70-200 IS2 sharp? Please help

A

adam*

Messages
916
Name
Adam
Edit My Images
No
Hey everyone,

Just got myself a new 70-200 is2 and after hearing so many stories about how amazingly sharp it is i'm not too sure about my copy - or i'm just being too picky. This is a focus adjustment chart shot after being adjusted +7 which seemed to be best on a 5dmk2. Also a shot of the set up to check it's all as should be. Is it odd that the left side isn't as sharp as the right? ? (Focus on the H of here)

original.jpg


original.jpg
 
It does look a bit off. What's it like taking pictures of real things?
 
Sharp in the middle at f/2.8 and decent when focusing to the right but not as good on the left if I focus with the outer point wide open.
 
Time for a replacement IMHO....
 
Although re-testing focus-recomposing it does appear pretty decent at 100% full size image hmm!
 
If its new and you are sure the Camera was squaare to the test page I'd print out the results you have above and take it into the store and ask for a replacement.
 
It's brand new, purchased through my agency but it's mine. Thankfully the shop is about an hour away so I might try pop in and hassle them.
 
There is lots wrong there, probably none of it down to the lens.

You are using a test for focus accuracy to check for a good copy. That particular test is flawed and notorious for throwing up spurious non-existant errors. On top of which, you're clearly not square to the target, there's more depth of field on the left, and what you're seeing is most likely a light reflection issue.

Do it properly before jumping to conclusions. There are numerous threads on AF microadjustment.

Whether you have a good copy or not is another question, which would not show up on that test unless the lens was badly damaged - it's hardly ever visible in the centre, only around the sides and corners. A quick test procedure for that is here, post #20 http://www.talkphotography.co.uk/forums/showthread.php?t=317381
 
Ok here are a few more 100% crops at f/2.8. Focus is on the A and placed in various areas of the frame. All taken at 200mm at the same distance from subject under even overcast light. Seems to be a clear difference in sharpness to me? I've done more tests focusing on smaller lettering - the centre is easily the sharpest again, the right it sharp, and the left side is like below - a bit soft i'd say?

Middle:
134642904.jpg


Top Left:
134642908.jpg


Bottom Left:
134642910.jpg
 
Last edited:
That was real world shooting of a box in my back garden, just as if I was shooting a portrait of someone in their back garden which I would call real world shooting. I'm not one for tests, never have been, but for all the rave reviews saying this is Canon's sharpest ever zoom etc i'm frankly not that impressed.
 
Obviously the above may be acceptable for some but I just spent alot on the lens and if it isn't perfect i'd rather go swap it whilst I can.
 
The problem is that I use this every day for work and it will probably become 'used' condition very quickly meaning I won't be able to take it back / i'll personally have to pay for it to be fixed.

My job may require me to shoot a box in a back garden as well.... ;)
 
Go to the garden (or the front room which seems to have good light), setup some props (include a newspaper or magazine, etc). Take some snaps from both focal ends and check back. :)
 
adam* said:
Sharp in the middle at f/2.8 and decent when focusing to the right but not as good on the left if I focus with the outer point wide open.
Click to expand...

Are you saying the lens is not as sharp using the far left focus point in the viewfinder?
 
Admirable - no sorry, it's the same if I focus / recompose as well.

mith: that was done in the back garden, 200mm is the critical point for me.
 
With the camera on a tripod, IS disabled, f2.8, using a timer or remote shutter release and using the centre focus point only with the subject square on to the camera, the images are soft on the left hand side compared to the left?

Have I got this right?
 
adam* said:
The problem is that I use this every day for work and it will probably become 'used' condition very quickly meaning I won't be able to take it back / i'll personally have to pay for it to be fixed.
Click to expand...

Not sure about that... Canon would still fix it in warranty if there was a fault.
 
admirable said:
With the camera on a tripod, IS disabled, f2.8, using a timer or remote shutter release and using the centre focus point only with the subject square on to the camera, the images are soft on the left hand side compared to the left?

Have I got this right?
Click to expand...

Yes
 
RichardtheSane said:
Not sure about that... Canon would still fix it in warranty if there was a fault.
Click to expand...


They could but you can imagine the hassle of being without my main workhorse lens for 2-4 weeks whilst it gets fixed... compared to going to the shop today and asking for a replacement.
 
Take it back to the shop.
 
Adam, there may be something wrong with your lens, but your tests are not showing it conclusively. You DoF is measured in mms, yet you are clearly not square to the box! Depending on exactly how you're doing it, that's probably what you would expect to see.

Why don't you follow the procedure outline in the link I posted above, and use a distant subject? The distance is to reduce focus error as far as possible. You need to get the test right because if you go and get another one, you'll find exactly the same issues.
 
If you focus on something at f/2.8 off centre and to the left of frame and it's not sharp then it's not sharp... however many times you test it? I'm concerned if it's sharp across the frame at 200mm not 130?
 
I'd echo what Hoppy says. These tests are difficult to set up accurately. If you've focused on 'Here' which is near the centre of the image, then the camera and/or the test chart only need to be skewed a millimetre or two from parallel to each other to throw one side slightly oof and it is only very slighlyt out.

You'd virtually need micrometer accuracy to set that test up properly. I seriously doubt there's much wrong with your lens - I'd question the test rather than the lens.
 
Last edited:
adam* said:
If you focus on something at f/2.8 off centre and to the left of frame and it's not sharp then it's not sharp... however many times you test it? I'm concerned if it's sharp across the frame at 200mm not 130?
Click to expand...

How do you know what the camera is focusing on? We can't see that from what you've posted and with the target at an angle and DoF at a few mms, there's no way of telling if its the lens or camera or user error.

Do the test at 200mm! As suggested on the link. You need to eliminate the potential for focus error. Follow the procedure exactly, it takes two seconds, and you'll see anything serious right on the LCD. Be careful that the area you check is exactly the same distance from the edge/corner each time.

Edit: why is the box in the first image so much bigger? :thinking:
 
Last edited:
OK here's your test, which is worryingly better. Opinions? Middle still definitely sharper.

Distance:
134643800.jpg


Middle:
134643801.jpg


Top Right:
134643804.jpg


Bottom Right:
134643798.jpg


Top Left:
134643799.jpg


Bottom Left:
134643796.jpg
 
Last edited:
I may be wrong but isnt the middle usually the sharpest point in the frame? The sides seem only slightly softer. Also the outer frame shots appear magnified more.
 
They're all 100% crops in photoshop. Usually but i've been told the 70-200 IS2 is just as sharp in the corners at f.2.8
 
adam* said:
OK here's your test, which is worryingly better. Opinions? Middle still definitely sharper.

<snip>
Click to expand...

Lens is fine :)

Yes, middle is better, as you would expect. No lens is ever equally sharp all over, especially at lowest f/number. If you did that test with a lesser lens, the corners would be miles behind but that's not the point. The point is, are all four corners consistently and equally sharp (or quite possibly unsharp). You're not testing for sharpness, you're only looking for equality to prove that the lens has been screwed together 100% correctly centred, ie a good copy.

Now check the AF with micro-adjust. Numerous threads on that, including this one http://www.talkphotography.co.uk/forums/showthread.php?t=310467&page=2 Try the cerial box with an angled ruler :thumbs:
 
adam* said:
They're all 100% crops in photoshop. Usually but i've been told the 70-200 IS2 is just as sharp in the corners at f.2.8
Click to expand...

You've been told wrong Adam. Lenses are typically sharper in the centre and softer towards the edges and corners. The situation improves somewhat as you stop down, but expect the poorest edge results at wide apertures - especially f/2.8 as in your case- that's a pretty big aperture for a tele lens.

You're also using a full frame camera which uses the whole image circle to project the image onto the sensor. A crop sensor is somewhat better in this regard as it only uses the centre part of the image.

Stop worrying and enjoy your new lens. I have the same lens and it's a belter.
 
HoppyUK said:
Lens is fine :)
Click to expand...
Agreed. 99% of the "My lens is soft/faulty" threads are really about the testing technique rather than the lens, and this is another such.

Adam, go out and enjoy using your lens. You won't regret it!
 
Cheers, good to have my mind put to rest :p I for one hate all these threads and then I end up with one myself haha. Just a case of changing over to Canon, being picky, spending a hell of alot of money and hoping it's all as expected.
 
The lenses which are sharpest edge to edge and corner to corner are macro lenses and they're very carefully and no doubt expensively ground and polished to produce those results, the reason being that apart from photographing bugs, they're used for photographing text documents, old line drawings etc, and also used for medical and scientific applications where edge to edge sharpness is very important.
 
adam* said:
Hey everyone,

Just got myself a new 70-200 is2 and after hearing so many stories about how amazingly sharp it is i'm not too sure about my copy - or i'm just being too picky. This is a focus adjustment chart shot after being adjusted +7 which seemed to be best on a 5dmk2. Also a shot of the set up to check it's all as should be. Is it odd that the left side isn't as sharp as the right? ? (Focus on the H of here)

original.jpg


original.jpg
Click to expand...

With the greatest of respect I question the integrity of the set up for the initial test. You can not be certain you were head on, a sofa cusion is not a typicaly flat surface..... I would suggest a more refined test process before passing judgement. This does not neccassarily mean "rear world" shots. From what I have read about setting up for this sort of micro adjsutment I'd suggest the camera is quite close to the test subject as well. I don't have my reference material with me as I'm in the office but I think it is recomended you set the lens to max zoom and position at max zoom x 50 from the target. In this case 10000 mm away, or 10 meters.
 
CT said:
The lenses which are sharpest edge to edge and corner to corner are macro lenses and they're very carefully and no doubt expensively ground and polished to produce those results, the reason being that apart from photographing bugs, they're used for photographing text documents, old line drawings etc, and also used for medical and scientific applications where edge to edge sharpness is very important.
Click to expand...

Macros are a good example for edge to edge sharpness. As a rule, longer lenses tend to be better at it, and the worst are wide angles, wide zooms, and particularly cheaper wide zooms, like kit lenses at the short end.

Kit zooms will probably be pretty damn sharp in the middle, but will fall away dramatically in the corners. With higher quality lenses, it's at the edges and corners where they score (aspherical element surfaces help a lot). Both will get a lot better as the f/number is raised.

If the OP tried that same test with an identical spec 70-200 2.8 Sigma zoom, those corner images would look very different, even if the centre was probably probably similar. But then the Sigma costs half as much.
 
You must log in or register to reply here.
Back
Top