Lens sharpness test, please help.

footman

footman

Messages
4,941
Name
Laurence
Edit My Images
No







I bought this lens on ebay a few days ago. Took it to the park this morning and I fancied that it was soft. Trouble is, once you get that into your head you can't shift it!

I took these test shots in the garden, the light is fading here. Camera on tripod,remote release used, the coke can shot was mirror up, the book shot was normal and I can't remember about the toilet duck shot!

To me they still look soft but I'd like some fresh-eye views please. No PP done except for some cropping and conversion to JPEG in DPP.

I'm not saying which lens this is to avoid any preconceptions on your part.
 
100% crops or resized? Look OK to me (but I'm no expert). Perhaps there's a lack of contrast which contributes to a feeling of softness....
 
Looks like back focusing to me, based on the can.
Do you have a body which has MA?
Also, as I'm sure you know you need to do some sharpening when converting from RAW to JPEG, did you use your default sharpness setting?
 
User Name said:
Looks like back focusing to me, based on the can.
Do you have a body which has MA?
Also, as I'm sure you know you need to do some sharpening when converting from RAW to JPEG, did you use your default sharpness setting?
Click to expand...

This body is a 1D Mark IIN and my other one is a 40D neither of which do MA.
As far as sharpening goes I suppose the default would be whatever DPP is using in its "convert and save".
 
footman said:
I've never been totally sure what constitutes a 100% crop! They have all been cropped from the originals.
Click to expand...

A 100% Crop is basically opening the original file, zooming in to 100% then cropping a section of that with NO re-sizing of any kind :)

They look "OK" to me though in most photo's there room for a bit of PP sharpening IMO.
 
They do look a little oof to me but only very slightly. Try in better light conditions.
 
footman said:
I've never been totally sure what constitutes a 100% crop! They have all been cropped from the originals.
Click to expand...
100% crop is a pixel perfect image. That is, you've taken a section out of the image and that section is reproduced pixel for pixel. In the case above, you need to go through to flickr and look at the original.

I'd agree with User Name - slight back focusing...

BTW, the naming of your test set gives the lens identity away ;) :D
 
I can't see any problem at all to be honest Laurence. If they're less than 100% crops which haven't been sharpened (which is mandatory) they're more than acceptable and would sharpen nicely anyway.

Even if they're 100 crops I'd still consider them good for an image viewed at 100%
 
Laurence, the lens looks fine, stop worrying about it..............:thumbs:
 
More interestingly Laurence what's the book like - looks a good read?
 
arad85 said:
The pics aren't 100% in the thread. The coke can is 1099 x 1820 for instance. Looking at the 100% image, it does look slightly rear focused.

I'd suggest setting up a test as in: http://www.talkphotography.co.uk/forums/showpost.php?p=2961973&postcount=22

You'll soon see if it is front/rear focusing....
Click to expand...

cheers, do I eat the stuff before or after the test?
Actually what I did do was go here

http://regex.info/blog/photo-tech/focus-chart

I'll try the test tomorrow.
 
Last edited:
footman said:
cheers, do I eat the stuff before or after the test?
Actually what I did do was go here

http://regex.info/blog/photo-tech/focus-chart

I'll try the test tomorrow.
Click to expand...
The problem with that is that it's so "busy" you have to be really sure what you're focusing on.

As to eating the cereal first... just give your lens 3 weetabix and I'm sure it'll focus perfectly ;) :exit:
 
Well how did you take it RAW or JPG ? I looked at it and here is a copy with some sharpening
5105433350_50f3f6cb8b_z.jpg
yours>
5105151472_60aabd8b0f_z.jpg

Think you might need to look at full size to tell.
 
@ chaz photos, wow, thats a real diffrence, what sharpening did you do? how much? prog? etc just keen to learn as that really makes the can pop (see what I did) now:thumbs:
 
cherokee1111 said:
would you mind explaining why for us inexperienced users please? might help us identify similar issues with our own lenses.
Thanks
Click to expand...
Because the writing on the side of the can is more in focus than the writing on the front of the can. The great thing about having some detail at different distances from the camera is that it enables you to see where the focus point actually is.
 
PS. the book looks out of focus to me at 100%
 
cherokee1111 said:
@ chaz photos, wow, thats a real diffrence, what sharpening did you do? how much? prog? etc just keen to learn as that really makes the can pop (see what I did) now:thumbs:
Click to expand...

Just un-sharp mask at 267% 0.9 pixels Threshold 0 in CS4 on the biggest pic on Flickr
 
After searching on google, I found that it is indeed possible to polish a turd. This would be another great example!

If it is out of focus, you have to get it in focus somehow, not just sharpen 300% till it looks good at 6x4".

Chaz Photos said:
Well how did you take it RAW or JPG ? I looked at it and here is a copy with some sharpening
5105433350_50f3f6cb8b_z.jpg
yours>
5105151472_60aabd8b0f_z.jpg

Think you might need to look at full size to tell.
Click to expand...
 
The easiest way to check a suspect lens is to compare it against another lens of known good quality. Shoot the same subject, at the same time side by side, same framing, same processing same everything etc etc.

For focus testing, the cerial box works well. Just don't shoot at artificially close distance (like the chart you linked to - tests like that should really be banned ;) ). While the shallow DoF helps you see shifts more easily, very few lenses are perfect at all focal lengths and all focusing distances and at close range you are pushing right to the edge of the tolerance envelope, and possibly beyond it.

The danger is that you see a slight error, correct it with micro-adjust, which then puts the AF out at normal range. Say 25x focal length, but no closer than 2m as a rule of thumb.
 
I do not understand people who get so worked up about lenses. They imagine all sorts of problems. Soft focus,back focus, front focus and so it goes on. Buy a test chart this will convince you all is OK. If a lens resolves a resolution of 80 LPMM at a stated distance its a good lens. Remember though, this is only one part of a lens test. A staggered line of AA batteries will show up auto focus faults. A brick wall for checking barrel and pin cushion distortion.
 
daugirdas said:
After searching on google, I found that it is indeed possible to polish a turd. This would be another great example!

If it is out of focus, you have to get it in focus somehow, not just sharpen 300% till it looks good at 6x4".
Click to expand...

The OP said nothing about out of focus, sharpen will NOT fix out of focus.
soft is nothing to with this if the photo was taken in RAW as it was you have to sharpen it.
A photo can be in focus and some pay lots of money for a soft lens.
Too many get soft and oof mixed up
 
I'm quite sure I'm getting worked up about nothing, probably because it's an ebay purchase. It's in my nature dammit! I'll do a simple 3 batteries test and maybe a bookend test today weather permitting.
Anyway, I was playing about last night turd polishing and came up with these two:



No PP other than DPP conversion to JPEG and resizing.



PP in LR3. I know it could with more NR. This was from that group in the park that I was convinced were not sharp. Thanks for all the help people!
 
I promise, this is the end of it, no more shots. 3 batteries at a 45 degree angle to camera, focus on the DU of the centre battery, shot at f8. No PP except raw export from LR3.





to me it seems that it's front focussing slightly. If I'm guilty of stupid pixel peeping then tell me, this is not grounds for rejecting a lens, right?
 
Laurence, what you need to do is either upload the full sized file to flickR and post us a link, or chop out an 800 pixel (longest side) section from the full file. If we're looking at images which have been reduced in size at all then it's impossible to determine anythng from them as they lose resolution (sharpness) during the reduction and will need sharpening.
 
The other thing Laurence is that your 1DMK2 doesn't have user micro adjustment for individual lenses which Canon have added to several later models, so in a worst case scenario it's possible the lens may need calibrating to the body, but with a 1DMK2 that's a job for a Canon tech unfortunately, but I don't think it's grounds for rejecting the lens quite honestly, even IF calibration is needed.

Even with user calibration though, the test needs to be set up very carefully and Canon recommend at least 3 exposures for each test to establish an average result. you can rely on. I've buggered about with user calibration till I've had spots in front of my eyes - it's easy to convince yourself there's a problem where there is none. Suffice it to say that my user calibration is set at 0 on all my lenses with both bodies and I have no issues at all with any of them.
 
You must log in or register to reply here.
Back
Top