PDA

View Full Version : New Lens is NOT sharp around the outside of the picture ? Please help



PerfectPicture
17-01-2011, 1:20pm
Hi All,

Recently purchased the Canon 10-22mm Lens from Citiwideonline last Wednesday and received it on Saturday (after driving down to the Derrimut depot in laverton). Turnaround time is quite good from O/S.

Went down to Williamstown and took some pictures, seems ok, but got home and noticed the centre focus seems to be OK? BUT all around the outside, like the left-hand side and right-hand-side its blurry??
And took pictures in different modes and spot metering on the day as well.
Is this a common thing with UWA lens?
Does it need micro-adjusting to get right?
Or .. did l get a dodgy lens perhaps?
If so .. what are my recourse of actions?? (l'm writing up an email to citiwide now, should l request another lens?? OR .. just request my full money back??).
geez..what a buggar!

Example of pictures to follow soon ...

Cheers
Robert

kiwi
17-01-2011, 1:46pm
should be sharp edge to edge, sounds like one of the elements is out of alignment - id return

PerfectPicture
17-01-2011, 1:58pm
should be sharp edge to edge, sounds like one of the elements is out of alignment - id return

Thanks Kiwi :)

should l ask for a replacement or try to get full money back, and try elsewhere?

kiwi
17-01-2011, 1:59pm
Up to you. Id ring them/email them and see what they say

PerfectPicture
17-01-2011, 2:57pm
Up to you. Id ring them/email them and see what they say

Thanks Kiwi, l sent an email to citiwide explaining the situation, and requesting to send this lens back to them ASAP and get a replacement OR my Full money back.
Now l have to wait for their response.

Have other people used Citiwide .. and are they reputable and do they provide a replacement or Full refund back if requested.

Has anybody else bought Lenses from overseas and if so, have they being SPOT On and clear in focus or dodgy first time around ??

Art Vandelay
17-01-2011, 3:01pm
Has anybody else bought Lenses from overseas and if so, have they being SPOT On and clear in focus or dodgy first time around ??

yep, no problem.

It's not an overseas vs local purchase issue. They're all from the same factory. Occasionally they'll spit out a dud.

Interested to see some samples.

PH005
17-01-2011, 4:07pm
I'm sure it's just a case of one slipping through the QC. Hope you get a satisfactory outcome very quickly. Keep us posted.

arthurking83
17-01-2011, 4:39pm
should be sharp edge to edge .....

Not nececellery!!

Robert!! .... the problem with diagnosing such an issue is that you haven't given any details as to what you did, what settings, what the scene is/was ... etc.

Best thing to do before you go an create yourself some fuss over nothing, is to take a few test shots and post them with full exif intact.

All UWA angle lenses suffer from massive distortion(except the Nikon 14-24/2.8) and therefore have a fair amount of softening at the image edges.. that is compared to the centre of the frame.
The other very well designed UWA lens of note is Sigma's new 8-16mm(APS-C only).

If you look at any test example of this particular lens, you'll see that the difference in sharpness of the centre of the frame compared to the edges can be as much as 50% difference in resolving ability.
eg. check Photozone's test results and notice that the centre of the frame is in the excellent range in terms of sharpness, whereas the borders are only struggling in the good range.

Also as I said before, you haven't yet explained what camera settings you've used ... in particular aperture value. Also what is hard to determine from any image that you post, is field curvature(which all lenses will have to a degree) and UWA are notorious for field curvature. So! if you simply shoot a brick wall even tho your focused correctly at the central spot, the issue of field curvature is going to place the outside edges of the frame well out of the plane of focus(stopping down a lot helps with that). if you fully understand/measure the amount/degree of field curvature that the lens is burdened with, you can take a perfectly sharp image across the entire frame of a curved subject.

Post a few sample pics(remember with full exif intact) and see how you go. Don't worry about going back to the store yet, until you are armed with all the relevant proof that this lens is faulty(unless you have time to waste).

PerfectPicture
17-01-2011, 4:40pm
thanks for responses guys ... here are a few examples >
HERE (http://img441.imageshack.us/img441/5177/mg3034b.jpg)
the centre of the image seems clear, but the ousides of it, is blurry? p.s. l'm using a CPL filter as well.
Just, click on the images to see larger file output.

and > HERE (http://img543.imageshack.us/img543/3443/mg3023b.jpg).

This one is the Williamstown beach, middle is OK, but near the horizon on both sides its blurry? > HERE (http://img257.imageshack.us/img257/6769/mg2976a.jpg)

kiwi
17-01-2011, 4:58pm
i agree with ak re distortion effects as opposed to corner softness.

both the examples you have posted will be more likely, and i actually think are more due to distortion becuase of the nature of what you are shooting here - i dont think anything in the horizon is sharp in #2 anywhere....focus seems in the front 3rd of the photo

more testing i think

William
17-01-2011, 4:58pm
No probs mate, Thats the way Wideys work, I have a Sigma 10-20 , Does the same thing around the edges at 10mm , There is a Technical word for it (Not Distortion) , But it escapes me now , They all give distortion at full wide , They can be a bit of a beast to get used to :) I think you'll find that thats the nature of the beast as they say :th3:

PerfectPicture
17-01-2011, 5:09pm
Thanks for the responses guys :)

So the sample pictures are fine to you guys??

What is the BEST WAY to run a test on a new lens to confirm sharpness, softness, etc?
also does it need micro-adjustment do you think? as centre seems OK?

l'm perplexed now.

peterb666
17-01-2011, 5:15pm
CPLs and ultrawides don't go together too well as the angle of light towards the edges is very different to that in the centre. If you want to check for sharpness or any other image defect, you need to do it without filters.

That said, the image sharpness does seem to be off but I am not familiar with the Canon lens so I will leave it to others to confirm or otherwise.

kiwi
17-01-2011, 5:37pm
Yeah, I'd take filter off and do the good old brick wAll test, tripod mount etc

soulman
17-01-2011, 6:31pm
i dont think anything in the horizon is sharp in #2 anywhere....focus seems in the front 3rd of the photo The first two look the same in this regard - focus point is quite close and nothing is sharp in the background. There is no EXIF on the images so I can't be sure, but it looks like you're shooting wide open or close to it. If this is the case then the lens could well be normal. The one I had needed to be at f8 or f11 to be guaranteed reasonably sharp across the frame. That's just the reality with UWA zooms. As noted by Arthur, resolving power at the edges of these kinds of lenses is never anywhere near what it is in the centre. You can make sharp pictures with these lenses but you have to work at it.

I @ M
17-01-2011, 6:43pm
I wouldn't concentrate too much on those images with the filter on. They don't look at all good on the edges but until you have given it a good test there can't be many conclusions drawn.
The one thing that does stand out badly is the amount of purple fringing in the first shot in particular.



There is no EXIF on the images so I can't be sure, but it looks like you're shooting wide open or close to it.

kuso exif viewer is showing F/7.1 for the first and F/8 for the second.

Allann
17-01-2011, 6:58pm
The 10-20 does suffer from soft edges up to about 5.6 at 10mm, but not as badly as shown here, especially since the first was shot at f7.1 and the second at f8. This style of shot would be best taken at about f11 however. The third however at f3.5 has produced blur at the horizon due to the shallow depth of field. Notice in that shot too, that the bottom corners are not as blurry as the first two.

I would suggest you try removing all other obstructions from the lens, i.e. ALL filters, and take some test shots. You can find a review here (http://photo.net/equipment/canon/efs_10-22/) which shows what to expect. If after doing your tests again, and them failing, then I'd contact the vendor.

arthurking83
17-01-2011, 7:19pm
....

What is the BEST WAY to run a test on a new lens to confirm sharpness, softness, etc?
also does it need micro-adjustment do you think? as centre seems OK?

l'm perplexed now.

for a quick test, to determine if the lens is decentred or not:

brickwall, using liveview(is best!!) focus on about 2/3rd to 3/4's of the frame edge. Live view should allow you to focus over any part of the entire scene.

Of course you'll be on a tripod, and you need to make sure that the camera is as squared up to the brick wall as possible. a millimeter out won't make a difference, but anything more than about 10° or so not square to the wall is going to show up as DOF issues, and not lens abnormalities. 10° out of whack is quite easy to notice by eye.

Try this at a few focal lengths.

Also note, with a UWA lens you do capture a massive field of view, so the temptation is to get closer to encompass only the brick wall, this also has side effects, where some lens designs work better at far distance then they do at closer focused distance, and vice versa.

if you do this test, not that some softness on one side can be easily seen relative to the other side of the frame, and this may be due to not perfectly parallel alignment.

Stop down to f/8, f/11 and f/16 and review the images, leave the camera where it is(if possible) and if there is any difference from one side of the frame compared to the other, move the camera to try to get a more parallel position.
That is, if the LHS is soft and the RHS(where you've focused) is sharper, then rotate the camera more towards the right(by a degree or two). Do the test again.. as so on and so forth.

PerfectPicture
18-01-2011, 12:11am
Thanks alot everybody ... l really do appreciate it.

Question, should l leave the Hoya Super HMS Pro1 UV Filter on, or 100% make sure and take it off as well?

l will do some test shots today .. and post the findings later that evening.

Cheers all.

p.s. l found the lens hood that comes with the Canon 24-105L IS USM lens is perfect for this lens as well, No Vignetting at all .. have to look for a cheap place to buy one for this 10-22mm lens as well.

ricktas
18-01-2011, 5:53am
take all filters off. You want to test the lens, not the filter. And to be honest the UV filter shouldn't go back on the lens, you don't need it.

arthurking83
18-01-2011, 9:06am
.... And to be honest the UV filter shouldn't go back on the lens, you don't need it.

:th3:

especially an ultra widey!!
If you had to put up with any IQ losses due to filters, you need to make sure that the filter is at least producing a result that balances out the likelyhood that it will also degrade IQ.

At least with a CPL or grads or ND's or whatever other filter used for a specific purpose, you get some benefit in using it, so you weigh up the pros and cons and strike a balance.

UV filters serve no purpose.

Arg
18-01-2011, 12:15pm
Would like to see a shot of the horizon stretching from side to side, focused on the horizon using Live View, at f11 and on a tripod. 10mm, 15mm, 22mm. then again at max aperture.

peterb666
18-01-2011, 5:29pm
:th3:

If you had to put up with any IQ losses due to filters, you need to make sure that the filter is at least producing a result that balances out the likelyhood that it will also degrade IQ.

Quite frankly, geometry is the enemy of filters on an ultra wide. While a lens can be designed to work well at a paricular focal length or even a range of focal lenghts, the poor old filter hasn't got a clue what lens it is going on. Is it a 10mm lens with a angle of view of more than 90% or a 500mm lens with an angle of view of 5 degrees? Even the best filter will be a compromise.

At 10mm, a filer is typically 2mm thick in the centre and light going through it is at 90 degrees.

Move to the edge and light is going through at around 45 degrees, travelling through approximately 3 mm of glass with a entry and exit angle of 45 degrees in opposite directions.

A filter that serves very little purpose such as a UV filter should be put to better use. A transparent drinks coaster comes to mind. :D

arthurking83
18-01-2011, 10:35pm
.....

A filter that serves very little purpose such as a UV filter should be put to better use. A transparent drinks coaster comes to mind. :D

Alternatively:
The subject of a strenuous testing regime for the purpose of determining, once and for all, if they serve any purpose as a so called protective lens. :th3:

PerfectPicture
18-01-2011, 10:52pm
Thanks for the replies everybody .. OK slowly uploading the pics to Flickr acct.
l have placed them into SETS e.g. "Canon10-22mm @ 10mm".
They are all straight from the camera, Jpeg's compress from 6MB files to 47% with faststone image viewer to around 660Kb each.
Shot with NO Filters attached, Shot with a Tripod and Shutter release cable, shooting the red brick wall for testing purposes of clarity, Sharpness, etc of the newly purchased Lens.

First Set 'Canon 10-22mm @10mm at different F/Stops' is > HERE (http://www.flickr.com/photos/robert2010/sets/72157625851982354/)

Second Set 'Canon 10-22mm @17mm at different F/Stops' is > HERE (http://www.flickr.com/photos/robert2010/sets/72157625726488881/)

Third Set 'Canon 10-22mm @ 22mm at different F/Stops' is > HERE (http://www.flickr.com/photos/robert2010/sets/72157625852086242/).

l hope you can see them and their viewable for you guys. if not. let me know.

Please let me know what you think of the images in regards to clarity, Sharpness and overall testing of the new Lens.

Thank you.

Arg
18-01-2011, 11:14pm
First impressions are OK but I can't see the 660kb files, I can only see 1024-pixel-wide images.

PerfectPicture
18-01-2011, 11:43pm
example Pic's .. taken directly after the brick wall test.
All pictures again taken with NO Filters attached, Tripod + Shutter cable release using LIVEVIEW zoomed in at x5 ...

first pic > HERE (http://img529.imageshack.us/img529/9604/mg3463a.jpg) at image shack (Full size displayed).

Second pic > HERE (http://img87.imageshack.us/img87/2007/mg3516a.jpg)

and this one, focussed on the foreground rock (again doesn't seem clear to me) > HERE (http://img171.imageshack.us/img171/7819/mg3512a.jpg)

4th image > again not really clear > HERE (http://img199.imageshack.us/img199/4984/mg3521a.jpg)

Again blurry and not clear at all?

kiwi
19-01-2011, 7:18am
The wall tests seem fine to me with pixel peeping

The rock focus I think is simply an indication of oof areas in the background. Run dofmaster to determine wha should be in or oof and look at the concept of hyper focal to determine where you should be focussing to get sharpness throughout the image

William
19-01-2011, 7:44am
I agree with Darren , The Brick Wall images look fine , I find for Landscapes at 10mm it's best to use f8 to f10 and get your focus points to lock on to about 1/3 into the shot when composing , Or have a read up on Hyperfocal focusing

Allann
19-01-2011, 8:10am
As stated, I think we can all conclude that the lens is fine. Yes a little soft at the corners when wide open but thats to be expected.

So why are you still getting blurred photos in the landscapes you may ask? I will have a go at the first image if you want.


f22 will introduce CA in the image, try to stay around f11-f16 for landscapes, use a DOF calculator (http://www.dofmaster.com/dofjs.html) to determine both the focus point and the DOF needed.
1/20 will create motion blur on anything in the scene that is moving, e.g. the birds, the boat masts, clouds, waves in the water, etc. Giving an overall blurred appearance
foreground and even the clouds are well underexposed introducing noise which can also reduce the crispness in both


For a better test try taking shots at 90 degrees to the sun, not directly into the sun, this will also help boost the colours more as the sun (and flare) will wash out the image.

All the images basically fall into the same issues. underexposed and too high aperture. Since it was reasonable dark (assumed since high ISO and slow shutter was used) this introduces a whole new set of techniques to get a good shot and will take practice to get right. Try taking them late afternoon around 6pm instead of 9pm, the light will be better and you'll be able to use a faster shutter speed and lower ISO. Landscape photographers will use as low an ISO as possible, rarely leaving ISO 100.

arthurking83
19-01-2011, 4:50pm
As Allann said .. ISO at 1600 is not a good way to determine sharpness in images, and neither is f/22.

Minimum aperture value you use should be f/16, as that is generally the point where the human eye can clearly see diffraction starting to affect the images(when compared to say f/8 and f/11).

Other points worthy of noting. try to avoid harsh sunlight, high contrast induces chromatic abberations and some fuzzyiness along sharply defined edges(so the edges look softer than they otherwise may).
exposure also makes a difference. slight underexposure has an effect that is similar to sharper images. They're not really, but the illusion is such. As sharpness is highly dependent on contrast, more contrast looks like sharper images(that's what USM bases it's magic upon)
if in low light(better) use a sturdy tripod, mirror lockup and a remote, all of which help to eliminate user errors.

If you need any help, don't hesitate to send me a PM to arrange for us to meet. It has to be next week tho, as I'm busy with kids up till Monday morning. Anytime on or after Monday evening I'm free.

So far, from these images, I'd say your lens looks to be about on par(or maybe just below). It really is luck of the draw if you get the best copy or the worst made by Canon at any given moment.

PerfectPicture
20-01-2011, 8:43am
Thanks guys for your responses and comments, l really do appreciate it.
Was shooting in AV mode so l can shot all pictures in the whole range of F/Stops available on the lens. Must of had Auto ISO selected, that why got ISO1600. p.s. Shots taken around 7:30 til 8:20pm~, so getting low light as well.

Thank you Arthur for the meetup offer, l would love to catch up anytime your free in the next week or so, either after work (norm, 5:30pm onwards or the weekends). Thanks alot mate.

l have sent 2 emails to Citiwide, they have responded, with fill in a word document, serial no, address, issues, etc. done that.
They now want to see explains of pictures as well via email before sending lens back, so when l get home tonight l will do that.
Hopefully l can get my money back! Has anybody else experience this or got a full refund back for overseas stores?

p.s. whilst l was shooting the brick wall, l also tested my Tokina 12-24mm lens as well, and reviewing the pictures late last night, you can see texture in the brick work!! l'm amazed this lens is much sharper then the Canon 10-22mm one?
like you mentioned ... must of being one that slipped through Canon's Quality control centre, as heaps of reviews say this lens is a good sharp lens. ohh well .. hopefully issue is resolved quickly...

l'll keep you posted.

Cheers
Robert

kiwi
20-01-2011, 11:40am
I'm confused why you bought a 10-20 when you have the highly regarded tokina 12-24 ?

PerfectPicture
20-01-2011, 4:22pm
Hi,
l actually thought the Canon 10-22mm is actually better then the Tokina 12-24mm, so it was an upgrade. but its didnt work out that way. and found out the Tokina is actually a much sharper lens, after all ! :)

Nice , pleasant surprise! :)

Arg
20-01-2011, 4:38pm
Ahem!
http://www.bobatkins.com/photography/reviews/tokina_12-24_f4_review.html ;)

The Tokina is obviously a v nice lens but if it is sharper than the Canon then you do have a dud Canon.

The thing with ultra-wides is that every extra mm is significant in terms of photographic flexibility. That's the main reason to pick the Canon, to get an effective 16mm instead of 19.2mm.

kiwi
20-01-2011, 4:43pm
I just spent 10 minutes looking at side by side reviews. Not sure what you've seen but they are fairly unanimous in that the canon has quality sample variations. That the Tokina is very good and every bit as sharp and better built , but if you get -a sharp canon that is has a big less barrel distortion and purple fringing wide open.

So if I were you I'd see if they could swap for a similar value lens in a range you don't already have and keep the tokina

If not well, you have two Uwa lenses. Why ? I dunno

arthurking83
20-01-2011, 6:13pm
Both SLRGear and PhotoZone have tested both lenses and found their sharpness performance to be close enough to equal across the frame.
To call either one better than the other would be a case of variable quality control from the respective manufacturers.

if it came to a choice between the two, I'd test them both in the exact same manner, find each lenses strengths and weaknesses, and offload the one that doesn't perform as well.
Note that the one stop faster aperture of the Tokina at the tele end would weigh in heavily if it were my decision!! ;)(but then again that extra 2mm of width is also nice to have).