Canon clarity issues

mrdarklight · Aug 29, 2012

Hi folks,

Well, I got a new Canon 50mm f1/8 II, and I took some test shots of a test pattern, and the image is a lot blurrier on one side than another. When the center is clear, the left side is slightly less clear, but the right side is quite blurry. So I thought I'd better make sure it was the lens and not the camera, so I took test shots with my Tamron 17-50 and my Canon 55-250. For both of those, the images are clearer on the right edge, but still blurrier than on the left edge.

The way the camera is situated, it is about 4 feet from the pattern, and views it straight on. The right side cannot be more than a couple of millimeters closer or further than the left.

I'm beginning to wonder if I have a sensor issue in my camera... maybe it is not correctly aligned? My camera is a 60D.

I'll post some test images shortly.

mrdarklight · Aug 30, 2012

OK, here is the first set of test images, from a single image shot with the 50mm f/1.8. The first is the full image of the test pattern resized down to 1200x800.

This is the center at 100% crop.

This is the upper left corner, 100%.

This is the bottom right, 100%.

The bottom right is significantly blurrier than the center, while the top left is only slightly blurrier.

The photo was taken at f/4, ISO 400, with a flash.

mrdarklight · Aug 30, 2012

This is with the Canon 55-250.

Top Left

Top Right

Center

Bottom Left

Bottom Right

It's not as bad as the 50mm, but it still seems blurrier on the right side.

mrdarklight · Aug 30, 2012

Well, I took a few more pics. I took some with the aperture at f/22 and they were sharp all over. then I took some at f/1.8 and they were much blurrier all over. Also, one more thing, when I move the lens, I can feel something moving in there. It's like one of the lens elements moves back and forth. Does anyone with one of these lenses know if that is normal?

Edit: OK, I'm pretty sure I know what's wrong. There IS something loose in this lens. I pointed the camera down and shook it a little, then took a pic with autofocus, then I took it off, turned off autofocus, pointed it up and shook it a little, and took another pic from the same spot. They were noticeably different in focus, despite being taken from the same spot.

Derrel · Aug 30, 2012

I'm not 100 percent confident that the camera is ***perfectly*** aligned with the test target. In the 55-250mm lens test, for example, why is the grid pattern not perfectly aligned with the edges of the frame? Are you counting on a wall being **perfectly** flat and true and squared-up? Because if this target is tacked or taped to drywall, it might not be perfect enough. Without more description of the exact methodology used, the equipment, the verification of the alignment,etc, it's hard to state with any certainty if the lens is out of whack. At four feet, at f/4, with a 50mm lens, it would not take much misalignment to skew the target's rendering.

Compaq · Aug 30, 2012

I've not done any testing, but I would not be surprised if it's just dof issues in your tests.

gsgary · Aug 30, 2012

user error, go out in the real world and shoot, i have never tested a lens in my life

mrdarklight · Aug 30, 2012

Thanks for this... useful info. I was taking pictures in the real world and noticed my shots were soft on one side. But the lens is definitely defective. It's going back today.

MLeeK · Aug 30, 2012

This image looks like the camera is tilted to the left and with something that close that is going to make a HUGE difference at any lower aperture.

mrdarklight · Aug 30, 2012

But given that the lines in the image are all pretty much straight... how off could it be?

MLeeK · Aug 31, 2012

mrdarklight said:
But given that the lines in the image are all pretty much straight... how off could it be?

A significant amount. It has nothing to do with the straight lines, it has to do with the distance from the sensor to the subject. If it's even a fraction of an inch at a wide open aperture and you are shooting that close it will throw the focus off considerably. If you were 3 feet from the subject, which I suspect you are not here you would have .07 ft of REASONABLY sharp focus and at each end of that the sharpness would not be quite as sharp as the center.

Notice how the samples on the left are much sharper than those on the right? That's because it's within that focal range. The right side might be within but it's at the far edge of the range of focus.

If you were closer the depth of field decreases dramatically and there will be even more of a difference. Your camera and your test chart must be perfectly square to be accurate.

I suspect you are closer to probably 1.5 to 2 feet. At 1.5 feet you have .02 feet focus or from 1.49 to 1.51 feet would be in reasonable focus. That's no much more than a few eyelashes.

cgipson1 · Aug 31, 2012

Derrel said:
I'm not 100 percent confident that the camera is ***perfectly*** aligned with the test target. In the 55-250mm lens test, for example, why is the grid pattern not perfectly aligned with the edges of the frame? Are you counting on a wall being **perfectly** flat and true and squared-up? Because if this target is tacked or taped to drywall, it might not be perfect enough. Without more description of the exact methodology used, the equipment, the verification of the alignment,etc, it's hard to state with any certainty if the lens is out of whack. At four feet, at f/4, with a 50mm lens, it would not take much misalignment to skew the target's rendering.

^^^THIS!