[imfinetoday]

I have a k20 and my primary lens is a Sigma 24mm 1.8. The other day,I was shooting in lower than ideal lighting, so instead of adjusting my ISO, I adjusted my f to 1.8 (1.7). When I reviewed the photo, the focus was horrid. The spot i focused on was definitely not in focus (the spot was maybe 50ft away).

I have noticed this more than this one time. Is this a front focus, back focus or completely different issue?

[mccarvindh]

With a wide aperture focussing can be difficult - so, if that out-of-focus issue was a one time occurence, it was most likely because of the wide aperture. If you consistently find a particular lens misfocussing then you can suspect a front or back focus

[mccarvindh]

I see that you have had this more than once. Ok, if that is the case, please check to see if the area of focus is behind/front of the area you intended to focus. Once you confirm that you can easily correct minor front/back focus issues with the menu in K20D. I can look that up and let you know later.

[imfinetoday]

I see that you have had this more than once. Ok, if that is the case, please check to see if the area of focus is behind/front of the area you intended to focus. Once you confirm that you can easily correct minor front/back focus issues with the menu in K20D. I can look that up and let you know later.

I've used the focal adjustments, but it still seemed to do it. MY thought is that since it's at 1.7 the DOF is just so small that it freaks out and and decides it's going to focus wherever. dunno if that's the case. seems a bit fickle to me

[Marc Sabatella]

It *always* just decides to focus wherever. If you let the camera ick the focus pioint, it pretty much literally focuses anywhere it pleases. if you pick the focus point, then it promises to focus somewhere "near" where you selected, but that's about it - you can't pinpoint the focus. and in the camera's terms, even a background that is literally a mile away is still "near" the place you wanted to focus if it is close to the selected point in the viewfinder. Without seeing sample images, it's impossible to say for certain, but my guess is that the camera simly chose to focus somethwere other than where you intended. That's neither front focus nor back focus - that's simply life with AF.

it's AF if the camera does actually chose to focus at a given spot, but misses and focuses slightly (normally no more than a matter or millimeters) in front. it's back focus if it focuses a couple millimeters behind the place where it tried to focus. But unless you've carefuly set up a controlled test, you really have no idea where the camerea tried to focus, so it would be impossible to conclude anything.

[imfinetoday]

But unless you've carefuly set up a controlled test, you really have no idea where the camerea tried to focus, so it would be impossible to conclude anything.

I plan on doing a full on test hopefully tonight. i've seen this suggestion or i've seen people use the focus test chart. any opinions on either?

[KjetilH]

Even drawing a thick black line across a sheet of paper and hanging on the wall would show if the focus is off, without the hassle of printing charts (though the suggestion in your link should work quite good as well).
Focus charts provide a scale (for easy comparison) as well as the black line, which ensures that the camera has one thing to focus on.

Adding to Marc Sabatellas comment on AF, the camera just "sees" contrast, and goes after the most prominent one. Especially with wider lenses (such as your 24/1.8), it is quite easy to get multiple "contrasts" in a focus sensor area, and you don't know what the camera will pick.

---

Last edited by KjetilH; 05-06-2010 at 10:42 AM. Reason: Darn translation.

[GeneV]

It *always* just decides to focus wherever. If you let the camera ick the focus pioint, it pretty much literally focuses anywhere it pleases. if you pick the focus point, then it promises to focus somewhere "near" where you selected, but that's about it - you can't pinpoint the focus. and in the camera's terms, even a background that is literally a mile away is still "near" the place you wanted to focus if it is close to the selected point in the viewfinder. Without seeing sample images, it's impossible to say for certain, but my guess is that the camera simly chose to focus somethwere other than where you intended. That's neither front focus nor back focus - that's simply life with AF.

It was not until I got this concept into my head a few years ago that I could learn to use AF. Try to split hairs and focus on a head in a group some distance away, and you have just as good a chance of having the camera lock on the mountain peak miles behind.

I have an mf switch on my camera, and, in the proper situation, I'm not afraid to use it.

[Just1MoreDave]

The object of either test is to give the AF system something it should be capable of focusing on, and eliminate the possibility that it simply grabbed the wrong thing. Then maybe distance is a factor - a lens might work differently close to its minimum distance. And you should be able to interpret the results quickly.

Look at your subject like the AF looks at it. I've seen people attempt to focus on something like their computer keyboard. Yeah, lots of edges and contrast, but the AF can't really differentiate between Y and 7. One key has nearly the same value as the other. Plus the keyboard is often on bad mixed lighting, fluorescent from the monitor and tungsten from 12 feet away.

I would recommend at least trying a test chart, following the instructions as well as possible. At minimum, it gives you some ideas about a controlled test.

One thought about your problem: at 24mm and 50 feet, something has to be off. Depth of field should cover a lot at that distance, and even though 24mm isn't impressively wide on APS-C, it is wide enough for subjects to be kind of small and appear sharp just because of that.

[imfinetoday]

ok so i'm revisiting this post... I'm wondering if this is an issue in the more low light + distance thing. Up close (not quite macro) comes out pretty good from what i've seen.

I'm going to test with an actual test chart just to do it. Is there an good one i can download? I've seen these few: this one, this one & this one

are either of these ok to use?

[Marc Sabatella]

Yes.

[imfinetoday]

revisiting this once again...

I ended up contacting Sigma regarding this and I'm wondering if I'm getting the runaround. What are your thoughts? Here's what I was told:

My technician was saying that to use this lens wide open and distance it was not meant to do.
You would have to stop it down to get sharp images at about 3- 4 meters and beyond….
This lens has a calibration spec of f1.8 @ 1meter… so @ 1 meter this lens wide open is good….this happens a lot with this lens , but it was not designed to shoot at more than 20 ft @f1.8 and get sharp results.

I do understand that the "sharpness" will be lacking at a distance (@1.8), but I keep running into an issue of the subject being out of focus.

[Marc Sabatella]

Without seeing your test images, it's impossible to say.

[imfinetoday]

3ft:


5ft:


10ft:


15ft:


20ft:


30ft:

[Marc Sabatella]

In the distant shots, it's pretty clear the camera is simply choosing to focus somewhere other than on your subejct. No big surprise there, since the subject isn't big enough from that distance to cover the whole focus sensor.