18-70 vs the 20-4 and the 50-1.8

I just completed a homestyle resolution test of the 18-70 vs my Nikon 20mm 4

and my AFD 50 1.8. All shots at F8 on my tripod mounted D50 and a solid mounted

newspaper as a target. Self timer used on all shots. At 20mm the 20 4 maybe,

just maybe had a slight resolution edge. If so it was very faint and almost not

worth mentioning. At 50 I couldn't tell the difference between the 50 1.8 and

the 18-70. This really surprised me. Multiple shots were taken to provide the

sharpest samples possible. BTW I was not cheering for any lens in particular.

After all I own them. Now I'm not saying the 18-70 is an ideal lens or ignoring

the fact that primes are better performers open or maybe exibit less

distortion, or in this case are smalller and lighter. Nor do I want anyone to

forget how handy F1.8 can be in available light. I just want to point out that

the 18-70 is a hell of a lens for the price. At least if you get a good sample

it is.

A lot of poor lenses can perform well at f8 or f11. What separates a great lens and a mediocre one is at wide open and at the extreme ends of the zoom range in the case of zooms.

 

I happen to like the 18-70 a lot becasue it is a very good deal for a $300 or so lens. However, I would compare it at f4 against the 50mm/f1.8 and 20mm/f4(??) also at f4. f8 is masking off some problems.

I was already suspecting among 'kit' lenses of various brands, this is the best. You confirm this.

 

It also has a perfect reasonnable and usefull range. On that aspect (low cost kit lens), Nikon as the edge for the budget minded buyer.

Agree. And I thought I made that point clear in my post.

Michael,<br>

<br>

Please forgive me but the test target and the likely test

distance are not suitable for general photography. What you need

are small individual targets that you place on a flat wall. For

general photography your test distance should be roughly 5m or 15

ft. A good brick wall isnt bad and though jokes are made

about those who photograph them very useful information is gained

from testing. The false assumption is that those who photograph

brick walls dont do anything else.<br>

<br>

Since you camera is an auto focus camera, focus error could be the

reason you dont see a difference between the 50/1.8 and 18~70/3.5~4.5.

Another problem is by f/8.0 diffraction is starting to take the

edge off the best lenses. Many think diffraction starts at f/16

or f/22 but its there all along. <br>

<br>

The sweet spot of a lens is where decreasing residual aberrations

meets increasing diffraction. Its typical to graph this and

a the chart shows a X as X marks the spot. The 50/1.8

will hit its sweet spot proper somewhere between f/4.0 and f/5.6.

Also the sweet spot is normally observed as a large central area

of the image. If youve seen the 12mm circle in cameras like

the Nikon F3 or F100 thats about the area that concerns me.

In careful testing Ive seen the central image area fall off

in resolution as the edge increase. The answer for this is a bit

more stopping down can be required to control residual

aberrations at the edge of the format as compared to the central

area. A lens that produces its best image quality at f/4.5 in the

broad central area might require ff/8.0 at the extreme corner. If

a lens for 35mm or DX requires f/8.0 to hit it's "sweet spot" I

do not consider that it has one. To put it another way the lens

really isn't too sweet. If you don't know a lens and want the

sharpest image shoot prime lenses at f/5.6 and zoom lenses at f/8.0. Some zooms do attain a sweet spot by f/5.6.<br>

<br>

Im not dumping on AF cameras here as I own three of these.

My preferred focus method is probably manual but sometimes, PJ

style flash photography for example, I rely exclusively on auto

focus. I also rely on an aperture of about f/5.6 to cover for AF

focus errors.<br>

<br>

There are so many factors to control in lens testing. I use a

camera with a 6x, high magnification waist level viewfinder and

focus by the light of a 1,000 watt quarts halogen flood light to

nail the focus. If the camera itself is out of specs this wont

give the desired results. I make the exposure by electronic flash

in total darkness using electronic flash. Camera alignment must

be meticulous or its all for nothing, etc.<br>

<br>

With modern AF cameras focusing a lens that is faster than f/2.0

requires a lot of skill or worse is just a crap shot. A lot

depends on your camera. Even a Nikon D2X or D2H is less than

ideal for fast lenses. I think with less than a 6x finder

critical focus with a lens faster than f/2.0 is not assured. At

any rate a Nikon F2 or F3 with the old coarse focus screens are

preferred for truly fast lenses and critical focus.<br>

<br>

To sum this up I recommend that you not place too much importance

on this test. Proper lens testing is very difficult.<br>

<br>

Best,<br>

<br>

Dave Hartman.

Thanks for doing the test. You've proved the old adage "F/8 and be there"! I tell folks who email me asking if the kit lens is "okay" that "sure, as long as you are going to take photos outdoors on a sunny day". I think your test lends some credence to this advice as well.

Michael, did you do any comparison tests at f16 and f22? I am particulary interested in these two f stops.

 

My 18-70 DX sample performs poorly at 70mm when compared to the 75-150 Series E. (at f8 and just about everywhere else)

 

I don't know if there is a rule of thumb that says that a lens is more likely to perform better wide than long, and perhaps some one with more experience could please help me out.

Micheal,

 

Can you post some comparative results? Here is my page of prime verses 18-70 and 18-35 lenses at equivalent focal lengths:

 

http://aaronlinsdau.com/gear/articles/lens_comparison.html

Exactly Phillip. That was the point I was trying to make. This kit lens makes a fine, reasonably lightweight travel lens when used at mid apertures. I would though like to adjust my findings on the 50mm comparison. More eyeball study is showing that the 50 1.8 is pulling down a fraction more detail. You really have to look hard to see it though.

 

David point well taken and respected. Guess I'm trying to say this lens isn't half bad.

 

Andy I happen to own a 75-150 which is an excellent little zoom. Sharp as a tack on my F100 but I don't use it on my digitals. Maybe I'll test it out a bit later. I only did the test at F8. Sorry.

 

Arron I try to post the results. I need to put the images together.

The 18-70 has really impressed me, but Dave's points can't be ignored. IMHO, the autofocus on the D200 isn't good enough for lens testing, though it's usually fine for actual photography. Nor is the focusing accuracy through the finder good enough for lens testing, at least not with my eyes and no magnifier. What you end up testing is the system, with yourself as part of that system. The sweet spot of most normal primes is about f/5.6, and as Dave points out, by f/8 you're losing resolution to diffraction. No idea about zooms, but if the lens design is really good, it will be an even larger aperture. Note that very small sensor P&S cameras are limited to f/8, because diffraction is so severe beyond that. Anyway, I agree that it's a great lens for the money. All I want beyond the 18-70 is a Sigma 10-20 (or similar), and 99% of what I shoot would be covered.

Lenses vary from sample to sample of course. My 18-70 at 50mm performs nearly identically to my 50mm 1.8 when both are at f4.5 (compared at normal subject distance), which is wide open for the 18-70. However, the 50mm performs well at 1.8, which is why I have it! The 18-70 performs very closely to my 24mm 2.8 and 28mm 2.8, but they both are pretty sharp at 2.8, so they are slightly more usable in darker rooms. The main advantage of the 18-70 is the zoom for all around shots in adequate light. I prefer a faster f-stop for portraits, however.

My 18-70 is in the shop because it started to simply come apart near one of the rings.

However, I was quite pleased with it for general use and it is indeed pretty darn sharp. I

think I had a pretty good sample, and I hope they don't wreck it at the shop, because it did

work very well throughout the range.

 

I don't agree w/the comment about the auto focus of the D200. It has been very accurate

for me. I do have to agree thought that some of what makes a lens good is its wide open

performance. Although the 18-70 gives a good center sharp image wide open through all

its zoom range, and indoors with a flash, (or outdoors for that matter), its out of focus

areas can be pretty darn ugly. Where the 50 1.4 I have wide or near wide open gets some

reasonably nice and smooth backgrounds.

I do not believe that very old E series (amateurish)zoom performs better than 18-70 ED lens. Not true. I saw how well Canon 50/1.4 performs with EOS 30D at 2.8 - very , very soft. Although 50/1.4 is a pro lens and much better then your E-series junk and... at 2.8 with film my Nikkor 50/1.4 is razor - sharp. The Nikkor 18-70 is designed for DX size. The 50 mm is designed for FF. At FF sensors 50 mm lenses are breathtaking (see EOS 5D).