Round Three ~ Nikon Nikkor 85mm f/2 Ai and Friends

[Please note: I failed to focus the 135mm Nikkor f/2.8 AiS carefully enough in this comparison. While what I write is true, it relates _only_ to this comparison.  I've gone back and re-checked the resolution wide open.  I've confirmed my earlier observations that my copy of the 135mm f/2.8 Nikkor is indeed just a touch sharper than my copy of the 105mm Nikkor-P. Which is to say, the performance of these two Nikkors are incredibly close to my copy of the Sony FE 55mm f/1.8 ZA.~ 27 December 2024]

I have bought and sold three of these over the years and I really need to stop the constant swapping and looking for "better" or "majick."  Really, I do need to stop.  So this is it.  Hopefully this will be the last Nikon Nikkor 85mm f/2 Ai I dance with.

Across the "interwebs" it's easy to read various people's comments and thoughts on nearly everything.  The word on the street is that the Nikon Nikkor-H/HC/K 85mm f/1.8 has more "character" than the f/2.  And the word on the street is that the Nikon Nikkor-P 105mm f/2.5 is sharper than the f/2 85mm.  The Nikon Nikkor 135mm f/2.8 AiS seems to be unloved.  There is no word on the street for that one that I can find.

I've at various times bought and later sold three Nikon Nikkor-H/HC/K 85mm f/1.8 lenses, so my closet is empty of those just now.  Some say it's sharper than it's younger f/2 sister.  I've not found that to be the case.  At the same time the f/1.8 "does something swirly" to the out of focus background.  If you like "swirly", that is.  Frankly, I'm a little ambivalent about "swirly" since the highlights seem to always have hard edge "cat eyes" which to me feels strange.

Using the famous Sony FE 55mm f/1.8 Zeiss as a reference (everyone who's anybody says it's fabulous ~ have I been "influenced"?) I thought I'd have another look at comparing four lenses.  Here's what I see.

Comments -

Wide open performance from best to worst:

• Sony FE 55mm f/1.8 Zeiss
• Nikon Nikkor-P (Xenotar) 105mm f/2.5 pre-Ai
• Nikon Nikkor 135mm f/2.8 AiS
• Nikon Nikkor 85mm f/2 Ai

The Sony FE 55mm f/1.8 Zeiss is a touch soft wide open.  Stopped down to f/2.8 and f/4 it is nice and sharp.  Here's the thing, I seldom shoot the Sony FE 55mm f/1.8 Zeiss less than wide open (see here for one example) and have been completely satisfied with the results.  Hold this thought.

The 105mm Nikkor-P is nearly as sharp as the Sony 55mm, with the 135mm f/2.8 AiS another eyelash behind these two.  Something I find interesting about this is that I'm looking at four different lens designs from three, maybe four different eras of commercial optical development.  Hold this thought, too.

The Nikon Nikkor 85mm f/2 Ai is the softest of the bunch wide open.  Stopped down one or two clicks cleans things up nicely.  In fact, all four lenses stopped down a click or two look remarkably similar in terms of "sharpness."  The 85mm f/2 stands out in this group of optics for it's lack of "sharpness" wide open.  It's clearly behind the other lenses seen here.

Which leads me back to held thoughts.

The 105mm Nikkor-P is the oldest design in this group and yet it is nice and sharp from wide open.  Conventional "interwebs" wisdom would have us believe it should be the softest of the bunch.  But that's not how optical design and manufacturing historically work.  How to make a sharp lens has been known and the manufacturing process well understood for over a hundred years.  To underscore this point, the Nikon Nikkor-H 50mm f/2 Gauss design is bitingly sharp from wide open.  It's not for the lack of design and manufacturing knowledge that lenses are "unsharp" wide open.

OK, then, what's going on with the pretty little Nikon Nikkor 85mm f/2 Ai? Why is it visibly softer wide open than the other lenses seen here?  After-all, it was designed sometime in the late 1970's perhaps just a couple months ahead of the very sharp from wide open 135mm f/2.8.  

Given that Nikon knows how to design lenses I have to believe that the 85mm's spherical aberrations at f/2 are intentional. The veiling spherical aberration softness at f/2 is nearly gone by f/2.8 and is completely gone from what I can tell by f/4.  If they wanted to, Nikon could have designed a bitingly sharp 85mm at f/2.

Bottom-line:   I'll bet dollars to doughnuts Nikkor 85mm f/2 was intentionally designed to be soft/sharp at f/2.  And I'll bet euros to beer that this lens was very specifically designed as a _traditional_ portrait optic.  Which is why I'm betting that at f/2 the Nikon Nikkor 85mm Ai is _meant_ to soften skin tones.  Now _here_ is a lens with "character."  The effect is _designed_ into the lens.  It's a portrait lens when shot wide open.  It's a general purpose lens when stopped down.

Have a look at the following and tell me I'm wrong.

Nikon Nikkor 85mm f/2 Ai at f/2
Sony A7RII
Electronic first curtain OFF
IBIS set to 85mm

Command Line Interface ~ Linux

Notes to self:  A few useful image processing commands for running in Linux.  All these are much faster to run from the CLI than using an app that's having to manage graphics at the same time. -

convert *.jpg -average <averaged-filename>.jpg – averaging command

convert *.jpg -evaluate-sequence median <output file-name>.jpg  - a different averaging command

mogrify -resize 1920 *.jpg – resizing command

mogrify -bordercolor black -border 10x10 *.jpg – adding a thin black edge to images

mogrify -bordercolor white -border 400x400 *.jpg – adding a white border to images

convert <filename>.<file-extension> -colorspace gray <output filename>.<file-extension> – command to convert a single image to black and white

for i in *.jpg; do convert "$i" -colorspace Gray  "BW_$i"; done – Bash script to convert a bunch of files into black and white

exiftool -a -u -s -G1 <file_name> - to read EXIF image file data

gmic -input <filename.file-extension> scale_dcci2x , cut 0,255 round output <theOutputFileName>.tif - command to perform a DCCI2x upsize

 

Digital Zone System ~ Validation Testing

I wanted to prove to myself something about a Digital Zone System that I've been working on.  

What I wanted to prove was that once I'd set up an input correction curve for each camera I own that the result would match the output of my other cameras, each with their own unique input correction curve.

The cameras I own span recent digital sensor development and manufacturing.  For this validation I chose two different formats, two different megapixel counts, and three different years sensor implementations.

In each case I spot metered the light area to Zone 7 and let the shadows fall where they will.  I did this because in digital I meter for the highlights and process for the shadows.  Keeping in mind, of course, this is the exact opposite of what we do with film.

Here's what I see ->

 

 

To my eyes this is an excellent match. Each input correction curve accurately matches the 1EV step as described by the original Zone System and adapted here to digital.

So here's a little game.  Want to win a free beer?  Tell me which camera made which image.

For me that's an easy bet for the house.  I seriously doubt anyone will be able to pass this test.  In fact, it's rather irrelevant.  Though it does make the point that if we understand our tools and set up our processing environments correctly, manufacturer to manufacturer variations, sensor to sensor variations, and generation to generation developments can all be "leveled" to the point good image making doesn't depend on tools as much as some would have us believe.