Nikkor 28mm F3.5 for LTM

[David Murphy]

Just tested my W-Nikkor-C 28mm F3.5 for Leica Thread Mount and I thought the results might be of interest. This lens is fairly common in Nikon rangefinder mount, but a little less so in LTM. The classic camera literature says 10,000 were made in all mounts.

I know a little about my particular example since I talked to the second owner (and from whom I bought it) who had owned it since 1953. He bought it from his friend in 1953, who had purchased it in Japan in 1952. His Japanese travelling friend also had a tour of the Nikon factory. Apparently Nikon was just all the rage among the Asia travelling set in 1952 :-)

The the first photo shows the hardware used, my trusty Bessa-R with a Canon 28mm finder. The film was recently expired Kodacolor Gold 200. The setting is the CalTech Campus in Pasadena, famed for its beautiful gardens and interesting architecture. Most of the photos are between F11 and F22 -- no atempt to probe the entire range of the lens, just a basic sort of not-to-scientific field test. I put fresh batteries in the Bessa and obeyed the TTL meter (it's a good one!).

The Kodacolor was processed by the local pharmacy and the negatives scanned on my Minolta Dimage IV film scanner. The only edits to the photos were mostly minor cropping, dust/scratch removal, and very small tweaks of color balance and exposure -- no sharpening!

Comments on the lens performance are welcome. I don't claim to be a serious photographer, so please don't bash the composition (etc.) -- this is basically a lens test.

[payasam]

I don't see how anyone can comment on such images on a computer monitor screen. I'd say the lens is not sharp -- but I know it to be sharp.

[David Murphy]

Thanks for the comment. One is forced by site policies to degrade the images somewhat to post them. However the much higher resoluiion scanned raw images (like about 2000 X 1500) do not appear tangibly different. Of course screen resolution has a role - my monitors are typically 1600x1024 -- very decent.

I can prove this by running one of the images through a sharpening filter and posting it at the same size and compression as these. I will do so later this evening and the difference is clear, incidating the computer digitizing/reproduction process is not an obvious limiting factor in the presentation -- the limiting factor appears to be intrinsic ray errors of the (vintage) lens solution. I may also put in a link to a hi-res scanned image which I think will also demonstrate this.

Here's my spin: I think this lens has reasonably good contrast, but overall it is not as sharp as modern 28's -- this is no huge surprise. I have owned four vintage 28's. I'd rank this one just behind the 28/3.5 Canon (which came 4 years later), but better than both the 28/2.8 Canon and significantly better than the 28/8 Zeiss Tessar (a novelty lens really).

FYi for those not having ever handled a 50's vintage RF Nikkor, the construction quality is extremely impressive.

[ferider]

It should be sharper than this, David. Maybe it's just me, but
when I look at the picture with the roses it seems that the background
is sharper than the foreground. Feels almost like a collimation issue.

I would do some test photos on a controlled object close up and see ...

Sorry,

Roland.

[NIKON KIU]

Nice.
Joe had a little go-around with one of these,I wonder what his thoughts are.

Kiu

[David Murphy]

I guess I can't really agree. I think this is what the lens does. I have not seen a wide angle rangefinder lens made before about 1965 (35mm or 28mm) that was as good as say a modern K mount Vivitar 28mm SLR lens. I challenge anyone here to disprove me! (in the friendly RFF spirit that is)

These vintage lenses are cute, compact, well-made, and collectible, but they don't perform like the modern glass because their optical prescriptions are simply not fully optimized. They are not fully optimized because the computer software and computers to do it were not readily available in the 1950's -- hand performed calculations are not sufficient. The laws of physics will be obeyed regardless of whether the names "Leica" or "Nikon" are stamped on the lens barrels.

Without meaning to put words in his mouth, the site owner here will usually claim that any CV lens outpeforms vintage counterparts (to the extent a comparison can be made). I agree with him. The counterpart here would be the CV 28/3.5. I have one in Contax RF mount -- just waiting to get my IIa out of the shop to try it. I'll post the results here for fun using the same conditions and subjects.

[ferider]

There are many outstanding wide angles from the late 50s and early 60s,
David.

Check with Vince, he uses the 28/3.5 Nikkor extensively.

Roland.

PS: to clarify: this lens, this Nikkor 28/3.5 of yours should be sharper than your pictures show...

[David Murphy]

I just went to the previous thread on the Nikon version of this lens with various examples and links. Here is the url:

http://www.rangefinderforum.com/foru...cm+f3.5+nikkor

I am still not convinced. The images posted there have a similar softness -- especially the colors ones. Of course these are rather uncontrolled tests, but I do not see a tangible difference. It looks like the shots of the children there use fill flash giving a certain (small) hardness and sharp contrast to the subject illumination adding to the "impression" sharpness.

[David Murphy]

I should add here that I have something of a background in computer aided optical design -- I'm not just carping when I talk about this. The optical prescriptions of complex lenses of yore where pretty good and impressive to study, but everyone of them can be tightened up with computer optimzation. Lens coatings are also vastly better now as well. All of this adds up.

[VinceC]

>>It looks like the shots of the children there use fill flash giving a certain (small) hardness and sharp contrast to the subject illumination adding to the "impression" sharpness.<<

Your scans looked a bit soft to me as well. Were you using any unsharp mask? I agree you can't really tell with a monitor and online resolutions.

My scans are straight WalMart scans out of a Fuji Frontier processor at 2 megapixels, but the sharpness of the 28/3.5 greatly exceeds that resolution. I did a lot of darkroom work with this lens 10 and 15 years ago and felt it was as sharp as SLR Nikkor wide angles. It's vintage characteristics are based on light falloff on the edges and lower contrast.

No fill flash with most of the posted photos (perhaps the first one below, but not the others). Usually I'm shooting portraits when the light is just right and I can convince my kids to stop rolling their eyes. I do burn and dodge. As an old wet printer, I can't help myself.

[payasam]

Grass, leaves, petals.... I think Roland sees what I see.

[MikeL]

Originally Posted by David Murphy

I guess I can't really agree. I think this is what the lens does. I have not seen a wide angle rangefinder lens made before about 1965 (35mm or 28mm) that was as good as say a modern K mount Vivitar 28mm SLR lens. I challenge anyone here to disprove me! (in the friendly RFF spirit that is)

I don't know how to disprove you, but I think the Nikkor 35mm f1.8 is very sharp. I couldn't tell a difference in an 8x10 between it and a 35mm asph summicron, except for when wide open. Color and contrast were different, but resolution seemed similar.

[VinceC]

I'm at work right now, so the best I can do is pull some detail out of a 2 megapixel scan. At home I have some high resolution flatbed scans of wet darkroom prints.

As shown below, detail well exceeds 2MP scan.

[David Murphy]

OK, Here are the links to two shots of the roses. These were resized and compressed from the original 3MB TIFF scans just enough to meet the file size restrictions of Photobucket. They are just a little under 1Mb in file size, whereas the inline photos above are under 75Kb.

Raw Roses (no scratch removal or cropping):

http://i160.photobucket.com/albums/t...es1024jpeg.jpg

Raw Roses (no scratch remove or cropping and sharpened to the point the smooth background just starts to show some "grain" - i.e. noise):

http://i160.photobucket.com/albums/t...rp1024jpeg.jpg

Note that if you are on a dialup line these downloads maybe painfully slow.

I can say I'm not totally enchanted with the Dimage film scanner -- I worry about negative flatness. I will have a flat bed scanner up and running soon and I'll do a comparison -- getting the negative not to curl or warp is the key. Mounted slides are much better in that regard, but for lens tests I can get C41 developed in an hour at the corner pharmacy - complete with scratches :-)

Lest I offend any Nikon lovers, let me tell you I still adore this lens -- just trying to decide what's best in my lens kit -- I'm always torn between new CV and these vintage beauties -- I don't have deep enough pockets to buy the new Zeiss M or Leica models

Sorry for the "stream of conciousness", but in studying these higher resolution pics on a really good LCD monitor here at work and comparing side-by-side to the lower quality images above, I'm beginning to agree with payasam that the computer process is degrading the presentation to some degree. The Dimage scanner may be another contribting factor. I really need to study what an 8X10 true enlargement of the negative at a good lab can do -- maybe that'll be my next move.

[VinceC]

This is a sharp lens for its era and is in my category of "plenty sharp enough for everyday work." I'm sure newer lenses are sharper, but you'll see by the files attached below that the sharpness of this lens can exceed the grain of TMax 400.

Depending on how you rescale the photos can also have an effect. Some jpeg resizings seem muddier than others on computer screens. There are a lot of other variables: Film type, scanner type and resolution, film flatness, sharpening and post-processing (all raw scans are dull, just as any raw digital file is dull).

That's why I am often amused when someone, based on a single image or two, will render an absolute verdict on a lens, without taking into account the steps involved in transferring that image from a latent film negative image onto a computer screen.

Even in wet darkroom there are so many variables -- enlarger lens quality and setting, flatness of negative, paper grade, accuracy of focus.

I'm attaching a couple of portfolio pictures and section blowups. The interior of the church was probably taken at f/3.5 or f/4. The grain along the edges does not appear to be in perfect focus on this scan of a 6.5x9.5-inch fullframe print. The portrait of the soldier, also a fullframe 6.5x9.5-inch was probably taken at f/5.6 and has a sharper print to work with as well.

[VinceC]

Couple more detail shots

[David Murphy]

Nice shots. I'd do anything to have that guy's job -- unfortunately I'm nearly 51 and they'd just laugh if I applied.

I agree the variables are many -- only controlled tests are definitive. I used to test lenses by shooting a roll of K64 and projecting the results -- damn few variables there. I may go back to that -- it's just takes a lot more time.

[VinceC]

>>Nice shots. I'd do anything to have that guy's job -- unfortunately I'm nearly 51 and they'd just laugh if I applied.<<

Thanks.

Montgomery Meigs was a 46-year-old brigade commander with 1st Armored Division when I took the picture in 1991. He went on to become a four-star general and retired a few years ago. He is a descendent of a Union general also named Montgomery Meigs.

[David Murphy]

Interesting. I just read his bio. About the only thing we have in common is that we overlapped at M.I.T. in the early 80's.

[raid]

David: Your lens looks great on the Bessa. I have the Canon 28mm/3.5 and not the Nikon. I wonder if anyone knows of a comparison between these two lenses.

Raid

[David Murphy]

Originally Posted by raid

David: Your lens looks great on the Bessa. I have the Canon 28mm/3.5 and not the Nikon. I wonder if anyone knows of a comparison between these two lenses.

Raid

Raid I used to own the 28/3.5 Canon. I foolishly sold it to buy a 28/2.8 Canon which was sort of a dissapointing lens. A sytematic comparison of the Canon and Nikkor 28/3.5's would be interesting.

[VinceC]

I'll bet the Canon and Nikkor are pretty close. According to Peter Dechert's book, the Canon 28/3.5 was introduced in 1951. The Nikkor 28/3.5 came out in '52.

[David Murphy]

At least some of my problem is due to my Dimage scanner and how i've been using it -- working on that issue now and getting some impoved results.

[VinceC]

Now I envy your scanner. I've never gotten around to getting a negative scanner. My Epson 3200 flatbed does okay with b&w negs but not well with color negs.

[David Murphy]

The best results I've had are with color slides -- especially E6 -- scanners sometimes get confused by Kodachrome. Color negs work pretty well, but flatness is always an issue. The slide mounts help keep the film planar.

I'm basically getting better results now with the Dimage, but the images are huge and can't be easily shared -- still working on that -- standby.