[jatrax]

Here are test shots of the DFA 24-70 @ 70mm and at various apertures from f/2.8 to f/22. I want to get opinions on the lens before I decide to keep it. This is a very expensive lens for me, and I usually only buy used but with the K-1 soon to arrive I felt I needed this to go with. Now I'm concerned I have a good one. Lens testing is maybe not my expertise so anyone who knows please take a look.

To me it looks a little soft at f/2.8 softer than I was expecting but maybe that is just me so opinions welcome. So what do you think? Sharpness? Decentering?

Shot on the k-3II so hopefully this will be good enough to see if I have a 'good' copy. We are not seeing the edges I suppose but for now this will have to do.

At f/2.8


At f/4


At f/5.6


At f/8


At f/11


At f/16


At f/22


---------- Post added 04-14-16 at 09:11 AM ----------

And here is one from the Pentax-F 50mm macro @ f/4 for comparison. I think this is the sharpest lens I own so good to compare.

[BigMackCam]

Hi John... Overall sharpness at f/2.8 looks good to me... about on a par with what I've read for the this lens wide open. The only thing that immediately concerned me was the slight softness and purple CA to the right hand edge of the blocks on the right side of the image... It's most noticeable at f/2.8, but I can still see it in your f/8 shot too, albeit just the tiniest bit. Slight de-centering, perhaps?

[nomadkng]

at f2.8 the right 1/3 looks a little soft. especially upper right quadrant.

I think your lighting is a little low but for a quick test though, it will work:

from what I can tell, you are getting 600ish out the 24-70 and maybe 650ish out of the macro example.

there's not a lot of difference between the 50 and 24-70 from what I can tell based on the comparison shot, but the aperture of the 50 shot is not listed. I think the only issue with your 24-70 is the right portion at f2.8 which seems to be gone by f5.6.

[quh86]

I am no expert in lens testing either, but two things struck me immediately:
first in order to evaluate a zoom it is helpful to also look at the wide end, not only the long end. second: your light source! the light is solely coming from the right, thus the lens seems a little softer on the right side with more pronounced CA. at first i thought your lens was decentered with a soft right edge, but a strong light from only one side could induce flare and CA, thus lowering resolution(?) and sharpness.

so i would suggest to place the light source on the left side and see if the lens exhibits the same, or you could alternatively just use a light that is pointed away from the scene you are shooting

[jatrax]

the aperture of the 50 shot is not listed

Aperture on the 50mm is f/4, about as good as it gets

your light source! the light is solely coming from the right, thus the lens seems a little softer on the right side with more pronounced CA.

Good point and possibly the cause of the issues noted above. This started out as a quick & dirty test, but I should have set up more professionally before I posted here. The room is well lit and I added a softbox with just the modeling light on the upper right. Probably not the best idea. I will re-shoot tonight with different lighting.

[FantasticMrFox]

First I looked at the centre of the f/2.8 image and thought "Soft?! That thing is sharp, sharp, sharp!" Then I had a look at the upper right corner and realised it's a bit softer than the other corners.

Looks like slight decentering to me.

[jatrax]

you are getting 600ish out the 24-70 and maybe 650ish out of the macro example.

Can you explain what this means? I have the test chart but honestly have no idea how to use it.............

[nomadkng]

Can you explain what this means? I have the test chart but honestly have no idea how to use it.............

the scale is line widths x100, which basically means look at the sets of parallel lines and find the set where you can no longer differentiate between lines. the number corresponding to that set is x100. obviously the higher the number, the greater the resolution. 600ish is pretty good. moire is actually kicking in at that range, but I think your lens might be resolving even higher than that. without major pixel peeping on my 4k monitor i couldn't give you a better number.

if you were absolutely retentive, you could order another copy and shoot them side by side and keep the best. but unless you will be shooting wide open a lot and shooting finely textured objects, I'd say that lens is probably very useable. yes there's a slight decentering on the right side, but it's not hideous.

[jatrax]

the scale is line widths x100, which basically means look at the sets of parallel lines and find the set where you can no longer differentiate between lines. the number corresponding to that set is x100. obviously the higher the number, the greater the resolution. 600ish is pretty good.

Thank You! I know that was a dumb question but I've got it now. I'm going to reshoot tonight with better, more even lighting and see if that makes any difference.

I'm heading to a wildflower site tomorrow at sunrise and would like to take this lens, but I won't if I'm considering sending it back. I've no problem testing in the studio but I figure once I've had it in the field it's mine.

[i5_david]

this looks like an absolutely brilliant lens. wow. so damn sharp open wide? a keeper and for me another piece of glass I need to save up to HAHA.
amazing!

[nomadkng]

Thank You! I know that was a dumb question but I've got it now. I'm going to reshoot tonight with better, more even lighting and see if that makes any difference.

I'm heading to a wildflower site tomorrow at sunrise and would like to take this lens, but I won't if I'm considering sending it back. I've no problem testing in the studio but I figure once I've had it in the field it's mine.

btw, was just reading about the Circle of Confusion and came across this:

For most people, the closest comfortable viewing distance, termed the near distance for distinct vision (Ray 2000, 52), is approximately 25 cm. At this distance, a person with good vision can usually distinguish an image resolution of 5 line pairs per millimeter (lp/mm), equivalent to a CoC of 0.2 mm in the final image....
A comfortable viewing distance is also one at which the angle of view is approximately 60° (Ray 2000, 52); at a distance of 25 cm, this corresponds to about 30 cm, approximately the diagonal of an 8″×10″ image....
if the larger final image will be viewed at the normal distance of 25 cm, a smaller original-image CoC will be needed to provide acceptable sharpness.

So basically what that means: the 5(00) line pairs in your test chart correspond to the avg human eyesight for an 8x10 image viewed at 25cm.

[kp0c]

First I looked at the centre of the f/2.8 image and thought "Soft?! That thing is sharp, sharp, sharp!" Then I had a look at the upper right corner and realised it's a bit softer than the other corners.

Looks like slight decentering to me.

Everyone assume the camera was perfectly perpendicular to the rest chart, which was probably not the case. In such case, that could account for the slightly softer right side. (Assuming focus was perfectly on the centre and dof extends a little further behind than in front of perfect focus, there is a possibility to get the left side in focus and not the right side, but yeah, probably unlikely.)

[jatrax]

Upper right has some light reflection on it that I did not notice when shooting. So that might be the issue. And though I eyeballed the line up, it might not have been 100% perpendicular. Will try this again tonight and fix those issues.

Thanks for all the comments and help! I'm impressed with the build of the lens, it just feels 'professional' but I want to make sure I have a good one.

[tibbitts]

Something you can do if your tripod and head allow is to flip the camera over, which will eliminate a lot of the alignment issues. Or, just do vertical in both orientations, which any ball head can do. It is hard to get the alignment and the lighting just right etc. so flipping the camera is generally informative. Plus if you only have one body, you never know if there's some odd sensor alignment issue.

Ultimately the lens manufacturers need to give us some guidelines here as to what to expect, or everybody is going to be sending back dozens of copies of every lens. By being silent about how they test and what their standards are, or what is actually possible in their production process, they just encourage us to make up our own rules and keep testing. The camera vs. lens manufactures are probably ok with that of course since we can burn through bodies doing nothing but resolution testing.

Speaking of which, you can't just test at close distances, you need more distant subjects as well. Well you can just test at short subject distances, but you may not always get the same results as at longer distances.

In another thread I mentioned that I asked Pentax (well, the customer service folks) what to do, and they said if you can see a difference from one edge corner to another, send it back or keep it and send it in for them to fix it. That's a little tougher than what I'd say, but it's as close to an official position as I've heard. Good luck on sending a lens back, of course, because they won't tell us what equipment they have at what facilities to actually fix problems like this, or what their standards are.

[Sandy Hancock]

In the f/2.8 image, the right edge is certainly a little soft as noted above (while the rest of the image is *impressively* sharp), but in the f/4 image the left edge is less sharp.

Are you sure your test chart was absolutely parallel to the sensor for all shots?