Results 1 to 18 of 18

Thread: 50D autofocus

  1. #1
    BPN Viewer
    Join Date
    Jan 2008
    Posts
    218
    Threads
    36
    Thank You Posts

    Default 50D autofocus

    I've had trouble getting tack sharp images with my 50D on any shot where the main subject is several hundred yards away or more. Subjects closer than that distance (flowers, birds, etc) I've been able to get tack sharp images. I believe my techniques are sound (tripod, adequate shutter speeds, mirror lock-up for static subjects, f/8 to f/11) and my tack sharp up-close images seem to verify this. I've found that this issue is true for my telephoto lenses as well as my wide angles lenses. As the result of a tip, I recently began using manual focus on my landscape images using Live View at its 10x magnification. The results have been splendid; tack sharp images consistently. While I like the result, it saddens me to think that I cannot rely on the 50D autofocus for farther out shots. I have not checked the different lenses to see if a microadjustment would help, because that would only throw off the autofocus on the up-close shots. Has anyone encountered this besides myself? I'm unsure what to do now when I'm trying to shoot wildlife in their environment (ie moose with a mountain backdrop) since there may not be time for Live View mf and I'd like to be able to use af. Any thoughts would be appreciated. Thank you.

  2. #2
    Lifetime Member Doug Brown's Avatar
    Join Date
    Jan 2008
    Location
    Albuquerque, NM
    Posts
    11,879
    Threads
    917
    Thank You Posts

    Default

    Have you tried focusing using the hyperfocal distance for your body/lens combination?
    Upcoming Workshops: Bosque del Apache 2019, Ecuador 2020 (details coming soon)
    Website -
    Facebook - 500px

  3. #3
    BPN Viewer
    Join Date
    Jan 2008
    Posts
    218
    Threads
    36
    Thank You Posts

    Default

    No Doug, I haven't. To be honest, I'm not even sure what that is. I'm shooting with a 18-55mm lens and 100-400 lens most often so I'll say I'm shooting at 18mm and 400mm with a 50D. How do I determine the hyperfocal distance, and how would that help with my problem? Thanks.

  4. #4
    Lifetime Member Doug Brown's Avatar
    Join Date
    Jan 2008
    Location
    Albuquerque, NM
    Posts
    11,879
    Threads
    917
    Thank You Posts

    Default

    Check out this link Eric and see if it helps. Please post your findings too.

    http://www.dofmaster.com/dofjs.html
    Upcoming Workshops: Bosque del Apache 2019, Ecuador 2020 (details coming soon)
    Website -
    Facebook - 500px

  5. #5
    BPN Viewer
    Join Date
    Jan 2008
    Posts
    218
    Threads
    36
    Thank You Posts

    Default

    Doug, Thanks for the link. That site is very informative and very clear. I will definitely use that information in the near future. However, I don't think I explained my initial problem very well. My main problem is that the autofocus on the 50D is not giving tack sharp results at the precise part of the scene I've focused on, unless I'm quite close to my subject. I don't mind using manual focus for some scenes, but often autofocus is needed for those wildlife in their environment shots. I was hoping someone had run into this problem before and could help me out. Hopefully I'm not making things even more confusing. :-)

  6. #6
    William Malacarne
    Guest

    Default

    Eric

    Could you post an example of what you are talking about? If you do be sure to post the tech data also.

    Bill
    Last edited by William Malacarne; 06-01-2009 at 10:41 PM.

  7. #7
    BPN Viewer
    Join Date
    Jan 2008
    Posts
    218
    Threads
    36
    Thank You Posts

    Default

    Attached Images Attached Images
     
    Here is a test shot that I took a few months ago. I focused on the backlit tree on the left side of the image using the center focus point. I recomposed my image and took the shot using mirror lockup and the 2-second timer. (and a tripod) The shot data is 1/200s, F/8, 310mm, ISO 400. I've cropped in on the image to show that the portion of the image that is tack-sharp is the cornstalks in front of the tree. This same scenerio has happened to me many times. I would use the micro-adjustment and adjust each lenses, but the lenses are dead on at close distances. Any suggestions are greatly appreciated.

  8. #8
    Forum Participant
    Join Date
    Aug 2008
    Location
    Piedmont, CA
    Posts
    179
    Threads
    40
    Thank You Posts

    Default

    Eric,
    Similar problems on my side - same combinations - although more aware of the issue on the 100-400. Micro adjustments did help. I started at -5 and have been slowly working my way up. As most of the work with that lens for me is at distance, I have not been plagued with the close-in. The 50D is a new camera for me, so I got the 50D guide by David Busch which is really a jumbo version of the manual, but did have some good tips for me - one was that the micro adjust may need to go as high as +/- 17! So I am working my way there. I am debating also changing the focusing screen to one with a split image that aligns when focused. I'll continue to follow this chain in case others chime in with good thoughts.

  9. #9
    BPN Viewer
    Join Date
    Jan 2008
    Posts
    218
    Threads
    36
    Thank You Posts

    Default

    Cal, I'm glad to see I'm not the only one with the problem. (Well, I'm not glad for you, but you know what I mean.) This has frustrated me to know end, sometimes I feel like going back to the 300D. At least I know I can take sharp photos with it. I might have to determine a microadjustment for each lens that I would set only when shooting distance shots. Not a good solution though.

  10. #10
    BPN Viewer
    Join Date
    Sep 2008
    Location
    London/Essex, UK
    Posts
    92
    Thank You Posts

    Default

    From the example in post #7 the problem looks to me like atmospheric haze and not necessarily a problem with AF. The contrast in that shot is terrible. The sky is completely blown, which in itself is neither here nor there, but reviewing the histogram there is nothing in that scene anywhere near black. I would expect shadow areas to be darker than the histogram indicates. Your foreground does not look overexposed so where have the shadows/blacks vanished to? I think the atmosphere is destroying distant detail/contrast and causing an apparent softness. That you mention the problem occurs at "several hundred yards" gives merit to the "haze" argument. How is the long distance "focus" performance in cool conditions - first thing in the morning or late at night or after fresh rainfall when all the dust has been taken out of the air?

  11. #11
    BPN Viewer
    Join Date
    Jan 2008
    Posts
    218
    Threads
    36
    Thank You Posts

    Default

    Tim, I agree that there was probably a lot of haze on that day. However, when I view the image at 100% the cornstalks are tack sharp (that might be obvious in my example). Since I was using my 100-400mm, I'm quite a good distance away from the scene. The amount of haze between my camera and the tree is not that much greater than between the camera and the corn stalks. Yet the haze did effect the sharpness of the corn stalks. That's one reason I keep thinking it's an autofocus thing, but I might just not be not understanding you correctly. Are you saying that the haze is affecting the quality of the image (contrast notwithstanding) or are you saying that on a hazy day the autofocus can be fooled by the haze and focus closer than intended. Sorry if I'm sounding ignorant about this. Thanks for your help.

  12. #12
    BPN Viewer
    Join Date
    Sep 2008
    Location
    London/Essex, UK
    Posts
    92
    Thank You Posts

    Default

    Hi, Eric. I take your point about the 100-400 compressing the scene. I am talking about more distance meaning more atmosphere and thus more haze and lower IQ, but yes, perhaps there is little difference in distance, in your scene. However, in my experience I do not find that the AF gets thrown off by shooting at longer distances.

    Funnily enough I was taking some practice shots this morning and have a couple of examples which I hope illustrate my point. One shot was of a microlight at perhaps 500m distance, but I find it hard to judge. Anyway, it was a long way away. It was, of course, a completely pointless shot, but I just can't help taking opportunities to practice my long lens technique and equipment. The other is of a magpie in flight at perhaps 70m. Both were shot with my 50D and 100-400 @ 400mm. Although the exposures are different by 1 stop, in each shot it takes +0.7 stops of Exposure adjustment in Lightroom before I clip the whites, so they are comparable exposures. I have tweaked WB for the microlight. There are no other edits to either image.

    What should be clear to see is that the microlight image is actually really quite sharp but the contrast is absolutely abysmal. The blacks/shadows are nowhere near where they should be. Of course, editing can fix that, but just look at the difference between the two images. The only explanation I can think of is that the increased distance is killing the IQ of the long shot due to whatever is going on in the atmosphere between lens and subject.

    Full frame images....


    100% crops....


    Clearly the "blacks" in the magpie look black. The blacks/shadows in the microlight look nothing of the sort, but I would not say the microlight image look soft/OOF, not even at 100% magnification. It's not as though the lens itself loses contrast when targeted at distant subjects, since it acquits itself well enough, for a zoom, for pictures of the moon taken at night when the atmosphere is cool and the dust/pollen has settled and the fumes of transportation have died down. As I said in my earlier post, your scene lacks contrast. If the cause is not a hazy atmosphere or a nasty cheap filter* then I don't know what it could be. You may well have a problem with long distance focusing, but that is not the only issue facing you in this scene. If you can eliminate the haze (flare?) from your test scene then you may do better at identifying whether you have an AF problem.

    * Actually I don't think you commented on whether or not you have a filter fitted. If you do, try removing the thing and then try again without the filter but with the hood fitted. Even though I own several high quality filters I rarely use them. They do not play well with the 100-400 and I am not keen to use them on my other lenses either. The CPL may get the occasional outing but the UVs stay packed away where they can do no harm.

    If anyone has any alternative theories about the problem with long shots then I'm all ears :)
    Last edited by Tim Dodd; 06-27-2009 at 07:59 AM.

  13. #13
    BPN Viewer
    Join Date
    Sep 2008
    Location
    London/Essex, UK
    Posts
    92
    Thank You Posts

    Default

    Regarding AF microadjustment, at what sort of subject distance are you performing your calibrations? I've seen two recommendations coming out of Canon....

    1. Calibrate at your intended shooting location, at your expected shooting distance, in the weather prevailing at the time. i.e. if you calibrated in your back garden at 2m distance in winter then that may not suit you well when it comes to shooting at 50m in the middle of the day during summer across a lake.

    2. Calibrate at a minimum of 50X focal length. For a zoom, perform the calibration at the long end. So for the 100-400, for a "general calibration" you should be setting up your focus target at 20m or more and certainly not at or near MFD.

    According to Canon, whichever approach you take, your focus target should be parallel to the sensor plane and not angled.

    What I would do for a field calibration, if I had a tripod with me, is to aim the camera at a tree trunk, or side of a building, with good contrast detail, at my expected shooting distance and use that as my focus target. I would AF on a contrasty area facing directly towards me and once focus was locked I would switch to 10X Live View. If the image looks sharp then job done - go shoot. If the image looks a little off then, very very slowly and carefully I would manually focus by turning the focus ring anticlockwise, at the same time watching the screen. If the image sharpened a little then I would know the camera was front focusing and I would need to dial in some (more) +ve microadjustment. If my manual focusing softened the image immediately then I would know that I was not suffering front focus. However, I may not know whether I could improve things by going the other way. I would defocus the lens completely, focus again using Quick AF withi Live View, and this time turn the lens very very slowly and carefully clockwise. If the image began to sharpen up I would know I was suffering back focus and would need to dial in some (more) negative microadjustment. That might mean reducing from +5 to +2, for example, rather than necessarily making the adjustment an actual negative figure. It all depends on where you started from.

    HTH,
    Tim.

    p.s. If you have a filter fitted, once again, take the thing off, both for calibration and shooting.

  14. #14
    BPN Viewer
    Join Date
    Jan 2008
    Posts
    218
    Threads
    36
    Thank You Posts

    Default

    Thanks for the tip on calibration Tim, I will definitely try that next time I'm shooting a static scene. I guess I wasn't aware that I would need to calibrate my lens separately depending on whether I'd be shooting close range, mid range, or long distance. My mid range to close range shots are tack sharp, so I was figuring the calibration was fine.

    I think I didn't explain myself clearly with my 100% crop example. It's not the compression of the lens that I'm talking about, it's just that the stalks in front of my intended focus point are tack sharp, while the intended tree is not. Their distance apart varies minimally when compared to their distance from me, so the amount of haze that should affect their sharpness is basically the same. I think this awful photo is not serving us well in our discussion, let me take some other example shots in clear morning air and post them.

    Btw, I'm not using any filters.

    Eric

  15. #15
    BPN Viewer
    Join Date
    Sep 2008
    Location
    London/Essex, UK
    Posts
    92
    Thank You Posts

    Default

    Hi, Eric. I understand what you are saying about the comparatively small difference in distance between your intended focus target and that which appears sharp, relative to your distance from both. I meant to go back and remove my reference to "compression", but forgot.

    Regarding calibration, there is this small bullet in the 50D manual at the top of page 181....

    It is best to make the adjustment at the actual place to be photographed. This will make the adjustment more precise.
    There is probably something lost in translation there, as it is not exactly a fully helpful piece of advice but what I think it should be saying is that exact subject distance and also environmental conditions can influence AF accuracy, so if you are able to calibrate in the field you should get the best results.

    Certainly the 50X calibration distance comes from Chuck - http://www.digitaljournalist.org/iss...tech-tips.html. Now, it is probably fair to say that 50X FL is not always appropriate. e.g. would you want to calibrate a 180mm macro lens at a distance of 9m? I doubt it. If you only shoot with your 100-400 at under 5m, for example, then calibrating at 20m (50X 400mm) seems a bit redundant. The results may turn out fine for your purposes, but they are unlikely to be better than the results when calibrating at the intended range. 50X seems a good general purpose figure in the absence of more specific knowledge of the shooting conditions and demands. Also, by calibrating at a distance that is a multiple of FL, your calibration target will remain a constant size within the viewfinder and as seen by the AF system.

    The thing to remember about AF microadjustment is that, compared to a full calibration by Canon, it is rather basic - you can only set one adjustment which has to cover all focal lengths and all focus distances. If you send the lens to Canon they will use software to update a table of values within the lens so that the lens itself can compensate for any performance anomolies. This calibration should cover multiple focal lengths and, I suspect, focus distances. It may be that your lens (or body, or combination of the two) requires a full calibration by Canon rather than a rudimentary AF microadjustment fix.

    HTH,
    Tim.

  16. #16
    BPN Viewer
    Join Date
    Jan 2008
    Posts
    218
    Threads
    36
    Thank You Posts

    Default

    Tim, Sorry it has taken me so long to respond. I have been too swamped here to address this issue and run further tests on my camera. Thank you so much for all your help and for your explanations.

    Eric

  17. #17
    BPN Viewer
    Join Date
    Sep 2008
    Location
    London/Essex, UK
    Posts
    92
    Thank You Posts

    Default

    No problem, Eric. I hope you get it sorted. FWIW, I rechecked my calibration a couple of days ago for my 100-400 and decided it did need a bit more adjustment. Here's my test target, which was set up close to 20m away, perhaps a little less....

    The CD is oriented to be parallel with the sensor plane - i.e. facing towards the camera exactly, or near as dammit - and represents my focus target. The CD is large enough in the frame to fill the spot metering circle and thus it is impossible for the AF to be drawn off to something else. The angled ruler is simply so that I can judge where my focus is actually ending up.



    100% crop....


    The CD is lined up with the 40" mark. As you will note, the 40 does look fairly sharp but things go soft pretty quickly just behind the 40, whereas to my eyes the 42 and 43 look to be about the sharpest numbers all in all. Thus I have a little bit of front focusing here. After firing off three shots at each of +1 (where I started) to +10, defocusing manually between each refocus, and checking the results I settled on an adjustment of +8.

    Here are is an example of the close performance at 400mm, 1/250, f/6.3, 400 ISO....

    Full image....


    100% crop....


    I'll pop long distance results in the next post.

  18. #18
    BPN Viewer
    Join Date
    Sep 2008
    Location
    London/Essex, UK
    Posts
    92
    Thank You Posts

    Default

    At a guess I'd say this pylon is about 300m away. This was shot handheld at 400mm, f/5.6, 1/1600, 200 ISO.



    100% crop, converted from raw in Lightoom with no edits - default sharpening etc....


    Run through Neat Image with conservative sharpening applied....



    I can't complain about that performance from a zoom lens and a camera with the merciless pixel density of the 50D.

    EDIT : Another example, shot this morning at 400mm, f/5.6, 1/1250. 400 ISO. Blacks deepened and sharpened in Lightroom....



    100% crop....
    Last edited by Tim Dodd; 07-17-2009 at 05:40 AM.

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  
Web Analytics