Why is my picture blurry on one side? Did I damage my lens?

[strantor]

I'm pretty sure I know the answer, and I already ordered a replacement lens, but I'm hoping someone will tell my I'm wrong and I can remedy this before Monday. Here's a sample of the pictures my camera is taking:

The camera was not moving.
The whiteboard was not moving.
The whiteboard was perfectly perpendicular to the camera.

I've been doing hack job pseudoscience stuff with this lens and I'm not sure if clamped it too hard in my 3D printed medieval head crusher or I put too much side force on it during assembly or maybe I just don't know what I'm doing. What specfic kind of abuse could cause this, so I don't do it again?

 

[Grandpa Ron]

I suppose if you have bent the lens barrel or mount the lens would not be parallel to the white board. I would stop it down to a small aperture and see what happens.

I would think that if the lens is not parallel, to the image being shot or the focal plane sensor, it might not focus across the entire sensor.

You can try projecting the image on a piece of paper, to see if it lines up with the lens axis.

Good luck

 

[strantor]

I suppose if you have bent the lens barrel or mount the lens would not be parallel to the white board. I would stop it down to a small aperture and see what happens.

Ok I will try that later today and report back.

I would think that if the lens is not parallel, to the image being shot or the focal plane sensor, it might not focus across the entire sensor.

I am focusing manually. I used 10x digital magnification in the LCD, in the very center of the image, to get the focus just right. Then I took the picture, and upon review of the whole image rather than just the tiny box in the center of it, everything from center to the right is in focus but everything to the left (that was out of view while I had magnification turned on), is out of focus.

You can try projecting the image on a piece of paper, to see if it lines up with the lens axis.

I'm sorry I can't picture what you're describing.

Good luck

Thank you! I appreciate the help.

 

[Grandpa Ron]

Strantor,

I am not familiar with your set-up. But, a lens focuses an image on the digital sensor or the film plane. The digital or film plane is in the camera, so you do not see it. Trying to duplicate it with a cardboard box and a frosted glass may be more trouble than it is worth.

So, the question is "Why is half the photo in focus while the other half is not?"

1. What happens to the image when you 10x focus on the right side?
2. What happens to the image when you 10x focus on the left side?
3. What happens to the image when you focus without the 10x magnifier?
4. What happens to the image when you mount your camera upside down and 10x focus on the middle?

Those are the variable I can think of offhand, that may give you a clue as to what your setup is doing.

I hope you are working with digital; the results are a lot quicker.

Good luck

 

[strantor]

I would stop it down to a small aperture and see what happens.

Ok, here it is as low F-stop it goes (F/5.6):

and this is as high as it goes (F/36):

That 2nd picture is actually acceptable, even if the lens is damaged. I've never tried setting it that high before and I don't know why it never occurred to me to try.

I am supposed to demonstrate my progress on this project Monday morning.

I am not familiar with your set-up. But, a lens focuses an image on the digital sensor or the film plane. The digital or film plane is in the camera, so you do not see it.

My setup is "highly irregular." It's a Canon 5D MkIV DSLR camera with a 28-135mm zoom lens. When you say "a lens focuses an image on the digital sensor" I don't know if you imply auto-focus or just the general function of the lens, but I am not using auto-focus if that's relevant to your point. The camera/lens is in manual focus and I am controlling zoom and focus with 3D printed actuators. A microcontroller external to the camera takes the reading from a laser rangefinder mounted below the lens and adjusts zoom and focus such that any two pictures of the same object taken from two different distances anywhere between 2ft and 15ft, appears scaled the same (ex: 800 pixels = 1").

So, the question is "Why is half the photo in focus while the other half is not?"

After setting F/5.6 per your suggestion it becomes apparent that things were worse than merely half the image being out of focus. But the question stands, and I don't know.

1. What happens to the image when you 10x focus on the right side?
2. What happens to the image when you 10x focus on the left side?

I'm not able to try that as the auto-focus switch is buried under layers of assembly that the camera is installed in. I could disassemble everything, flip the switch, and, put it all back together, but I am going to have to do that anyway when the new lens arrives.

3. What happens to the image when you focus without the 10x magnifier?

The result is the same, I just now tried it. The magnifier I think is just a function of the camera's onboard LCD and doesn't have any bearing on the actual photo, but there's a good chance I'm wrong about that; I am posting in the "Photography Beginner's Forum" after all. If you know better, I'm here to learn.

4. What happens to the image when you mount your camera upside down and 10x focus on the middle?

Here are pictures I just took at 90, 180, and 270 degrees:

Those are the variable I can think of offhand, that may give you a clue as to what your setup is doing.

I hope you are working with digital; the results are a lot quicker.

Good luck

Thanks again!

 

[wfooshee]

You say your external microprocessor is adjusting zoom and focus through servos. How does that external processor know the focus setting is "correct?" Earlier you said you're focusing with Live view at 10x magnification, so the external processor is not actually focusing? I'm not understanding what part of this is processor-driven and what part is not.

Also, in your very first image, the numbers at the bottom left are sharper than the numbers at the top left. Also, the surface imperfection on your board are sharp at the top right, but not at the bottom right or bottom center. I don't think your board is as flat, or as parallel to the focal plane, as you think it is.

 

[Grandpa Ron]

As you know, clarity with a high f stop, does not solve the problem, it just takes advantage of the greater depth of field.

I assumed you were using manual focus when experimenting, because the techno-guru that designed the cameras auto-focus algorithm, would have no idea of what folks are doing when they deviate from the normal picture taking.

Shoot the scene with another lens, or the same lens removed from is mounting fixture, to see if the camera/lens will take a normal photograph.

It is hard to solve the problem without first finding the error source, from experience I know it is easy to start chasing your tail.

 

[strantor]

You say your external microprocessor is adjusting zoom and focus through servos. How does that external processor know the focus setting is "correct?" Earlier you said you're focusing with Live view at 10x magnification, so the external processor is not actually focusing? I'm not understanding what part of this is processor-driven and what part is not.

Ok, sorry; that was confusing how I worded it. To get the focus correct at any given distance I have to calibrate it at many points between 2ft and 15ft. The process looks like this:
1. Set camera 24" from target, zoom in 10x on live view (focus is at 90 degrees, image is blurry).
2. Create an entry in the microcontroller's look-up table for 24" input and 20 degrees focus rotation output, save lookup table. Focus actuator rotates to 20 degrees. Image is less blurry.
3. Adjust the 24" entry in the lookup table to 0 degrees, save lookup table, watch live view. As the motor turns focus from 20 degrees to 0 degrees, image gets clearer and clearer, then blurry again. I have passed the correct # of degrees for 24".
4. Try modifying and saving the lookup table a few more times with different values while watching in live view (still zoomed in 10x on center) until the perfect number is found, for example, let's say it's 4.5 degrees.
5. Move the camera 30" from the target and repeat steps 1-4 for 30". Let's say the right number for 30" is found to be 6 degrees. I save that.

Now if I move the camera back to 24" the focus will rotate back to 4.5 degrees. And moving it back to 30" will rotate focus back to 6 degrees. At any point between 24" and 30" the focus ring will be at a calculated corresponding degree between 4.5 and 6. For example 27" would be 5.25 degrees.
6. Repeat above at 6" increments all the way to 180".

After doing that, you can place the camera anywhere between 2ft and 15ft and the image will be focused (still, 10x zoomed in on live view).

I spent hours doing that. Then, I took it out of 10x zoom for the first time and saw the full image for the first time, realized the left side was totally out of focus at all distances.

It would be simpler to just set 24" at 4.5 degrees and 180" at 58 degrees, and let every distance in between be calculated, but the distance:degrees relationship isn't linear.

This probably seems like a lot of effort to replace what is already there, autofocus. But the environment in which I'm taking pictures is all one color, and almost featureless. There is nothing for autofocus to focus on. It doesn't work.

Also, in your very first image, the numbers at the bottom left are sharper than the numbers at the top left. Also, the surface imperfection on your board are sharp at the top right, but not at the bottom right or bottom center. I don't think your board is as flat, or as parallel to the focal plane, as you think it is.

Ok I will check this. Thanks.

 

[wfooshee]

Keep in mind that the focus ring might not be a mechanical connection in the lens. I know that on my ancient 18-55 kit lens, the manual focus is a mechanical link, but on my 70-300, it's not; I can turn the focus ring forever and ever either direction. All the ring does is tell the focus motor in the lens, "Go that way."

In other words, a given position on the focus ring is probably not repeatable; putting it here might not mean focus at 28 inches. If the ring doesn't have a hard stop at the end of the focus range, then you're wasting your time. If it does have a hard stop, then carry on!

 

[strantor]

Keep in mind that the focus ring might not be a mechanical connection in the lens.

Thanks for mentioning this. I had to figure this out the hard way.

I know that on my ancient 18-55 kit lens, the manual focus is a mechanical link, but on my 70-300, it's not; I can turn the focus ring forever and ever either direction.

I can turn the focus ring forever and ever as you describe, but in my case it is still a mechanical connection, just with some slip, like a clutch. I know it's mechanical because I can adjust the focus even with the lens off the camera. You can feel the difference between when the focus is actually moving, and when it's slipping. It presents a slight but noticeable difference in torque to move it once it reaches the end of travel.

All the ring does is tell the focus motor in the lens, "Go that way."

In other words, a given position on the focus ring is probably not repeatable; putting it here might not mean focus at 28 inches. If the ring doesn't have a hard stop at the end of the focus range, then you're wasting your time. If it does have a hard stop, then carry on!

While it doesn't have a hard stop, I have found it possible to make results repeatable by intentionally overtravelling during the homing sequence that it does on startup. The focus turns about 130 degrees before the ring starts to slip so I designed the actuator to rotate 140 degrees. On startup, it will travel to the home position at +140 degrees, and it starts slipping right before it gets there. Then it goes to 0 degrees, and starts slipping right before 0. Once at zero, all the slip is biased in one direction and as long as I don't let it go past 129 degrees, there is no problem.

 

[wfooshee]

That is how my 70-300 behaves as well. As long as you know you're hitting one end of the focus and move from that point as a reference, you should be OK. I was pointing out that if you moved past the limit of focus and didn't account for the ring not stopping, then backing the other way 15 degrees might not be where you thought it was. It'll be 15 degrees off the limit, but if your feedback didn't realize it had gone past the limit, it would lose track of its absolute position. Sounds like you've got it handled.

 

[strantor]

Here's an update. New Lens arrived (a week late) and this is how it looks:

F/3.5, close up, zoomed out:

F/5.6, far off, zoomed in:

So I guess the moral of the story is don't squeeze the hell out of your lenses with weird clamps.

Thank you both for the help!

 

[wfooshee]

Definitely better!