[stevebrot]

A much better test is for a well-defined single object in good lighting, well-exposed, and well-focused.

...and not at huge distance. The article on this site recommends 30m to 100m using good light and a subject with defined and obvious texture.

https://www.pentaxforums.com/articles/photo-articles/how-to-check-your-lens-...centering.html


Steve

---

Last edited by stevebrot; 05-16-2019 at 08:32 PM.

[Pikselisiirto]

Those photos have been focused to the infinity with live view contrast AF at the center. All subjects should be in perfect focus because they are very far like two kilometers or so. No refocusing between shots. I used more cloudy weather to reduce CA, so that sharpness is more clear. Infinity focus makes sure that camera reposition does not change focus range at all. I use infinity focus a lot, so if there is problems, closer focus range does not matter. This lens should be excellent in all focus ranges. I'm starting to think that this lens produce normal sharpness only at the center and top middle edge. Then it would be excellent, if top middle edge sharpness happens in all edges.

I have already sent the lens to the Sweden for checking. If they think that this is normal Pentax peformance, I probably sell the lens and the camera. These quality issues are just ridiculous because I already sent first lens and got decentered one back. Then the camera also has obvious shutter shock issues starting at 1/250s and slower. That one I can avoid with ES, but it makes IBIS useless. My older Nikon D3200 camera had zero problems.

[rogerstg]

...and not at huge distance. The article on this site recommends 30m to 100m using good light and a subject with defined and obvious texture.

How to Check Your Lens for Decentering - Articles and Tips | PentaxForums.com

Steve

And, the final paragraph "Be Realistic" bears mentioning. IMO, the article's "good" lens seems to test worse than the OP's though it's not easy to tell since the OP's are not very high contrast scenes.

[beholder3]

No. 1: "sharpness" and "resolution" are effectively 100% contrast measurements. So using any poor contrast / low light scene is nonsense, to put it mildly.

No. 2: the article calls for 30-100 m distances because it is far away to avoid DoF issues in most cases but it is near enough to also limit seeing issues. Take 10 images of any long distance subject and atmospheric movements will change the results on this pixel-level more than any lens will.

Major user skill issue.

[stevebrot]

All subjects should be in perfect focus because they are very far like two kilometers or so.

Evaluating texture at two kilometer distance* with a 50mm lens, even an excellent 50mm lens on a high resolution sensor is difficult. The examples you shared are not capable of showing whether you have a problem or not. If there is a problem with your lens, I hope that Pentax is able to make it good for you. If not, I wish the best with the sale of your Pentax kit and finding better performance quality with a competing product.


Steve

* Distance helps with this test because there is less problem with parallax. Too much distance makes accurate focus much more difficult due to limited focus throw; it is easier to focus to moderate distance. There is also the issue with atmospheric haze and temperature gradients. 30 to 50 meters would have been quite adequate.

[stevebrot]

Lens Rentals Blog | Testing for a Decentered Lens: an Old Technique Gets a Makeover

Good discussion.


Steve

[Pikselisiirto]

Are you guys really looking test image at full resolution which is 1458x1458? There is nine photos and each is 486x486 at 100% crop pixel level. I did this so that the forum would not resize the image, but it still does that, if I'm logged in. Must be my personal settings. Image can be opened in separated tab to view at 100% then.

For me, varying softness is obvious. If I label each photo like upper row is 1,2,3, middle row is 4,5,6 and botto row is 7,8,9, then I see that only 2 and 5 are excellent. 7 is worst, then 9 and 8

I am full aware about how weather conditions can affect sharpness. These same edges always are worst, no matter how many times I shoot and no matter what conditions. It gets worse when opening aperture more. At F5.6 it is quite minor problem and at F8 pretty much gone. That would be expected from a lens which would cost like a hundred bucks or less. Not from a lens which costs a grand.

If excellent sharpness can not be obtained at infinity in every edge but only some, what is the point testing in closer focus? The lens is unacceptable anyway then. Haze would be visible in every part of the lens. It should be obvious. I have shots at F1.4, F2.0, F4.0, F2.8 from this same place. Same problem edges everytime.

And BTW, the first copy did not have this problem, when I checked couple old photos. It had cosmetic issues like small white spots inside the lens barrel and importer agreed to give another copy when I sent photos from those spots. It feels like I should had kept that even if those white spots look annoying and might produce a bit flare in extreme situations. It just leaves bad taste when you spend that much money and then see flaws, first cosmetic and now optical flaws.

[clackers]

. All subjects should be in perfect focus because they are very far like two kilometers or so. No refocusing between shots. I used more cloudy weather to reduce CA, so that sharpness is more clear.

?

This is not how you test a lens.

[stevebrot]

?

This is not how you test a lens.

...and not how one attains "perfect focus".


Steve

[mikesbike]

Believe it or not, infinity focus is not really accurate focus, even with the best of lenses. Some will actually focus slightly beyond infinity. When you get into that territory, a dependency on DOF is often in play. I do not know why you refuse to run an alternate and more accurate test method, which is simple, and is recommended by experts for all brands of lenses.

I do not have this particular lens, nor am I interested in it. While it is nice to have good sharpness wide open at f/1.4, it is not a focal length I use all that much. However, I do have numerous other Pentax lenses, including some DA* models, all of which came out fine when checked for decentering. It does happen across all brands, as evidenced in testing labs, but I have been fortunate in my purchases.

[MJKoski]

Lens with high micro-contrast resolves details just fine even when the light is flat and dull. Infinity focus is just as suitable for decentering test as any other distance if the lens is capable of resolving clear details wide open.

What I do is focus lens wide open to very extreme corner where a clear object exist. Then I compose the same object to other three corners. Instant results with a decentered lens. Easy as cake.

[mikesbike]

Of course. You are achieving good focus on a single object, then composing that object into different areas of the frame to compare resulting clarity. It is far better to use a single, well defined object and reposition the camera. Nothing wrong with your method of doing this, that I can determine.

Even using a flat brick wall is fallible, as the bricks are not identical, and you'd need to make sure of being at exactly right angles to the wall. Otherwise, the right section of the wall right be at a slightly different distance from the left section.

Yes, a good lens is capable of resolving fine detail when lighting is low, flat and dull, but accurate focus can be more difficult, and results are not as easily visible.

As to accurately focus by slamming the lens all the way to infinity, assuming focus will be accurate due to the distance of the scene, to me is taking accurate focus for granted rather than achieving specific focus.

[stevebrot]

Lens with high micro-contrast resolves details just fine even when the light is flat and dull. Infinity focus is just as suitable for decentering test as any other distance if the lens is capable of resolving clear details wide open.

Two kilometers of hazy air?

How many seconds of arc would one expect the sensor to resolve at 50mm focal length and that distance even without a cloud of particulates? FWIW, part of doing tests of this sort is to make it easy for the system in general to do its work such that true optical problems are easily distinguished from problems with the subject.

As I was typing that last, I was remembering another thread a few years ago where the OP was quite adamant that they must evaluate their problematic DA* 50-135/2.8 at infinity and were frustrated because they could not get the AF system to lock on objects several kilometers away. As it turned out, the lens had gone quite wrong and was unable to render sharp at any distance. The PDAF lapse was a symptom of a problem that was obvious at 20 meters.


Steve

[Pikselisiirto]

Got the lens back from the service. Took almost a month. First it went to Sweden and they sent it to Germany. Service report reads that they did focus point adjusment and optical check. Now those soft edges are much better. Still it feels like that my copy is not quite 100% what it should be because top edges have more sharpness than bottom edges. I tested it with F2.8 for infinity focus sharpness. Not really using that much those wide apertures for infinity focus though and closer focus distances do not produce that obvious differences what I have seen. Center area is really sharp and there it seems to perform 100% and top middle edge is quite close. That is probably the intended quality for all edges.


A bit shame that premium lenses can have sloppy quality control. What I learnt from this that I do not trust anything what manufacturers say about premium quality control. Finnish shops have 14 day return policy for web purchases, so I will do these tests first no matter what brand or price it is.

Here is some comparison shots. Right top image is the worst edge before service which is left bottom edge. Left image is the same after service. Left bottom image is the best edge which is top middle edge. Right bottom image is right top edge which shows how much better that corner performs compared to worst corner. These are 100% crops.