24-105 : soft or not?

Hi guys

Been using/testing my 5DIII with 24-105 for a few days now. I realise stepping up from a crop takes a bit of adjusting, especially when it comes to DOF. But I have a feeling my 24-105 isn't always as sharp as I would expect it to be, especially when NOT zooming in completely. This one is an example. Center AF, focused on her eyes, single shot AF. This is what comes out (click for fullsize) :

IMAGE LINK: http://www.flickr.com …nsvondercrone/8​702365680/

IMAGE LINK: http://www.flickr.com …nsvondercrone/8​702391584/

It looks pretty good to me but if your concerned test it. An easy way is to manually focus using live view and zooming in 10x to get the focus perfect and take a picture, then do so with the AF and check for any differences. If it needs a little adjustment, that's what MFA is for.

wide open?

ed rader wrote in post #15890299
wide open?

Yes, most around f4 or f4.5 (used Tv mode to avoid subject movement, so set that to 1/400th)

Took some more in her bedroom. Sooc JPG, no adjustments. Fixed focus point, but manual (so not always center AF to avoid problems with recomposing). What do you think?
(click for fullsize)

IMAGE LINK: http://www.flickr.com …nsvondercrone/8​702423238/

IMAGE LINK: http://www.flickr.com …nsvondercrone/8​702423968/

IMAGE LINK: http://www.flickr.com …nsvondercrone/8​701302975/

IMAGE LINK: http://www.flickr.com …nsvondercrone/8​702425398/

The second photo is sharp on the eyelashes and at 105mm this close the DOF is paper thin...

You can sit there guessing the whole day.

There may be two issues for a soft photo.

a) The lens is soft throughout the distance range at any given (or all, which clearly is not the case here) focal lengths.

b) The lens is front/back focusing at specific (or many) focal length.

I'd suggest you get the target from this site and test for front/back focusing. Read the instructions carefully before you test, so you can get reasonable results/conclusions.

In the process, if you have a lens which is soft throughout (say at 25 mm), you will see it in the scale.

http://www.peleng8.com​/how-to-detect-back-focus.html

Here is the test for my nifty fifty as an example

IMAGE LINK: http://s37.photobucket​.com …/50mmf18_100cro​p.jpg.html

Fact that your second pic is much sharper than the first one means it's not body/lens fault.

First pic was shot @ ISO1600 vs 2nd pic @ ISO1250.

Besides you can afford to cut shutter speed by half at least, that will in turn cut your ISO by half.

Also Single shot AF isn't recommended for living subjects, use al servo instead.

EDIT - just saw your second lots of samples, if those aren't sharp, I honestly don't know what is...:p

that looks about right for wide open. a few things you've got to be careful of with this lens.

90mm to 105 can be softish especially wide open. f4 is soft (some will claim "tack" sharp). I would always stop down to at least f4.5 and preferably f5.6

also distortion at 24mm can be very noticeable, especially shooting buildings or anything with vertical lines.

if you are "testing" for sharpness I would suggest you do so outdoors preferably in good light.

Ok thanks for the info. What I found pretty strange is that it seems sharper (wide open) at tele than it is wide. Would expect the other way around. I know, it's not the ideal situation to test a lens but then again, most of the pictures I take are not in ideal situations either so I rather go for a 'out in the field' test instead of a scientific one.

I've also noticed quite some vignetting and (less so) distortion when importing the (raw) files into lightroom. Guess that's another thing that I'll have to get used to with a FF (after shooting over +15yrs with a crop). I know this is an issue far less likely to occur on crop (due to the 1.6x factor).

Thanks again for your input guys.

I find that mine doesn't really get super sharp until almost f8. Now, that is on a 1.6 crop stupidly small pixel sensor but the point is that it does need to be stopped down to get super sharp (as is the case with most lenses).

Your results actually look pretty good to me.

Looks crazy soft to me! I will gladly take it off your hands for a very discounted price!

Your lens is fine.

Here's the problem with your test:

1. Poor light
2. You're hand holding your camera, thus introducing your movements into the equation
3. You're shooting a moving subject (no matter how still you try to be, neither you nor your subject are static)

If you want to do it right, shoot a completely static subject in good light at a reasonably fast shutter speed with the camera on a tripod, mirror locked up, and a shutter release used to fire the camera. You can use a chart tacked to the wall, or use a front/back focus type setup (batteries lined up, the chart above, etc.). Shoot systematically at progressive apertures. But, before you do that, micro adjust the lens to the camera.

Also keep in mind when scrutinizing images for sharpness that there is only one plane of focus. Regardless of depth of field there will be only one plane of space that is perfectly in focus. The larger you view that image the more that will become apparent.

That first one the focus is just off. But others do have a very clear point of focus in them.

it looks ok. I think your biggest issue with the first shot is that you've got pretty thin DOF, and at 105, it's the lenses weak spot, and on top of that, you're shooting ISO 1600.

if you want razor sharp images, use more lighting and/or stop down a bit

Both have a narrow DOF. While it's nice to have the background OOF, for shots like this I would be at at least f5.6 on FF. This will also sharpen up the lens a lot. The lens is pretty darn sharp after f5.6 IN THE CENTER, but even at this aperture sharpness falls off quite rapidly.

However, I think the shots look sharp enough to be 'good' assuming you are happy with the DOF. The wide-angle ones you could probably get away with a tad extra PP selective sharpening in the face if you wanted.

Then if I want more 'pop' I selectively sharpen in PP (face, eyes) and perhaps desaturate the background slightly.

I like the first wide-angle bedroom shot the best, nice pose and nice complementary colors.

I don't see any problem at all. I don't know how much sharper you want it. If you printed a normal sized print you wouldn't be able to see any softness.