closely how the conversion from RAW in CaptureNX was different from the in-camera conversion I decided to stay with the in-camera converted JPGs. So forget about the last paragraph))
What I will present you with in the next post is a series of 100% crops linked to the original pic's homepage at Flickr. We have three comparable focal lengths: 14mm, 17/18mm and 20mm (unfortunately I didn't match the 17mm on the Sigma, but that has no influence on the reading of the grey disk)
For each focal length
I'll show you 4 pics
(1) the Nikon wide open @F/2.8 and then (2) closed to the widest aperture the Sigma could muster (F/4-F/5.6), then (3) the Sigma wide open and (4) the Sigma at F/8.
I need not show you the Nikon at F/8 as this is not necessary (you'll see). All apertures and focal lengths are at Flickr for the Nikon
and the Sigma
To complete the series I'll include the 24mm Nikon results and the 10mm Sigma results plus some comparison of the Nikon at f22 to show you how the res deteriorates due to diffraction to satisfy your every appetite.
To be continued...