NikonGear'23
Gear Talk => Lens Talk => Topic started by: Moritat on November 19, 2016, 16:47:19
-
I'm looking for opinions on the 18mm 3.5 AI-S. I understand it has CRC, but I would be using it more for distance shots (architecture, etc). I'm interested in it's performance at f8 or f11. Does it have good edge to edge sharpness at these apertures? Thanks in advance for your responses.
-
I'd say "yes" to your question. Sharpness across the frame is in fact very good from f/5.6 onwards.
The contrast of the f/3.5 is quite higher than of its predecessor, the 18 mm f/4 Nikkor, and colour rendition is excellent. However, unlike the older lens, there is some barrel distortion for intermediate to distant focus.
-
Don't be mislead by the term CRC (close range correction), it means the lens should perform well at infinity and at close range.
How does the AFS 18-35 compare? Moritat may prefer the manual lens, but the zoom is only slightly heavier, a bit more bulky, just as wide and fast, plus it zooms and is more compatible with modern cameras...
-
Bjorn... Thanks for the feedback on the sharpness and the examples. There doesn't seem to be a lot of info out there on this lens. I think most photographers purchased one of the 20mm lenses instead. Roland... I appreciate the info on the definition of CRC. And yes, I would prefer to stay with an older Nikon prime rather than a zoom. I'm quite fond of AI & AIS glass. My other options would be one of the 20mm lenses (2.8 3.5 4.0), but I've read that they aren't the greatest at distance/infinity. I've also seen some great defished shots from the 16mm 3.5 AI, but I think some edge sharpness would be lost in the process. I think the 18mm 3.5 might be worth a try.
-
I think the 18/3.5 is underrated mainly because it is 18 mm and thus falls between the ultra-wide 14/15 mm class and the very-wide 20 mm lenses. For its time, the f/3.5 was an excellent design capable of quality results. On a digital camera, there tends to be some lateral CA, but nothing that cannot easily be removed straight off the RAW converter. If you prefer JPG in-camera instead, then the behind-the-curtain processing of the jpgs usually reduces CA to insignificant levels any way.
The 18/3.5 has a near limit of 25 cm, meaning you will be able to poke the lens almost into the subject. Don't think it being a short focal length will supply you with infinite depth of field up close, though (see attaced picture below). Although the CRC is very effective at close range, it cannot prevent some mild corner smearing by coma in the out-of-focus background when you do close-ups.
The examples below are both film (unfortunately, Kodachrome; my apologies but we didn't know better at that time believing in Kodak's hype. About one year later I had dropped Kodak and went for Fuji films instead)
-
Here is an example of the 18 mm f/3.5 on a digital camera (Nikon Df). This image was at f/22 thus the lens tolerates reasonably well being stopped down all the way. One will observe that the background is not entirely sharp, despite the small aperture setting. This is to be expected as depth of field gets severely restricted once you move into the close range, virtually no matter what lens is deployed for the capture.
On a digital camera it is advantageous to CPU-modify the 18 mm so all EXIF data is automatically recorded (incidentally that is why I knew this image was done at f/22 ...). The modification of the 18 mm Nikkor is documented here http://nikongear.net/revival/index.php/topic,32.0.html.
-
One additional bonus, if one is thus inclined, is the 18 mm f/3.5 does excellent service for IR photography. Again, keeping image quality even when the lens is well stopped down prevails.
(Nikon D5300 IR, 18 mm f/3.5 Nikkor @f/16)
-
I do not own this lens. A competent French tester, J.-M. Sépulchre (who publishes in "Le Monde de la Photo") rates this lens higher than the 18/2.8 AF-D. Measurements attached. Used for landscape around f/8, it should be fully satisfactory - note however that the sharpness scores relate to a D700 (12 MPix).
His main rant is against the "moustache" distortion. Make sure you got a correction profile from somewhere.
-
The close-up picture of the daffodils, posted here, show clearly the geometric distortion on the background building. I didn't try to eliminate it, though.
Some subjects might require a different approach and either a lens profile, or a capable RAW converter such as Photo Ninja with its versatile tools for geometric correction, will be helpful to straighten things out.
Thanks for the test reference, Airy. My French has become too rusty, unfortunately, but the essence of the report comes over. My subjective impressions obviously agree with these findings.
-
Yes, definitely. Mr. Sépulchre's tests do not cover all relevant domains (e.g. bokeh, flare, LoCA...) but they are made seriously and consistently, so they help comparing rather than "judging".
-
Bjorn, thanks for all this information. I'm really impressed with the detail at apertures of f16 and f22. With most lenses I'm hesitant to stop down past f11. It's also good to know that the 18mm is a good candidate for IR as I have a converted IR camera. And Airy, thanks for the chart and additional information. I like what I'm seeing here.
-
I think the 18/3.5 is underrated mainly because it is 18 mm and thus falls between the ultra-wide 14/15 mm class and the very-wide 20 mm lenses.
Is the 18/3.5 underrated or just unknown? This focal length was never a big seller, probably due to the high price. Most photographers opted for the cheaper, and nearly-as-wide 20m focal length instead. It did sell in larger numbers than the 15mm model from the same period, which was even more expensive and much more bulky. The 15mm is probably more widely known since it was one of the "flagship" lenses in the Nikon lineup (if we ignore the 13mm...)
Reference: http://photosynthesis.co.nz/nikon/serialno.html#13_18
-
KEH lists one for $415. But I saw it uses 72mm filters :(. Bummer. I've got so many different filters now I refuse to buy a different size than what I have.
But Bjorn's examples are terrific. I think my gear horns just got a little too excited!
-
I would describe the 18/3.5 as both underrated and unknown, for the reasons that Roland has outlined.
It is a smaller lens than the Nikon units that preceded and succeeded it.
Nikon deprived it of a fully satisfactory lens hood solution: only a shallow slip-on was offered.
I can echo Bjørn's description of the color and contrast 'improvement' over the earlier 18/4.
-
I notice most reviewers use a small chart for checking distortion. Isn't this where distortion is at it's worst for most lenses? Why not the old brick wall? Maybe the small chart also. If I'm going to use a lens for photographing architecture do I need to see the distortion it gives when focused to 1.0 meter and less?
Dave Hartman
-
Same applies to sharpness or CA charts... on the other hand, I could see with the 20/3.5 UD for instance that there is definitely a correlation between short-and long-range distortion, and I am happy that LR provides a correction profile. It does not take the distance into account, as no such information is conveyed, but it provides a decent correction (calibrated on infinite ?) 100% precision is anyway neither realistic not needed.
I consider all chart info as relative (i.e. easing comparisons), but not telling much in the absolute. In absolute terms, I'd trust some reviewers and not others, after having made my opinion (correlation between their rating and my observations or satisfaction).
-
I notice most reviewers use a small chart for checking distortion. Isn't this where distortion is at it's worst for most lenses? Why not the old brick wall? Maybe the small chart also. If I'm going to use a lens for photographing architecture do I need to see the distortion it gives when focused to 1.0 meter and less?
Dave Hartman
I have several times wished to do a brick wall distortion demo with a very wide lens at a large distance, like 30 or 40 feet(10-12meters) . However, it has been impossible for me to find a brick wall that can be photographed with the camera dead square to the wall and at large distance and cover the camera frame with the bricks. Maybe one of those old Gitzo ladder tripods that stands 12 feet tall would be needed.
-
Keith, you could shoot at a normal angle of incidence and capture half the frame. Then flip the camera upside down and get the other half :D
I agree that testing distortion on close-up subjects, like most reviewers do, is a poor substitute for a properly conducted test of geometric distortion. For a lens with CRC, in particular a wide-angle lens, the likelihood is large that you will document barrel distortion that simply does not exist at intermediate to distant motifs.
-
Bjørn,
Thank you for confirming my suspicion.
Dave
-
Keith, you could shoot at a normal angle of incidence and capture half the frame. Then flip the camera upside down and get the other half :D
Excellent idea. For distortion, I'll probably leave out the bottom half since distortion is presumed to be symmetrical. Sharpness would be different, of course.