A900 Corners

[PhotoTraveler]

I'm shocked we are a few days into the Post A900 world and I haven't seen a single post on this.It's really a lens question too, but it's caused by the A900s existence.

So how is the fall off on the A900 to the corners, especialy at wide angles? We have heard Cropped Frame Pundits on their soap boxes for years that FF will have terrible corners and lenses will need to be replaced, so on and so forth. But even further, there has been all the talk about FF and SSS not being possible due to corners in the minds of these people. Now we have FF and SSS, how is it looking at the corners?

From what I see from some of David shots, things are looking great, but of course none of them look to have been test shots for this. I haven't seen a single person talk about this.

Also, how about that 16-35? We haven't heard anything on it either. Is it showing much fall off? from Davids shots it doesn't look it. Is this a lens that is every bit as good as the Nikon 14-24?

Does the new Sony body and lenses show that Olympus has barked up the wrong tree for these years?

[David Kilpatrick]

I am hesitant to comment because when you use lenses labelled DO NOT USE it's already breaking some of the groundrules for testing anything!

The sample A900 body I used shows a consistent tendency to softness at the right hand end. This affects all lenses, including the 16-35mm. I have had exactly the same thing happen with a full frame Kodak body in the past and the cause was a very slight lack of trueness to plane-parallel positioning of the sensor. Every lens projects a slightly curved field, and if the sensor is out of parallel by a tiny amount, one end of the shot will appear a little sharper than it should be, the middle will be largely unaffected, and the other end will seem to fall off more than it should. The A900 I used showed exactly this type of result, so I conclude that the sensor was not perfectly collimated.

The 16-35mm DO NOT USE sample had about the same level of distortions and CA as the 16-80mm CZ on APS-C at full frame and 16mm, but of course, the coverage was equalling the 11-18mm at 11mm - and the lens was better than that for illumination and CA. However, it shows a strong compound distortion which I experimented with curing using LensFixCI and (oddly enough) the correction for the 16-80mm at 16mm did quite a good job. The Nikon 14-24mm needs no correction; it is effectively rectlinear, as good as using a Super Angulon 75mm on 5 x 4. The 17-35mm D shows a little more distortion than the 16-35mm, and the corners fall off suddenly into softness, just like the corners of the 16-105mm at 16mm on APS-C.

Here, just to show that the 17-35mm is not a no-go area for the A900, are some small examples:

First, 16-35mm at 16mm (f5.6) straight shot:

16mmunfixed.jpg (74.05 KiB) Viewed 3872 times

Next, the same shot processed using Kekus LensFixCI and the geometry adjustment for the 16-80mm CZ APS-C lens:

16mmlensfixed.jpg (73.9 KiB) Viewed 3873 times

Next, the rather warmer and more sympathetic colour from the 17-35mm D, uncorrected, at 17mm - the viewfinder image from this lens was amazing, brighter than the CZ, or it looked it because of the warm colour rendering, maybe the sun was also brighter and bouncing off the hotel wall - I see a reflection on the buildings - so not all down to the lens:

17mmunfixed.jpg (76.4 KiB) Viewed 3872 times

Finally, applying the Kekus correction to this file also works pretty well to sort the wavy geometry. No vignetting correction was given. With a proper lens database addition to the LensFix plugin, full correction would be easy:

17mmlensfixed.jpg (67.67 KiB) Viewed 3872 times

One reason I have now ordered an A900 body is that I can see from brief tests 'regular' lenses will work pretty well. I have some stunning resolution from the CZ 135mm f2.8 in a couple of shots but the contrast of the subject I shot was extreme, and I can't show the result without access to a raw processor along ACR lines.

For regular lenses like the 28-75mm D, illumination and corner quality becomes no issue at all at longer lengths, but a smooth fall off is present at shorter settings:

28-75mm at 75mm, f8:

[email protected] (64.35 KiB) Viewed 3874 times

28-75mm at 28mm, f9. Please note - the bright sky at the top left was composed so that the 100 per cent viewfinder only JUST touched it. There is slightly more in the final shot! My thoughts about 100 per cent being difficult to achieve are probably true, and in reality, it's 100 per cent coverage but 98 per cent accuracy:

[email protected] (72.85 KiB) Viewed 3873 times

Hope this helps. Having buildings with natural vignetting (a lighter cleaned stone in the middle!) does not, of course, aid assessment.

David

[Dr. Harout]

I'm shocked we are a few days into the Post A900 world and I haven't seen a single post on this.It's really a lens question too, but it's caused by the A900s existence.

So how is the fall off on the A900 to the corners, especialy at wide angles? We have heard Cropped Frame Pundits on their soap boxes for years that FF will have terrible corners and lenses will need to be replaced, so on and so forth. But even further, there has been all the talk about FF and SSS not being possible due to corners in the minds of these people. Now we have FF and SSS, how is it looking at the corners?

From what I see from some of David shots, things are looking great, but of course none of them look to have been test shots for this. I haven't seen a single person talk about this.

Also, how about that 16-35? We haven't heard anything on it either. Is it showing much fall off? from Davids shots it doesn't look it. Is this a lens that is every bit as good as the Nikon 14-24?

Does the new Sony body and lenses show that Olympus has barked up the wrong tree for these years?

I told you I was glad to have you with us

[Greg Beetham]

Those 16 or 17mm shots sure show the extreme perspective distortion at that angle of view on FF when used outside, take a look at the size of the people over on the other side of the street (that doesn't look all that wide), they are a tiny fraction of the size of the people on the near side.

David, that darkening on the bottom LH corner in both the 17-35 and the 16-35 shots, is that the lens based vignetting, or was the scene like that naturally, a shadow or something?
Greg
btw I forgot to ask, were there any filter's involved?

[David Kilpatrick]

It is a result mainly of lens vignetting, but the subject does exaggerate it. Any wide angle like this for full frame will display this level of vignetting, regardless of sensor microlenses etc. It is just an optical fact, they displayed it on film, they still do on digital. It can be corrected in raw conversion or crudely in JPEGs. The 17-35mm is clearly worse than the 16-35mm as you would expect, but for my type of shooting, this vignetting is not a minus point. I feel it often adds to the character of shots and when I used to work with b/w in the darkroom, I shaded in the edges of my prints heavily.

David

[Winston]

Did you shoot any CRAW to get file size? I calculate that they would be something a little short of 25 MB.

[David Kilpatrick]

No, I only shot RAW, to be safe - given that some converters still can't handle A700 cRAW!

David

[Greg Beetham]

I was wondering... what if the AA filter had a slight ND effect added to the center, that gradually faded too a clear state by the time it got out towards the edge of the sensor, but then it occured to me (turkey) that'd be no good, it would effect every lens...maybe an add on filter would do the trick if there was such a one.
In any case as you intimate, it probably doesn't matter that much anyway, those sort of shots are heavily biased towards the artistic spectrum, not so much the clinical accuracy aspect, so one could say it was an intentional effect, the only problem is you are going too get it in every shot using those lenses, whether you want it or not.
Greg

[Dr. Harout]

By the way, were there any questions about the other sensor (reversed one, or whatever)? Will they be used on Alphas?
And how would they react on the corners (logically, I think that issue will be corrected by itself - or should I better keep my mouth shut )

[Greg Beetham]

Yes Doc,
That reverse sensor sounded like a real advance, but it seems to have gone all quiet on that front for the time being, I think the cost was fairly extreme tho.
Greg