50 1.8 II Focus Problems?

All,


I recently purchased a second hand (but new-in-box) nifty fifty. After taking some shots around the house and outside, I was beginning to question it's autofocus. So I setup up an impromptu focus test. I used a tripod and cable release taking a picture of a magazine article on the ground at about a 45 degree angle (3-4 feet from the lens). I had the camera set on Av Priority at f/1.8, ISO 100. Most of the pictures were 1/40 sec, a few were 1/30.


After each photo, I reset the focus to infinity to make the lens search for focus. I took 6 photos with the center focus point, 3 photos with the far left point, and 3 photos with the far right point. BTW, I have a 50D.


To summarize, the focus was all over the place. In case I am mixing up the terminology, when I say it front-focuses I mean it is focusing on the paper closer to the camera than where I want; and the opposite for back-focusing. For the 6 center focus point shots; I had 3 back-focused, 2 front-focused, and one dead-on. For the 3 left focus point shots; I had 2 dead-on and 1 front-focused. For the 3 right focus point shots; all 3 were back-focused.


If my test is flawed, please let me know what I need to do different for a more accurate test. Otherwise, what's going on with my lens?


Thanks,


Braden

Try a faster shutter speed you might be having handheld problems


Try the lens at 5.6


It is tough for a lens to start at infinity


The depth of field of the for that lens at 3 feet f1.8 is


near limit 2.97


far limit 3.07


for a total depth of field of .07ft (.03 ft is in front of the subject and .04 ft is behind the subject)


you might be asking too much at this close distance


try it at a longer distance


(or you might try a 10foot laptop, works for me)

Well done, you brought together two separate posts. I imagine some people will be lost by your last comment, but I got a nice chuckle out of it. [:D]


It was on a tripod and I was using a cable release to eliminate any IS problems.


I was intentionally looking for a narrow DOF so I could determine any focus problems. It seems if I decrease the aperture and increase subject distance, it would be very difficult if the lens was focusing accurately or not. Like I said, I have not done any tests like this before nor have I read how they are supposed to be done, so I may be completely wrong.


Let me figure out how to resize and crop and I will post a couple examples.

I was trying to focus on the indent of the paragraph beginning with "While travelling..."





Front Focus:


[img]/cfs-file.ashx/__key/CommunityServer.Components.UserFiles/00.00.00.38.96/Front-Focus.jpg[/img]





Back Focus:


[img]/cfs-file.ashx/__key/CommunityServer.Components.UserFiles/00.00.00.38.96/Back-Focus.jpg[/img]





Good Focus:


[img]/cfs-file.ashx/__key/CommunityServer.Components.UserFiles/00.00.00.38.96/Good-Focus.jpg[/img]

I stil think you are trying to focus too close for this lens (minimum focal distance of 1.5feet) especially for a 100 dollar lens.


I rarely ask any of my lenses to start autofocus from infinity.


You got any shots from a bit longer distance. Or maybe a 5.6 or 11 with the same experiment.





The dof for you is .07 x 12 inches (maybe this is too much for this lens)


You might try it at TEN FEET.

I did sort of the same test with my Sigma 50mm 1.4 and I also did it at about 3 feet so that shouldn't be the problem.


You can download a test chart which makes it more accurate and reliable. They are free for download and available in multiple types. Just use google. Often these charts come with a good manual, so you'll figure it out.


I think you're perhaps having trouble with the extra sensitive AF point around the centre in this occasion. I'm not sure, but it might be your problem. There's too much around the centre to focus on.


Try your test again with a proper focus chart and see what it does.


Hope you'll be more lucky then,


Jan

I have a nifty fifty that I use on a 40D and I encountered the same problem when I tried to take a pic of a menu at my nephews wedding in Jamaica. My settings were similar to yours and the pic was taken at roughly 45 degree angle too about 2 - 3 ft away. The depth of field at Av 1.8 is so crazy shallow that words one or two lines up or down were getting blurry. I ended up standing up and taking a pic straight down but even then the centre focus was sharp and the peripheral got softer to the outside.


[img]/cfs-file.ashx/__key/CommunityServer.Components.UserFiles/00.00.00.33.12/menu-2.jpg[/img]





[img]/cfs-file.ashx/__key/CommunityServer.Components.UserFiles/00.00.00.33.12/menu-1.jpg[/img]

For AF testing with that lens, you're correct in doing it at f/1.8 and at ~4 feet.


By focusing on a target like magazine text at a 45° angle, you're probably challenging the AF system too much. At that distance, small printed text is not offering a lot of contrast at that angle (yes, black on white text is high contrast, but the size of the contrasting elements is too small relative to the size of the AF point sensor). Keep in mind that the AF sensor at a given AF point is actually larger than the representation of that AF point in the viewfinder.


This image from Bryan's 7D review shows an approximation of the actual sensor size vs. the size of the box in the viewfinder (the top one is the relevant one, since the 50D doesn't have Spot AF). Also, I believe that the 7D's viewfinder representations on the transmissive LCD are bigger than the etched boxes on the 50D, meaning the actual AF point will be even larger relative to the VF representation on your 50D than in the image below.


http://media.the-digital-picture.com...7D/Spot-AF.jpg


A cross-type AF point (all of the 50D's are of that type) will use both vertical and horizontal features for AF - so, looking at your images, the AF system may be using the thick vertical line to the right of the column of text in it's AF determination. Since that line extends all the way up and down the page, it 'fools' the AF system a bit.


Personally, I use aLensAlign Profor my AF testing - it has a contrasty focus target that is straight-on to the camera, and a ruler at an angle just off to the side, to show the DoF. Printed AF test charts such asthis oneare similar, but the focus target is a thick black horizontal line with white space around it, so it's an adequate target with the chart at 45° (older versions of that same chart had a separate cut-and-paste focus target so the target was parallel to the camera and the ruler was at an angle, like the LensAlign tool).


I'd try the test again with the printed focus chart. Let us know how that goes.

Thanks for the insight and printable focus test. I will see if I can make time to give it a try tonight and see how the lens does.


I'll post the results

Thanks John, that's exactly what I meant to say [Y]

Thanks Jan and John. I used the printable focus test and retested the lens. I took 14 shots. 12 were dead on and 2 were slightly front focused. Is that an acceptable rate? Or would you expect even more consistent accurate focus? One focused 10mm short, the other 14mm short. So they weren't way off, but it was very noticeable.


Thanks for all the help.


Braden

Quote:

---

Originally Posted by TucsonTRD

I took 14 shots. 12 were dead on and 2 were slightly front focused. Is that an acceptable rate? Or would you expect even more consistent accurate focus?

---

I think an 86% hit rate is quite reasonable, especially for a $100 lens that lacks the faster and more accurate USM motor found in the more expensive Canon lenses.

I'm with John on this one. 12/14 seems like a good rate.


Those 2 are pretty far off I must say, I think this isn't a focus-accuracy problem (so it's not fixable by AF micro adjustment) but more an inconsistently fault.


And last but not least, if it focusses 12/14 photos dead on in real life use at f1.8 you are a very lucky man [;)]

Thanks for the input and advise. I was a little worried after my own uneducated test. I feel better now and I hope to get some good photos out of the 50.