I don't know what is wrong...help!

[benstanley]

<p>I've been really unhappy with the sharpness I've been getting in a lot of my photos. Probably only 1/4 of them are sharp. This is in situations where I would expect a much higher rate, and I think I am having focus problems rather than camera or subject movement. In some cases I can see that the focus is in a different place to where it should be. I'm getting very unsharp photos even at 1/500+ with IS... I don't know if its the camera, lenses, or technique. I use centre point focus in general<br>

I'm not really a beginner but it sounds like a beginner question doesn't it...<br>

Please see the 2 photos:<br>

http://users.tpg.com.au/liznben//17mm.jpg<br>

http://users.tpg.com.au/liznben//55mm.jpg</p>

<p>These are with 17-55mm/2.8 IS on 60D, tripod mounted, without moving the camera between shots. I used only centre focus point, IS was off, and the camera film plane was as parallel to the target as I could get it. The target is flat, give or take 1-2cm, and I was about 1.5m away.</p>

<p>Notice how in 17mm.jpg, the right hand side is blurry. I get quite a few of these at random times. <br>

Notice how in 55mm.jpg, the left hand side is a touch blurry wheras the right hand side is sharp.</p>

<p>What is going on?!</p>

[benstanley]

<p>ps. the above 2 shots are with MLU, 2 second shutter delay, flash lit with 430exii, overpowering ambient light.</p>

[hjoseph7]

<p>If you are using center point focus then the sharpest area is going to be located in the center. If you want the left corner, or the right corner to be in focus then you need to move the focusing point to those areas.</p>

[justin_stott]

<p>Remember your plane of focus is not necessarily really a plane, but often somewhat of a curve made of points that are equally distant from the camera. At f2.8 it would not be unusual for there to be lack of focus at the edges of a flat plane this size 1.5m away.</p>

<p>As to why your out of focus areas are on opposite sides at each end of your zoom range, does a lens element rotate during your zoom? (Sorry, I don't have experience with this lens.) </p>

<p>After a little search it seems that this is a pretty high end lens, it may not hurt to have it checked out while it's still under warranty. In the mean time, stop down a couple of stops for extra depth of field, and hope someone more knowledgable replies!</p>

<p>Good luck!</p>

[Tim_Lookingbill]

<p>If you REALLY established perpendicularity of the lens to the surface in your linked shot then you shouldn't have the sides out of focus while the center is in focus.</p>

<p>If this was established then you need to get a better quality lens or get another copy of your current kit lens because there are differences in manufacturing within the same model. I've seen this with my own Pentax K100D 18-55mm kit lens where my sharpness results were much better than yours even hand held where someone with the same model lens and camera as mine got similar results as yours.</p>

<p>Another thing to consider is turning off IS when using a tripod. That's what's instructed in my camera's owner's manual with IS in the body of the camera and not on the lens.</p>

[derek_hofmann]

<p>I see three problems with the 17mm image:<br>

1. It was shot wide open at f/2.8. You won't get the sharpest results from a lens at its widest aperture, especially (but not only) at the edges, even if your focus is perfect. Try f/8 instead.<br>

2. The image is heavily compressed. It's 2.4 megabytes, when it should be around 8-9 megabytes from that camera. Better yet, shoot RAW and post a PNG.<br>

3. It's slightly underexposed. The lack of contrast may make the photo look not very sharp.</p>

[john_deerfield]

<p>It is not uncommon for <em>any</em> lens to be "weak" on the corners when used at the wide end of the zoom. And especially wide open. The 17-55 is no exception:</p>

<p>http://www.photozone.de/canon-eos/178-canon-ef-s-17-55mm-f28-usm-is-test-report--review?start=1</p>

<p>Using a UV filter can also exaggerate the effect. Also, if you are shooting from a tripod, IS should be OFF. Even if you repeat the experiment with IS OFF, an aperture of f/4, taking off any filters and raising the tripod (you look a little below center to me), the corners of the image are probably never going to be as sharp as the center. And, you are going to reach a point where the camera/lens simply <em>can't resolve </em>what you want it to. Here is a link to medium format camera review:</p>

<p>http://www.luminous-landscape.com/reviews/cameras/phase645.shtml</p>

<p>If you scroll about 2/3rds of the way down there is a full frame image with a 100% crop on text. Now that is some resolving power. But you aren't going to get than from a cropped sensor camera!</p>

<p>Finally, don't forget that our eyes can play tricks on us:</p>

<p>http://www.pixiq.com/article/visual-perception-and-photography-2</p>

<p>Yes, the corners are soft and there is some vignetting happening. So now we see the corners as even weaker.</p>

<p>All-in-all, I don't think there is anything wrong with the camera/lens. There are a few things you could do to improve the test image (as outlined), but at some point you may want to shoot wide open, without a tripod... at which point I might suggest adjusting your expectations!</p>

[Mattman944]

<p>Expanding on what Justin said. It is the goal of a lens designer to have the area of sharp focus to be a plane. However, there are many compromises in zoom design. Minimizing barrel/pincushion distortion, flat field of focus, size, weight, cost, etc. Compromising on the field of focus flatness is a parameter not normally noticed by the buyer or even the user. The field of focus can be slightly concave or convex and may change as you zoom. I can't be sure this is what you are seeing, but it is true.</p>

[William Michael]

<p>The Stock Exchange is always blurry at the edges, Ben.</p>

<p>***</p>

<p>Seriously:</p>

<p>The two test shots are very likely in not related very much, to the problems you are having in the field.<br />Taking the test shots firstly – as mentioned zooms are a compromise of design (all lenses are, zooms more so). You have chosen the two extremes of the zoom where the lens is arguable weakest.<br />At 17mm you used F/2.8 and at SD = 1.5m: those specs and even 1° shooting error would result in an issue with DoF at the edges, not to mention the fact that you should expect the lens to be softer at the edges and also that the Plane of Sharp Focus is NOT a plane anyway, especially at FL = 17mm</p>

<p>The same applied to the 55mm, but the shooting angle is more relevant as for example 1° out of plumb would be even more magnified apropos the resultant working DoF at that Shooting Distance, again considering the Plane of Sharp Focus is not a Plane, anyway.</p>

<p>I think analysis of real example of the “problem", will reveal other issues that describe your ill content.</p>

<p>The first aspect I would look at would, be the AF MODE you are using: especially if you use AI Servo or AI Focus often.<br />It is necessary to establish if there is a commonality apropos shooting specs and or shooting conditions with the shots with which you are unhappy.</p>

<p>***</p>

<p>I note this is NOT the kit lens.<br />I note that IS was turned off.<br />I also re-iterate that the Plane of Sharp Focus is <strong><em>not a</em></strong> Plane.<br />Also at 17mm at F/2.8 I would not expect the edges of the lens to be as sharp as the centre when shooting at 1.5mtrs.<br />I have used thisn lens.<br />The vignette on ten 17mm shot, is appropriate also.</p>

<p>***<br>

You do not state that you re-focussed between shots, but if you did not, you should have.<br /><br /></p>

<p>WW</p>

[benstanley]

<p>To answer a couple of questions:<br>

-IS was off (tripod mounted)<br>

-Yes I refocused after zooming between 17/55<br>

-For these tests, AF was One-Shot and I normally use one-shot in actual use. I have tried AI Servo for kids who are moving a bit but that rarely seems to give good results either<br>

-I would presume some lens elements do rotate, because the lens needs to refocus after I change the zoom. In fact, it has to refocus *a lot*. This lens is a very, *very* long way from being parfocal. At 17mm the distance window is totally useless. Infinity becomes 2-3 meters.</p>

<p>So...<br>

- I'm definitely not expecting the edges to be as sharp as the centre especially at f2.8. <br>

- I know that the plane of focus varies in shape between lenses, but I would expect that it would be symmettric, so if one edge is out of focus the other will be too. But in the 17mm shot one side is very different to the other. Also... consider the DOF below. Its almost 90cm deep.<br>

- Obviously it would be easy for me to not be shooting exactly flat with the surface (I did my best), but on the other hand at 17mm @ 1.5m @ f2.8 gives me a depth of field starting 32cm in front of the focus point and 57cm behind! Pretty sure I wasn't that far out!! And if I was shooting at an angle, I would expect one edge to be back focused, the centre to be sharp, and the other edge to be front focused. But what I have is a sharp centre, the left side is showing natural f2.8 blurriness but seems to be in focus, and the right side is very messed up.</p>

<p>I'm just disappointed I'm getting so many shots that aren't sharp. I get the occasional one which makes me go "now that's the way its *meant* to be" which makes all the others look bad. In some cases, it is so unsharp, post processing sharpness has no effect because there's no edges to sharpen!!! My 70-200/4L IS has a much higher keeper rate, and I'm not talking about small differences in sharpness here. Its really annoying that I've done my best to get reasonable gear, and either I (or the equipment) is not really performing that well, and its frustrating because I can't really say for sure what the issue is...</p>

[benstanley]

<p>ps. Yes the image was originally 8-9mb jpeg, I just resaved it in photoshop with quality 9 or 10 instead of 12 to upload. Normally I shoot raw, but for this test it didn't seem necessary.</p>

<p>Just to reiterate - I don't mind it not being as sharp in the edges/corners, that is normal. Its the fact that one side, and one side only is really really bad.</p>

[scott_ferris]

<p>I personally think there is an element misalignment or internal issue of some sort with your lens.</p>

<p><em>"Remember your plane of focus is not necessarily really a plane, but often somewhat of a curve made of points that are equally distant from the camera."</em> This statement is quite simply wrong, the plane of focus is a flat plane perpendicular to the sensor, unless specialised tilting lenses are used. Some cheaper lenses, and very old, non planar lenses, do suffer from a noticeable amount of field curvature, but the 17-55 is not one of them. At 1.5 meters and f2.8 the newspaper should all be in focus, not as crisply as the center, but much better than the posted images. The fact that it is repeatably bad on one side and that changes with focal length can really only mean there ia a lens issue.</p>

<p>I would suggest repeating your tests, once set up rotate the camera to get the center and the right side (at 17mm) sharp, see what that does to the left side. Keep making small changes to the camera rotation and then when then 17mm images are as sharp as you can get them zoom to 55 and do it all again.</p>

<p>It is a shame the 60D does not have micro adjustable focus, but I am not sure it would help in this case, however modifying <a href="http://www.northlight-images.co.uk/article_pages/cameras/1ds3_af_micoadjustment.html">this technique</a> would help to eliminate possible misalignment between camera and target. If the interference circles are bigger on one side either you have a lens issue or the camera is not perpendicular to the screen, if you have to move the camera to get the best, and equal, image at the different focal lengths you have a lens issue. Of course you can do all the testing without actually doing, or having, micro focus adjustment.</p>

[William Michael]

<p>Scott are you absolutely sure about Plane of Sharp Focus of the 17 to 55 <strong><em>at 17mm</em></strong>? <br />I am talking only fractions of mm here of it NOT being a perfectly non planar lens - adde d to that the camera being only a few degrees off perpendicular.</p>

<p>I agree 100% with your re test procedure </p>

<p>Regards and Happy new Year, Scott</p>

<p>WW</p>

<p> </p>

[scott_ferris]

<p>For a 60D and a 17mm lens........<br /><br />Subject distance 1.5 m<br /> <br />Depth of field <br />Near limit 1.18 m<br />Far limit 2.07 m<br />Total 0.9 m<br /> <br />In front of subject 0.32 m (36%)<br />Behind subject 0.57 m (64%)<br />Circle of confusion 0.019 mm</p>

<p>So 0.9m or 35" of DOF, that should easily cover a small misalignment or focusing errors.</p>

<p>Very warmest wishes for the New Year to you too William, have a great one.</p>

<p>Scott.</p>

[William Michael]

<p> HA HA. very nice!</p>

<p>I didn't think of you answering my question that way. I didn't even think of playing the actual DoF figure through and the relevance of them.<br>

My Excuse - the brain is still in holiday mode, I will need to sharpen up!<br>

Your Point is taken.</p>

<p>WW</p>

<p>(and the Plane of Sharp focus is a little tiny bit curved - but not much . . . :)<br>

2011 score card:<br>

Scott 1 :: Bill 0)</p>

 

[john_deerfield]

<blockquote>

<p>To answer a couple of questions:<br />-IS was off (tripod mounted)<br />-Yes I refocused after zooming between 17/55<br />-For these tests, AF was One-Shot and I normally use one-shot in actual use. I have tried AI Servo for kids who are moving a bit but that rarely seems to give good results either<br />-I would presume some lens elements do rotate, because the lens needs to refocus after I change the zoom. In fact, it has to refocus *a lot*. This lens is a very, *very* long way from being parfocal. At 17mm the distance window is totally useless. Infinity becomes 2-3 meters</p>

 

</blockquote>

<p>How about the question: is there a UV filter on the lens? If so, take it off and see what happens.</p>

[john_tran14]

<p>Try manual focus to see if the AF focusing is the problem or not. If you can not do manual focusing well, take many shots with little by little changing focusing. If one of the MF photos is much better then you can focus only on AF</p>

[benstanley]

<p>Sorry, no there's never been a filter on the lens I don't use them.</p>

<p>I can't MF better than the AF. Live view focussing seems about the same as well.</p>

<p>I'm going to take it back to the shop today to compare with another one.</p>

<p>I did some more experimenting last night at different distances with different lenses, and noted down any problems I found. In summary<br>

- My 50/1.8 is much more blurred on the right side right up to about f8, particularly when focussing on something further away. Above f4 the rest is nicely sharp.<br>

- At this point you'd start blaming the camera (or me) but my 70-200/4L at any focal length does not have the problem, my 17-40L is a touch softer on the left not the right at 40mm and a touch softer on the right at 40mm, and my 100/2.8L does not have the problem. The camera was sturdily tripod mounted and not moved when switching between any of these lenses. Although I still often miss focus with these lenses (one-shot, center af point, most contrasty point of target - eg I usually focus on the eyes or is this wrong?). And I've now done the tests on 2 different occasions with the same results.<br>

This time around I also used a tape measure to ensure the camera was at the same height as the target point this time, and as close to centered as possible.</p>

<p>Scott I'd like to try your test, but I'm not sure what I should expect to happen?</p>

[scott_ferris]

<p>Ben,</p>

<p>What you should see, if your camera and monitor are perfectly aligned, and your lens is ok, is an even pattern of interfering circles across the width and height of the screen. If they are not aligned, but your lens is ok, then you will get an uneven pattern across the monitor, but it will be equally bad both sides of center if your camera is pointed directly at the center, it should be close to a mirror image either side of center. If your lens is out of alignment internally, you will see an uneven distribution of circles, the areas that are more out of focus will be more heavily patterned, despite being centered and aligned with your lens, the monitor, will show more patterning on the right hand side at 17mm if your lens is damaged.</p>

<p>Hope this helps, Scott.</p>