18-55mm woes (long)

Hi all,

Please bear with me while I describe a possilbly phantom problem with
my 18-55 mm verison II D40 "kit" lens. I've never been at all convinced
by its wide angle performance, so I've finally got around to doing some
primitive resolution tests. Lacking real test charts, I simply printed some
text at different font sizes and stuck the piece of paper on the wall...
Photos are off a tripod with 2-second shutter delay at f 8 or f 9.

With anything more than about 1.5 metres (4 feet) away, the lens at 18 mm is really soft both centre and edges, no wonder real pictures disappointed....

But... when the subject is closer than about 1 metre, or better at 50 cm (18 inches) then things sharpen dramatically (see the two exciting pictures at this link for a comparison)... The largest text is 12 point, then 11, 10, 9...

http://picasaweb.google.co.uk/blubbo99/LestTest

At 55mm, everything looks really sharp at both distances. This behaviour
is quite baffling, and totally at variance with, e.g. the results from

slrgear dot com, where the edges are soft at 18 mm, but the centre is very sharp, not significantly different from its 55-mm sharpness.

http://www.slrgear.com/reviews/showproduct.php/product/1033/cat/13

Is the lens faulty? Am I misinterpreting something?? If anybody could do
a similar test and report back, it would be enormously useful. My 55-200 mm

seems to work fine, so I presume(?) its not a front/back focusing issue, but please feel free to correct me on that.

Thanks,
Bill

I downloaded your large file and it appears to be sharp...except for the typed copy on white sheet. The words did not resolve. At f8 you are stopped down over 2 stops so you should be in the lens' sweet spot. Everything else along the plain of focus, appears to be sharp.

Why did you use ISO 400 instead of something slower? Still, this should not have made any measurable difference.

Thanks for the reply... But that's just the problem, why are
the words blurred when the image is evidently in foucs?
Maybe this is just the normal performance of the 18-55 mm at
18 mm, but if so, it's not great. I will do some more testing
at ISO 200 and report back...

Bill

Neocon said:

I downloaded your large file and it appears to be sharp...except
for the typed copy on white sheet. The words did not resolve. At
f8 you are stopped down over 2 stops so you should be in the lens'
sweet spot. Everything else along the plain of focus, appears to
be sharp.

Why did you use ISO 400 instead of something slower? Still, this
should not have made any measurable difference.

Click to expand...

Bill, at ISO 200 with camera on tripod you can shoot at various f-stops. But, I am certain your real world photography does not involve having a typed sheet as your main subject. Instead try tripoded shots of subjects you normally photograph. If you get sharp, contrasty prints...then don't worry about it.

I'll be blunt not to be rude but to save you worry

The test has no real validity

Try again but this time the framing should be the same in each shot. In other words tape 3 sheets together to form a vertical strip. Take both shots with the 3 pages just touching the edges of the frame top and bottom.

In you test as you go wide to 18mm you made the text 3 times smaller, no wonder you can't read it
--
http://www.photo.net/photos/JohnClinch
http://www.flickr.com/photos/john_clinch/

Lets face it, any modern DSLR will do the job.

I don't quite get what you're doing. So there are two cases:

Case 1) You did something reasonable. Then you may have a focus issue at the wide end. Could be, but not very likely due to large DOF at wide angle.

Case 2) You're mixing up sharpness with some other unknown property. At wide angles, less detail will be captured since the "individual piece of detail" becomes much smaller. You can use a horribly soft 400 mm lens and a exceedingly sharp wide angle lens, and you'll be able to read texts far away with the tele but not the wide angle...

I don't know what kind of disappointing shots you got at 18 mm. Maybe you should post one of your real world samples.

BG

....I don't take many photos of text, but these don't seem soft to me.





-this one not 18mm but is cropped and at 100%



--



Unlike a painter, a photographer starts with something finished and works backward - Anonymous
http://rlatt.smugmug.com

I'm not surprised by your findings. I've extensively tested two 18-55 lens myself. One was the version II and tho other one was the original. Both were new. There were very real and disturbing differences between the performances of both. The version II had better corner sharpness overall but less center sharpness. For some odd reason, at same exposure, white balance setting and aperture, the version II's color and contrast rendition was very different from the original. I could focus closer to a subject with the version II at various focal length's. In the end, I kept the original because I couldn't get used to the unnatural color cast of the version II.

I wasn't happy with either lens at the 18mm setting. The 55 mm setting is satisfactory. When you think that most users will probably use this lens at the 18-24mm setting, it hardly makes sense. Then again, the lens is really inexpensive. For not much more money, the 55-200 VR is so much better.

I'm saving for the 17-55 F2.8 because the 18-55 as it is now just won't cut it when a wide angle focal length is desired.
--

http://www.pbase.com/michelfleury

I am also not very happy with amount of details 18-55 can resolve. Here are couple of pictures with foliage and it is very fuzzy as you can see. The pictures are crappy (blown highlights and the 1st is probably not the easies for autofocus to deal with but at 18mm F11 it should not really matter). Anyway this rendering of foliage is consistent with my other pictures. Looks like the lens does not resolve details very well.

http://picasaweb.google.com/roytmana/Temp

to download full size photo click on the photo in the album and then on "download photo" link on left hand side

At the same time results of my Utah trip were very pleasing
http://picasaweb.google.com/roytmana/Utah2007Mix

but they were of large objects at huge distances where fine details are not that important

Alex

What you say runs contrary to what most users of this lens experience. It's quite good at 18 mm, most say its better than the 18-70 DX at this focal length. There are even some that bought this lens saying they'll use it as a 18/3.5 "prime". I think this lens is a bargain for the price - but I guess there is indeed some sample variation. I'm happy with the copies I've used though...

BG

Couldn't download the images in full size. However, none of them is 18/11. One is 18/6.3 and one is 32/11. Also, are you sure there is no motion blur? If there was any wind blowing at that time, the one with the foliage could easily suffer blurring from movement of the leaves at 1/60 sec.

BG

I found the lens to be reasonable when stopped down to at least f8 at 18mm otherwisw it's extremely blurry, full of CA and generally nasty looking most of the time. Wide open it's hopeless unless you crop down at least 500 pixels on each side. Now, whether this is sample variation I don't know but mine isn't overly impressive and I don't use it unless there is plenty of light available to stop it down.
--

Gary

Nikon D40
50-1.8
18-55
55-200VR

Can you post any examples?

I too was unable to download a full size image, 912x606 is what downloaded. I don't think the shot looks all that soft. The DR (Dynamic Range) is a little too much and both the shadows and the highlights run off the histogram. I brought the shadows up just a bit and compensated for the AA-filter by applying a little USM to the luminosity channel and the photo looks pretty sharp. Again this is working with a fairly small image. Basil might be right when he says there may be some motion blur present in the foliage; 1/60s would certainly be slow enough to allow for some motion to creep into a picture with even a small amount of breeze.





--
Brooks

“Dodging and burning are steps to take care of mistakes God made in establishing tonal relationships.” “Twelve significant photographs in any one year is a good crop.”
Ansel Adams

It should be noted that the optics did not change from the original lens and the MK II version. The MK II version may have a marginally tighter inner tube, but the main difference appears to be cosmetic.
http://www.photozone.de/8Reviews/lenses/nikkor_1855_3556_II/index.htm
--
Brooks

“Dodging and burning are steps to take care of mistakes God made in establishing tonal relationships.” “Twelve significant photographs in any one year is a good crop.”
Ansel Adams

Thanks for the many interesting replies. Neocon's warning about
me becoming a pixel peeper was very timely ;-)

The following, I hope, shows the very same lens that I was moaning about really doing rather well at 18 mm (processed from raw with lightroom including sharpening).

http://picasaweb.google.co.uk/blubbo99/Trees/photo#5082710848078569234

Of course, the trees are much further away than my silly test chart, so perhaps(?)
there is a focus issue, which I will look into further. However, my
primitive assessment of front/back focus (take pictures of a tape measure on the
floor from about a 45 degree angle) seem to suggest that there isn't a problem.

However, I will now dig myself back into a hole, by trying to calculate the

"absolute resolving power per pixel (arppp)" of the D40 sensor, which may be a limiting factor here.

The D40 sensor is (of course) 3008 x 2000 pixels. Pythagoras gives us
a diagonal pixel count of 3612 (of course the diagonal will cut through these

pixels at an angle, but it's close enough). The diagonal angle of view at 18 mm

for this lens is 76 degrees (from slrgear.com). Thus, by rudimentary trigonometry (has to be for me) we can work out the arppp at various distances. At 2 m (6 feet), the arppp is about 0.86 mm (about 3/100 of an inch). The 12-pont letters on my piece of paper have a diagonal pitch of about(!) 2-3 mm, so maybe 3-5 pixels worth. Viewing at high maginfication suggests this is not a million miles from the truth. So, as I think Basil said, I was expecting the impossible from the lens and it has very little to do with sharpness, per se.

For the trees, at about 20 metres (60 feet), the arppp is about 9 mm, or 1/3 of an inch, enough to resolve individual leaves, which it can do with ease.

By the way, the difference between the basic jpg out of the camera and the fine jpg from raw is huge, but that's another story...

Bill

Neocon said:

Can you post any examples?

Click to expand...

Sure, they are big though so rather than direct linking I put a space after the period, before the jpg. Copy the whole line and paste into your browser, then remove the space before the jpg.

20mm f3.8 1/100, ISO 200

http://img295.imageshack.us/img295/3273/dsc2155lf9 . jpg

18mm f10 1/100, ISO 200

http://img295.imageshack.us/img295/8595/dsc2249fc7 . jpg

Now, as you can see, little of the first pic is in any way useable for anything when shot wide open. The overall image quality is abysmal and IMO the lens is effectively an F5.6 Max at all focal lengths. At f8 it cleans up nicely and at f10 is pretty good. I've had much worse CR results than this but didn't keep them as they are essentially worthless.
--

Gary

Nikon D40
50-1.8
18-55
55-200VR

Gary, thanks for posting the two shots. Second shot is much better in terms of crispness (not including depth of field), control of CA and optical distortions. In terms of bang for your buck, this little kit lens is pretty good, if you can live with stopping down at least two stops.

I quickly stepped out and took few test shots from my porch

there are 9 full size pictures taken with D40.

18-55 II at 18mm and F8, F9, F11
Sigma 24 Ultrawide II manual focus F8, F9, F11
18-55 II at 24mm and F8, F9, F11

All pictures are RAW straight from the camera converted to JPEG with Capture NX with no PP saved with JPEG quality level 90

http://mysite.verizon.net/vze3zy9k/test/

Please check it out. 18-55 details seems to be not too bad but softish. I would greatly appreciate your assessment as to sharpness and amount of details

Thank you

Alex

Nice test.

The 18-55 holds up pretty well overall. Once stopped down it isn't a bad lens, a little soft and bland with more distortion and CA going on than I'd like but it's a cheap kit lens designed as a starter/beginner product and made to a price point to be economically viable at reasonable quality. My beef is that it's hopeless at low light where you need to open it up more. You have to choose between higher ISO noise or high distortion. I'll probably end up replacing the 18-55 with either primes or a better quality zoom as many of my shots seem to be in low light.

--

Gary

Nikon D40
50-1.8
18-55
55-200VR