XF 16-55mm - Do I have a bad copy?

canetsbe wrote:

Here are a couple more photos, one manual focused using 100% magnification on the camera LCD and focus peaking, vs AF. You will find that both methods are yielding the same result.

AF

MF

Also some closer shots, also yielding same result regardless of AF or MF. This one I compared f/2.8 to f/8.

f/8

f/2.8

A couple of questions...

Where is the focus point in the landscape shots? They still look crappy to me.

Are these SOOC jpegs or RAW conversions?

If they are from RAWs, what software are you using, and what processing has been applied?

canetsbe wrote:

Joachim Gerstl wrote:

No, you most likely have no bad copy. Don't send it to Fuji yet before you give it another try.

Try to do focus manually using the loupe or simply use the AF on a subject which is about 5 to 10 meters away. To crank the lens to infinity doesn't work. Many lenses focus behind infinity. There is no way to get such a crappy image quality without a serious user error.

Lol, ok. That's helpful. /s

What Joachim said.

Most likely a focusing issue.

Bob

norjens wrote:

Erik Baumgartner wrote:

canetsbe wrote:

Here are a couple more photos, one manual focused using 100% magnification on the camera LCD and focus peaking, vs AF. You will find that both methods are yielding the same result.

AF

MF

Also some closer shots, also yielding same result regardless of AF or MF. This one I compared f/2.8 to f/8.

f/8

f/2.8

A couple of questions...

Where is the focus point in the landscape shots? They still look crappy to me.

Are these SOOC jpegs or RAW conversions?

If they are from RAWs, what software are you using, and what processing has been applied?

From the context of the conversation here previously, I'd guess he focused on the trees. Regardless of exactly where it was focus, he states that it was manually focused at 100%, so it should be sharp somewhere. Clearly it is not, and no raw converter can do such damage to a file as we see here.

@OP: These are good samples, I think we can conclude from these that there probably isn't any other issue than the lens that is causing these results. It doesn't exactly look like softness, it almost seems more like some form of reverse coma (which isn't a thing, but that's what it looks like). I see a bit of it in my copy as well, but nothing like the amount seen here.

I agree, I don't think mine could look that bad at f/2.8 no matter where I focused it. Looks like it might be even worse on the right side in the landscape shots.

OP, my 16-55 is very sharp at all apertures and focal lengths. It's best at the wide end, but still very, very good at the long end - yours should be too. Send it back.

canetsbe wrote:

Rod McD wrote:

Hi,

The f2.8 image looks awful to me, even if it wasn't the ideal aperture. Nothing seems to be in focus anywhere in the image. Which leaves various possibilities.....

• Was it focused too far (beyond infinity, if that is possible on the 16-55 - it is on many AF lenses). And how/why?
• Was it focused too near? Seems unlikely.
• Was it on a tripod with the IBIS left on?

Other than those ideas, I'm stumped I'm afraid. A suggestion.... Put the lens on a tripod, IBIS off, and re-shoot using manual focus assisted by 10X magnified view. Try several shots refocused between each. If you get the same results, it's probably the lens optics, if you nail sharp focus, it's not the optics and you may have an issue with your camera's AF accuracy at that distance.

Then try it (still on the tripod) with the IBIS on, and see if that re-induces the loss of sharpness. If it does, it's the IBIS generating movement and it will confirm that you need to turn it off when using a tripod.

You could also repeat these tests at closer distances and see what results you get.

Regards, Rod

Thanks for the suggestions, Rod. Those are good ideas for troubleshooting. The photos above were shot hand held with IBIS enabled. I am inclined to believe that this is indeed a problem with the lens itself, as I get the same results when shot using my X-Pro 2. And other lenses seem to focus just fine on these bodies, without any issues focusing at distant landscape subjects.

Hi again,

You're right that there's clearly something wrong, and the fact that you're getting good result with other lenses suggests that it's something with that camera / lens combination. Remember that they are a system working together - it's the system that's giving you the results and it's not the same system with other lenses.  However, if you're thinking of returning it to Fuji and arguing that it should be replaced when their service response is that it tests OK, you'll probably need to be able to demonstrate more than belief.

I'd still advocate the process of elimination that I suggested above.  To which I would add, for landscapes, an additional test of shooting it slightly closer than infinity and at a distance well beyond the hyperfocal distance for 55mm and each of your selected apertures.  That test should nail whether it's focusing too far (beyond infinity).  If a lens manually focused beyond the hyperfocal distance for the FL and aperture doesn't produce a reasonably sharp infinity, there's something else wrong.  (That's with a standard default COC in the DOF calculator.  If you're going to pixel peep, print large, or otherwise be very particular, halve the COC in your DOF calculator when you calculate the hyperfocal distance you use for the test.)

All tests on a solid tripod of course.  With IBIS off.  Record notes carefully and post the images to tell us what results you get, and we'll help if we can.

Hope that helps,

Cheers, Rod

canetsbe wrote:

rwbaron wrote:

canetsbe wrote:

Joachim Gerstl wrote:

No, you most likely have no bad copy. Don't send it to Fuji yet before you give it another try.

Try to do focus manually using the loupe or simply use the AF on a subject which is about 5 to 10 meters away. To crank the lens to infinity doesn't work. Many lenses focus behind infinity. There is no way to get such a crappy image quality without a serious user error.

Lol, ok. That's helpful. /s

What Joachim said.

Most likely a focusing issue.

Bob

A focusing issue with the lens, yes; you do realize the point of this thread is that the lens is behaving incorrectly?

Could be the lens or the body. Have you confirmed that the lens produces an acceptably sharp image at infinity using manual focus? If it does then try a series of 10 shots off a tripod focusing on a good contrast target at infinity and see how many are in proper focus.

Bob

norjens wrote:

rwbaron wrote:

canetsbe wrote:

rwbaron wrote:

canetsbe wrote:

[snipped

What Joachim said.

Most likely a focusing issue.

Bob

A focusing issue with the lens, yes; you do realize the point of this thread is that the lens is behaving incorrectly?

Could be the lens or the body.

- OP already said that other lenses produce good images with the same body.

Have you confirmed that the lens produces an acceptably sharp image at infinity using manual focus?

- OP has provided images at close, medium/far, and infinity focus. All are confirmed to not be acceptably sharp for the 16-55, although the close focused ones appeared closer to acceptable.

If it does then try a series of 10 shots off a tripod focusing on a good contrast target at infinity and see how many are in proper focus.

- Why take 10 shots when he is at 1/5000 second exposure time? They would be 10 shots with the exact same (zero) motion blur, wouldn't they?

Bob

It would seem to me that you haven't read the full thread, because this doesn't appear to make much sense in that context. I don't mean to start an argument here, it just appears that OP already did the tests and got the feedback sufficient to conclude, so there's no reason to ask him for another round unless you can first explain a fault in the previous test and conclusion.

You're correct in that I did not read all the posts in the thread but have since.

The OP is attempting to determine if a problem exists with a lens but from what I've read and seen there is little to no attempt to eliminate variables. First, the scenes/subjects are not well suited for this and it appears he's doing this handheld. I've been evaluating lenses for my own use for decades and have a system for doing so and have learned the pitfalls and shortcomings of not doing this correctly.

He needs a better scene for evaluation. A city scape if available is ideal for infinity focus where a specific building or structure can be the target for focus. Ideally the scene should fill the frame and do not shoot over water or on humid days. For close focus a simple lens testing chart is best and if not available then small targets can be printed from various sources on the WEB and attached to a vertical flat surface. If all else fails then newspaper taped to a wall or garage door can work. When evaluating a lens for centering/tilt defects it's imperative the sensor plane is parallel to the target. This can easily be done with a small mirror in the center of the target but for purposes of what the OP is looking to evaluate it's not necessary as long as he realizes the corner/edges may show varying levels of softness.

The camera/lens needs to be on a solid tripod, not handheld. IBIS/OIS needs to be off and a remote release or timer should be used to trip the shutter. There's no need for raw capture as fine Jpeg's at default settings are adequate. The ISO should be a low as possible and the target/scene should have good contrast and be in good light. The camera drive  mode and focus mode selectors should be be in single frame for testing the AF.

The AF point should be of medium size and not larger than the target used for focus acquisition. It should be centered and not "focused and recomposed".

Start with manual focus using the magnifier and focus peaking. At minimum refocus and shoot 3 frames with 5 being better especially at wide open apertures. Then use AF to capture another 3 to 5 frames defocusing the lens each time. For zooms do this at each marked FL at both wide open and f5.6~f8. When shooting targets at close focus the camera/tripod needs to be moved between each FL. Compare the results.

The OP may very well have an issue with his lens as I found many over my years of testing, mostly Canon. I hate to see him send it in only to have Fuji repair return it as "in spec". I had a Canon 28-70f2.8L that after the above testing proved to be defective and even then Canon's service facility returned the lens to me twice saying it was "in spec". I requested the lens be sent to a different Canon repair facility and it was returned repaired with a note "replaced faulty focus module". The point is the OP should have more information with the highest possible confidence when the lens is sent in.

Bob

One of my test setups

Using a mirror to achieve proper alignment

One thing that can make it easier to see what is happening it to shoot a point of light.  Here is the setup I used, drilling a hole through a thin piece of aluminum and placing it in front of a light bulb:

setup for photographing pinhole

I photographed the pinhole from a distance of 34 feet, placing the pinhole under either the center focus point or one of the four corner focus points. Below are 100% crops from the raw images (converted using dcraw), with the aproximate pixel location of the pinhole in parenthesis:

center (3016,2016)

top left corner (681,805)

bottom left corner (626,3215)

top right corner (5444,807)

bottom right corner (5434,3254)

You will notice that the center and corners on the right pretty much make the pinhole look like a pinhole, but on the left hand side it is smeared around.  That's probably astigmatism.  To rule out camera shake, I would have to take several images on the left side and verify that the smears are all the same.

I used manual exposure to take these pictures, which was necessary because most of the frame has to be black to avoid overexposing the pinhole.  In theory they should all be the same brightness (except that the corners might be darker than the center due to vignetting).  I'm guessing that the differences in brightness are a camera issue.

Hi,

Looks much better.  Good outcome!

Cheers, Rod