No sharp images

Discussion in 'Canon EOS' started by peterridding, Dec 5, 2008.

  1. Ater weeks and weeks of homework, I have just purchased the 450D and 17-55 f2.8 IS USM, as well as the 70-300 f4/5.6.  I have taken about 700 shots using every combination that comes with the camera and 17-55 lens.  I am absolutely disappointed in the results.  Blurred, soft, some over exposed, but even on a tripod, faster shutter speed, I still do NOT have a sharp image at 100%.  My S5 IS takes a better image.  Obviously I thought it was driver error, but after reading all the advice and experimenting, with no better result, I may have a bad camera body.  Is this possible?  Could it be the 17-55 lens?  I am totally at a loss what to do, and the combination was expensive in AUD. I have been into photography for years, film slrs and compact digital, now dslr.  It comes down to driver error or equipment error - any help will be greatly appreciated, preferably before I hang myself.
  2. Peter, can you post a couple of examples of what you mean?
  3. I have a 40D and finally pulled the trigger on the 17-55 and have been very happy with the lenes.. Very sharp and much better than any non L's I have had. The only thing I have that may beat it is the 70-200 2.8L..
  4. There are several differences in the way dSLRs work (compared to P&S cameras)
    • dSLRs do a much less aggressive internal postprocessing (leaving more room for human processing). This often leads to the (wrong) assumtion that the images are less sharp. The aren't ... in fact, they are less sharpened. Play with the sharpness parameter in you camera menu and see whats happening. Also, try to play with the image in postprocessing a bit.
    • Same applies for saturisation.
    • Depth of field is a very important difference. While P&S (due to their tiny sensor) usually show the whole picture (from foreground to background) sharp, dSLRs do NOT! There is a technical reason for this, and the photographer should be aware of it when taking the shot.
    The other thing is, that having so much money spend, one usually expects immediate and enormous improvements. But very often there isn't. The S5 IS wasn't a bad camera (by the way) ... but compare two lowlight shots taken at 800iso ... just to give you something immediately visible.
  5. Peter,
    My suggestion is to "take a step back" and work through a few tests to help identify the cause of the issue. First up, I would suggest taking a shot of something contrasty and well defined (eg black writing on a white page). Mount the camera on a tripod - use manual focusing with liveview set to 10x, and preferably with a remote release or self timer.
    If you get a good result then repeat the test, but with the AF switched on to see if it's an AF issue.
    Be aware that when shooting RAW, it's essential to apply sharpening in post processing; Canon recommend amount = 300%, Radius = 0.3, and threshold = 0 as an excellent starting point - this makes BIG difference when viewing pixels at 100%. If you're shooting jpeg then default sharpening is typically smaller amounts and larger radiuses, which make an image printed at a typical size more defined, but don't improve it the same way when examining pixels at 100%.
    Other things that enter in to it are AF mode (1-Shot or Servo) - AF point selection - Depth of Field - shutter speed (in relation to camera shake) - filters - subject.
    I'd suggest trying the above, and let us know how you get on.
  6. I also couldn't get a sharp image (with the 50D). I sent it back to Canon 2X for correction but the result was "no improvement" each time. I returned it to the camera store. They replaced it with another faulty camera so they replaced it again. Now, I think that I have a camera that works.
  7. Peter, I have the exact same set up as you and have noticed similiar results. I have noticed a few things that may be of some help. First is, as stated above, the results can be improved dramatically in post processing and there is a menu function to increase sharpening. Most of my soft shots in retrospect have been traceable to me by some error. For instance forgetting to stop back down after shooting at f2.8. I have also occasionally forgoten to turn off image stabilization when shooting on a tripod. I have also noticed that what looks a little soft on the monitor looks sharp as a tack printed and could be an issue regarding screen resolution.
    Having said that however, I still am often disapointed in the results with this combination of lens and body. I often get images I feel are unacceptably soft when there is no apparent reason. For insance shooting at f11 at a 500 shutterspeed on a bright sunny day. I'm not sure if this is true but I have noticed that when shooting objects up close or within say 50 ft. I get very sharp results. When shooting landscapes at wide angles the horizon line and things around the edges are soft, especially any details in trees. I can usually get the results I want but I have to check and adjust constantly, for what seem like very normal and unchallenging situations. This is with the 17-55 EFS IS I am refering too. The 70-300 seems pretty sharp except fully zoomed at 300mm. Always a little soft there.
    My biggest complaint with the camera is the sensitivity to highlights. I have found it necessary to underexpose a stop or two in any outdoor situation to keep the highlights from blowing out. This is a constant struggle that requires extremely close attention to preserve detail. I'd rather underexpose and adjust in post processing than lose highlight detail that is impossible to recover. I have enabled the highlight tone priority function but to little avail. My other big complaint is a lot of noise at any exposure longer than a second. Any shot taken at night seems to have a ton of noise that looks like orange snow. I have poured over the manual and enabled noise reduction, but it seems to do little and it is so severe I am unable to eliminate it in post- proccesing. I have considered upgrading to a 50D or a 5D, because I want a camera that handles these issues more effectively. I have to do some homework though, to make sure it doesn't have the same sensor. I'd consider trying Nikon but I don't want to buy new lenses.
  8. Again, this is all conjecture until the OP shares a photo - that would likely narrow it down quite a bit.
  9. Peter -- I'm actually in a similar position, but I haven't ruled out operator error [in my case] just yet. I have yet to try a few things, but I also posted a thread ( -- just a day before yours, incidentally!). I haven't been getting many blown highlights, but I'm not sure if my issues could be fixed by calibration.

    I'm waiting for the weather to clear up a little bit before I try some shots with the 17-55 outside; I've been shooting with the 50 prime and the results aren't ideal, operator error or not.
  10. Can't see the 8 responses to my question.
  11. I don't know what happened, either -- I posted a reply and this is what happened.
  12. I have fixed the thread. Please do not use Microsoft Word's HTML mode to try and create posts on It causes massive problems.
  13. I didn't think I used Word at all in my reply, but something must have gotten screwed up in the process... I'll be more careful next time.
  14. Hi again, I noticed I have 8 responses, and thank you to those people. It was extremely hard to read as they don't show up as my answer above does, but in some word pad with all the heiroglyphics associated with it. Wonder what happened? But can read enough to take the advice on board. It was interesting to read about the difference between my S5 with the smaller sensor and the 450D with larger sensor. I can understand why the S5 pics look better. Thanks. I have tried the 1 shot, AI focus AF, AI servo AF, with the automatic and manual AF point selection and evaluative metering, but again the landscape features are not just soft, but more like blurred. I used higher shutter speeds handheld with higher ISO to try and get a little depth of field, no good. I used the tripod (a good one) but did not turn the image stabilizer off - the manual says to to save battery power - so I will do that. I have noticed that even with the 2.8 lens I need really strong light so I can use 100 ISO and get some depth of field, so have to have a fairly low shutter speed 30th to 60th sec but consider the IS would handle this. I'm into landscapes and will be tramping around NZ early next year so won't be carriyng a tripod with this rather heavy setup. I haven't put any post processing into the pics, and taken them all on large jpeg (highest setting). I'm getting desperate as I just can't get my pics sharp. Would it be possible to post the comments again so I can read them properly and print them out? I know I'm asking a lot! To be fair, I did capture a few decent pics of a seagull on sports mode with the 300 lens, and a few decent evening shots with the 17 lens, but again at 100% they show some softness. I captured one sharp shot yesterday, indoors, handheld, of a bauble on the Xmas tree. Not bad out of 700 shots! So this must be driver error, if 5 in 700 come out clear? I won't even go into the blown out highlights this time, just the exposure/sharpness. I wait patiently for those good people to rewrite their comments. Thank you so much.
  15. Thanks Josh, I just hope it wasn't me. I can see the responses now, thanks again everyone, but anymore feedback is still most welcome.
  16. Peter ... without a few sample images showing the problem, further answers are useless. ... And by the way ... I'm not going to rewrite my comment.
  17. Thanks Rainer, the comments are all above now. I will post some samples.
  18. Here we go
  19. Aaaaahhhh, was trying the link so you could get a bigger image but didn't work so will a resized one be ok?
  20. If you are using a UV filter for Protection, Take them off. I did, and my photo sharpness has improved 100%...
  21. Thanks, but no filters.
  22. From Peter, front yard resize. As well as everything soft, the highlights aren't good either.
  23. <p>2 things peter, is `IS` on or off? if on when you focus the IS takes time to settle, if you clic too quick t
    he lens will still be moving internally, if `IS` is off, when you focus&nbsp; the USM is very quiet have h
    ear d of the MF/AF switch not on right so move it back &amp; forth, then when you focus on a close object
    (book) by half press the shutter then something in the backyard, half press again (1 shot mode) does it move th
    e elem ents and look sharp in the viewfin
  24. Hi Chris, I did wait for the IS to settle in and yes, the lens automatically focuses near and far, I can hear it, and the images look sharp in the viewfinder. Is it best, for landscape say, to have the automatic (9 dots trying to focus) or the manual setting of using one dot better?
  25. there is a huge difference bewteen a dslr and a p&s. first, in a p&s it supplies all the thinking it is setup, from the factory to give good to very good pics straight from the box. and does. a dslr will not do this YOU have to supply all the thinking; or it is very likely that the dslr shots will be worse than the p&s.
    i can give you some general guidelines-
    shutter speed. keep it at 1/125sec or faster(you can get away with 1/60sec if you have to). this depends on subject movement. 1/125 will stop a person walking or running but not a car.
    fstop. use f5.6 to f11.0 if possible. below f5.6 you will get a pic ok but the lens will begin to get soft and the dof will drop off. above f11.0 you run into diffraction distorsion. thius is what light does when it passes through a small hole. does this mean at f12.0 the pic is fully distorted? no. but it is beginning. max IQ and performance is between about f5.6 and f11.
    IS and tripod. turn IS off when the camera is on a tripod. it has nothing to do with the battery. it is because the IS will continually check for shake, and that check will actually cause shake when there was none there. so it is best to turn it off alltogether.
    kit lens. it is not the best but certainly capable of give reasonable pics.
    be sure to use proper stance and han dholding technique. what you can get away with with a p&s is not what you can do with a dslr. the dslr demands better technique or it will show in the pics.
    iso. use the lowest iso you can that gives a reasonable shutter speed and fstop. the lower the iso the bettter the IQ.
    exposure. if you know how to use a histogram try not to let the histoine hit the right side of the hisrogram box before it hits the bottom. if you do the highlights will be blown.
    below is some prewritten remarks on dslr and p&s.
    i wrote the following posts some time ago, they may be of interest.

    no matter which dslr you buy.

    heavily consider the following. there are NEW DSLR owners' writing in all over these forums on this subject.
    when changeing from a p&s to a dslr, there is a huge difference.

    when you take p&s out of the box add a memory card and a fully charged battery you can now shoot and take very good pics.
    BUT, you cannot do this with dslr. the camera HAS TO BE SETUP first. you have to adjust the contrast/saturation/sharpness/shooting modes(color style or whatever it is called) to your likes. if you don't it is quite likely you will disappointed with results. your p&s will likely outshoot the dslr.
    to setup-you have shoot a test shot make ONE adjustment reshoot check pc screen readjust, until you are satisfied. and you do this with each of the adjustment types. then you have all the custom adjustments in the menu to check and if wanted change.
    when done you can put the camera into AUTO or PROGRAM and get reasonably nice shots. i would advise at first staying with jpeg. as you learn about the camera and photography you can then go to the other shooting modes and try RAW if you wish.
    dslrs are made to see the shot through the optical viewfinder not through the lcd. this is true of almost all dslrs including the k10d.
    dslrs and color.

    if you mean heavy saturated colors then no dslr is going to do that. they are not made to give strongly saturated colors. they are made to give ACCURATE COLORS. not heavy saturated colors.
    this is not the same thing at all. too many people who come from a p&s are very disappointed in the dslr colors, because they are not bright and saturated. this is because they are and have been using a p&s which has been giving them saturated and incorrect colors for so long that they think it is the right look. nothing could be further from the truth. the p&s colors are wrong, wrong. the camera manufactures know that the public buys high megapixel and heavy saturated colors and is what they make and sell to the public.
    but the slr/dslr is a whole different world. for the dslr accuracy of the scene in terms of view and color is a religion rpt religion. you want accurate color that is what you are going to get with dslr. but they will not be the bright saturated colors of a p&s. ytou can with adjustments in the menus up the color is dslr, but it will not look the INACCURATE CARTOON COLOR of the p&s.
    if you are wishing to buy a dslr for more and brighter color, save you money the p&s is what you want.
    not too long ago a new owner of a dslr was on these forums talking about the poor color of his new dslr. it seems as if he was shooting on an overcast day. many many people replying to him told him that cloudy day shots give the most accurate color, which they do. he couldn't believe and get over that idea. he also owned a p&S previously.
    you might be interested in this; which i posted a while back.;

  26. Limit yourself to 24 shots a week. You'll figure it out before the second week is over.
  27. Hi, I`d rather use one focus point as the camera will choose a FP which covers a point with the most contrast, normally the one you don`t want. Try centre point only on something like the darker tree in yard. with `1 shot` 1/2 press then reframe, also with such wide angle it maybe better to zoom in focus and then zoom back out. 200ISO in `Av` mode at 5,6~f8 should yield a sharp pic, otherwise :)
  28. Chris, I've tried a similar test but to no avail... I've been taking pictures almost exclusively with center point AF and the camera still doesn't get focus bang on. (I started a separate thread before this one -- "Questions about Focusing" -- in this forum.)

    Peter, that being said, you should try setting the AF point to center -- hold down the right-most button (top right corner, in other words) while rotating the selection dial until the center point is highlighted. Try shooting with this one point (and AF set to "One shot") on a contrasty subject; another possibility is to do a focus test. You can grab a pdf version of the site here:
  29. Peter: All I've seen here is your resized image, shot at 1/125th of a second at f3.5 at 17mm.
    A few changes will likely result in a sharper image. First of all, I would bet that the 17-55 lens, although it has a good reputation (I don't own one) almost certainly is somewhat soft, particularly near the edges of the frame, when shot at the limit of the zoom range (17mm). Secondly, it is highly unlikely that it is as sharp as possible when shot at f3.5. Even on the L series of lenses (including primes), noticeable improvement in sharpness is seen as you progressively stop the lens down to at least 2 stops smaller than wide-open, so I'd expect f/5.6 or f/8 to give you a better image
    Also, though I can't judge your original from the downsized image yo provided here, it is clear that you can improve the apparent sharpness of the posted image by applying a sharpness filter in Photoshop. There are of course lots of good suggestions posted already, and the advice is good.
    Good luck!
    Although it may seem counter intuitive, when resizing (or "downsizing") to a smaller image size, you have to sharpen for the new image size.
  30. Thanks Xingyuan, I don't really understand your directions about moving the AF point to centre, but I did download the focustestchart and the result, at 100%, is not too bad. You can certainly read it, but a bit soft.
    Chris, I had forgotten that oldie of zooming in and focussing, then zooming back out to take the pic, on manual focus. It works. But I really want the autofocus if possible. I will use the single point and recompose, but check out this link for recomposing, it's very interesting. Also, I noticied that my pics of the dreaded front yard were a tiny bit sharper with manual, but they were much better, not blown out or overexposed on manual as they were on AE. Why could this be? The AE is still soft, but I will work in the parameters you and Gary have given. The camera and lens set up certainly doesn't like contrast.
    Gary, thanks so much for all that valuable information about the two types of camera, and the hints on the dslr.
    Do you want samples of the latest experiments?
  31. Thank you Richard.
  32. Peter, how are these compared to yours? (These are mine.)
    IS was off on all four tests, I cranked up JPEG sharpening to full (I did, anyhow, before I got an ERR 99, but an off-on reboot solved that. I didn't double-check to see if the reboot undid my settings).

    Auto-focus was was center point; manual focus was through 10x LV with the LV mode set to Live View (rather than quick). Taken at an approximate 45° angle (angle between camera lens and subject plane) on a tripod. Mirror lock-up enabled.


    Ruler test: center AF point was set to the 6" mark.

    Ruler test -- AF;

    Ruler test -- MF through 10x LV;


    Nikon D70 focus test page: focused on as the instructions said... on the black line in the center of the page.

    Focus test -- AF

    Focus test -- MF
    What I meant by moving the focus point: the 450D has nine AF points. In complete AF, the camera picks which points it thinks it can focus off of and then uses those to focus. By the suggestions of choosing one specific focus point, we mean that it's possible to select any one of the nine and use that specific point to try to obtain focus. There's a command dial on the top of the camera just behind the shutter. When you push the focus button on the top right corner of the back side of the camera and look through the viewfinder or at the screen, you'll see the nine AF points... rolling the control dial will scroll through the points. The suggestion was to try to the center AF point and do the test that way.
    Take a look at page 61 of the manual -- it explains the process with pictures, which I suspect will help quite a bit.
    Richard: what you see or suggest from the focus tests in my post?
  33. Hi Xingyuan, fully understand now. You think this will make a difference to the whole scene by focussing on a particular spot within that scene? Do you manually or auto focus generally? Did you read the recompose link? Here is the auto LV of the test. I will then post the manual view. They are nowhere as good as yours with the ruler.
  34. and again
  35. The first shot (the autofocused one) shows a noticable amount of frontfocussing. Redo this test with 10 images and see if this behaviour is consistent over all (or at least) most shots or if it varies from front- to back-focus.
    The second shot (the manual one) is slightly backfocused. But with the screens in the AF dSLRs its quite hard to aquire proper focus manually.
    Also ... redo one of your outside shots ... set the lens to infinity *manually* (and of course switch to MF mode) and take shots at several focal lengths. Compare those to the shots you have. Make sure you use a tripod and shot at f/8 (one shot) and at f/2.8 (another one). Those shots should turn out ok ... the f/8 should be sharper over the complete field than the f/2.8 shots.
    If you can, redo the AF test with a different lens. If the focustests show consistent frontfocus and if the manual focused tests of your outside shots are ok, I think the lens+body combination should be sent to Canon for adjustment.
  36. Peter, I'm not entirely sure why your image is softer as a whole than mine -- I think the biggest reason could be that the image you shot was at a focal length of 55 -- try zooming in a little less, if possible, and then see if things sharpen up a little bit. Your ISO is pretty low, so there shouldn't be much noise coming in, and you also shot at a high enough shutter speed for the image not to be motion blurred. I'm not sure what else would cause the softness. (Did you leave the sharpening in-camera as default?)
    When you shot MF, did you use 10x LV or just what was displayed on the screen? One option you could try is to shoot LV, but zoom in with the same button that you use to zoom in when reviewing a photo (again, this is the same button as the one used for changing the AF point). Pressing that button once will display a 5x magnification; pressing it again displays a 10x magnification. This should make it easier to tell exactly what's in focus and what's not. I suspect that you'll get a better MF result through this option.
    When you shot the images, did you try using just the center AF point? The technique would *not* be to focus and recompose -- it would be to center the focus test so that the center AF point is on the black line. You would then just focus and shoot.
    As far as your questions go, "making a difference to the scene" is hard to quantify. With a properly focusing camera (which yours appears to be -- mine was out of focus by quite a bit more), it shouldn't necessarily matter which points the camera chooses. Whichever one should be fine. However, for focus *tests*, it's good to use one focus points because it eliminates the possibility of the camera picking a point that isn't really in the same plane as your intended subject. An example would be focusing on a human face -- the camera might think that the ears are a good focus point, so it would pick those points -- but when you look at the pictures afterwards, all you'd see is that the face was soft and the ears would be prickly sharp. Choosing to focus on some part of the face (eyes, for instance) using one focus point would tell the camera not to choose on its own.
    I saw the focus and recompose article, but the key takeaway remains true for most aspects of photography -- there are some things that work and others that don't, but few techniques work for (or against) everything. As the author mentioned, if you're shooting across a football field, the effects of F&R would be less seen than if you're doing portrait photography. In landscapes, there should be enough contrast for the camera to pick out an optimal focus without needing to focus and recompose.
  37. Thanks again for the responses. Rainer, will do those outside tests, but too late tonight, will try it tomorrow. Xingyuan, left the default settings, did the 10x zoom with LV and the centre point. Get back to you later. Many thanks.
  38. Gadzooks! Go find another Canon user or check a local photo club. Ask for help checking your camera out. Put your lens on the other guy's camera and see if you get the same results. Or go to a camera store and ask to compare another camera and lens to yours.
    Whatever your problem is, most cameras do not exhibit the same. Your posted images look to me to be soft overall, as if sharpening was not applied. You may have a lemon, but is it lens or is it camera, only putting your lenses on another camera and other's lenses on your camera will tell the tale. If all images come out soft, then it will be both, if one set is better, then you will have the culprit.
  39. could be backfocus and most likely the lens unless you have a camera that has microadjustment I'd suggest returning your lens and going to the store with the focus chart and test 5 lenses and buy the best one with the least amount of variation.
  40. Hello, sorry to be so long replying, but have been getting used to the camera, reading manual, practising, etc. Can't get outside to try "real world" pics as has been raining for days. Have been reading lots of other posts and taking advice from there. I still do think my lens is front focussing, is this a real problem and worthwhile returning the lens? Adding 2 images to show. Many thanks for the patience and help.
  41. Back again, focussed on the LG sign. Here is the whole pic. Thanks for any comments.
  42. What were the lighting conditions for the pictures? The second picture was taken at 1/6s (it seems from the EXIF data), which might be a little fast for handheld (even with IS).
    Also, what AF did you use?
  43. It was natural light coming through the window immediately to the left of the phone. I used single point on the LG name. I agree, could be some camera shake there. I really hate to keep on about this, do you think I'm paronoid about the focus bit? Here is another "real life" one. I used f8 @ 1/60 @ 200 ISO @ 55m, AF using single middle point and focussed on the red car.
  44. and the crop
  45. Thank you everyone for your answers. I will persevere with the lens as I have taken a few sharp pics and believe driver error has a lot to do with it. Will post a few decent pics later on and hope to read some comments about them. Thanks heaps again.
  46. Peter, I must admit that I am quite impressed by your decision to persevere -- I am very much in the same boat and look to do the same, but I haven't yet convinced myself that there's nothing wrong.
    Hopefully the weather will cooperate this weekend and let me get some photos in...

Share This Page