Front & Back focus in the one lense: any ideas??

Giphsub

I just tested my bigma against a focus chart, and found that the lens tends to backfocus at 500mm and front focus at 50mm, with a gradua change from front to back in between. This sucks! Has anyone come across this type of thing before? Any ideas? I might have to adjust for it in the d300 for the 500mm focal length because that is the focal length I realy want it for...

Giphsub

Why would this be happening?

ziggy53

I've decided that I'm not a big fan of the focus strips that purport to show you front or back focus. The problem is that autofocus mechanisms are not as point-like as people seem to think they are.

In reality, single point autofocus is a "region" of the scene. On the Canon 40D, for instance, I have noticed that the center point autofocus dot's region seems to extend to the limits of the circle in the middle of the viewfinder.

Most of the focus strips have too much "clutter" around the area you are supposed to point towards to make for an accurate target for checking autofocus.

A more appropriate target strip would have a large central target at the same plane as the film/image plane and then any other markings would be off to the side, away from the target area. The center target should also have a clear and sharp cross shape to give the autofocus something reliable to gauge.

I have something constructed that I'll get a photo of to show the concept.

Tee Why

Can we see the shots?
What was the test methodology, was the lens on a tripod?

ziggy53

Tee Why wrote:

Can we see the shots?
What was the test methodology, was the lens on a tripod?

Hopefully tomorrow I'll set it up. It's not that the parts are that big but I put it on an outdoor table and it's kinda spread out. Weather and time did not work out today.

Yes, I use a tripod to make sure that the focus point is where I want it to be.

Tee Why

ziggy53 wrote:

Hopefully tomorrow I'll set it up. It's not that the parts are that big but I put it on an outdoor table and it's kinda spread out. Weather and time did not work out today.

Yes, I use a tripod to make sure that the focus point is where I want it to be.

Sorry for the confusion, I was asking the question to the OP.

Giphsub

Yeah I can put the shots up later. I'll get to it when I get a moment.

Thanks for the reply Ziggy. The one I was using did have a strip very similar to the one you mention. The central area for focus was uncluttered and the measurement lines were well off to the side.

I'm pretty sure the focus changes from front to back as I change between 50-500mm focal lengths, I am wondering why this might happen with a large zoom such as this.

ziggy53

OK, I got stuff set up. Here is an image using the 70mm focal length of a Canon EF 70-200mm, f4L IS UMS (IS turned off).

This is just a white resin table with the target stuck to a paper plate turned upside down and a yardstick off to the side.



The full resolution image is here:
http://ziggy53.smugmug.com/photos/304235536_NFVkK-D.jpg

Notice that in this zoom setting, the image appears front-focussed. It is probably the autofocus picking up the front edge of the paper plate. In other words, the autofocus may have selected an edge other than what I wanted it to select simply because that edge was probably too close to the target and too close to the circle of sensitivity for the central point autofocus. It does not help that this is an f4 lens. An f2.8 lens might be different in how it works on the camera.

At 200mm things look more as I would expect:



This full resolution image is at:
http://ziggy53.smugmug.com/photos/304235536_NFVkK-D.jpg

This time the autofocus is spot on the target, mostly because this time the target is fairly filling the central circle of the camera, i.e. there is no distraction for the autofocus.

Giphsub

Thanks Ziggy. I will have another look at the images I took. I think there was enough space for the autofocus to work properly, but will check.