Clash of the Titans: Printing Nikkor vs Repro Nikkor f/1.0

Have questions about the equipment used for macro- or micro- photography? Post those questions in this forum.

Moderators: rjlittlefield, ChrisR, Chris S., Pau

Chris S.
Site Admin
Posts: 4042
Joined: Sun Apr 05, 2009 9:55 pm
Location: Ohio, USA

Post by Chris S. »

Lou, capital test! :D (But am I biased because the Titan in my lens park won?)

Couple of thoughts:
  • 1) You tested the RN85 at its best aperture (f/2.8 ) against the PN105 at f/2.8. At 1x, my PN105 specimen improves notably when stopped down half a stop to between f/2.8 and f/4. Your test showed the PN105 as better but close to the RN85; if your PN105 is like mine, might not comparing the two lenses at each one's best aperture show the PN105 to be better by a larger margin? (No disrespect--lens testing is bloody hard, while nitpicking tests is easy. :? )

    2) For your 100 percent comparisons, it appears you used screen captures of the images displayed in Photoshop. This probably works fine when done as you did--shots from both lenses treated exactly the same way at the same time. So any vagaries from your computer's graphics processing should effect both images the same way. No problem.

    But I've been looking for a chance to urge against this becoming a widespread standard, so here goes: Screen grabs are impacted by a particular computer's graphics processing, and if less meticulous practitioners than you use this method with less care, it may lead to false conclusions.

    I had a private discussion on this topic with a forum member who does wonderful lens tests. He and I each took his original images and did screen grabs on our separate computers, displayed in Photoshop, at agreed-on magnifications; our results were dramatically different. I attribute this to differences in how our computers' graphics systems rendered.

    In contrast, we obtained much more comparable results when we each used Photoshop to crop his images to 100 percent segments, saved these crops as images, and compared the resultant image files. Going this route, we felt, permitted more useful comparisons between photographers.

    No criticism intended about your wonderfully-conducted test. You made certain that the variable created by computer graphics applied equally to both images. But for comparisons between photographers, or between images processed on different computers, pixel-level crops saved as images make for a more extensible comparison. I'm jumping on this chance to speak in favor of this becoming a standard.
Lou Jost wrote:I am sure there are amazing lens discoveries floating around out there. I have a bunch of strange things I bought over the last year that I have not tested carefully; some seem extremely good at first sight.
Occasionally, when we kiss a frog, it turns into a lovely princess. But most of the time, it remains a frog. To my mind, testing unknown lenses for macro is similar--odds are, one will purchase and test lot of junk lenses to find an occasional outstanding lens. For an inveterate lens tester, money and time spent purchasing and testing frogs will exceed the money/time cost of purchasing a known princess.

So agreed--hats off to the hard-working, cash-spending frog-kissers among us! What a service they do us. They probably find frogs most of the time. Then, when an odd frog turns into a lovely Kate or Meghan, they tell us about it. And then we few, we happy few, we band of brothers, can purchase such a lens! :D

Cheers,

--Chris S.

Beatsy
Posts: 2105
Joined: Fri Jul 05, 2013 3:10 am
Location: Malvern, UK

Post by Beatsy »

Chris S. wrote:...And then we few, we happy few, we band of brothers, can purchase such a lens! :D
At rapidly inflating prices! :D

Well said (to your whole post) though - hear, hear!

I won't be buying any frogs in the near term due to my recent splurge on a new PC and monitor (dead lucky I pressed the button mere hours before Ray P offered those Printing Nikkors for sale too :) ). But I found my foray with the Heidelberg Linoscan lenses a very enjoyable activity in it's own right and will definitely be tempted to dabble again - should an attractive "ribbit" sound catch my attention...

Lou Jost
Posts: 5942
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

Chris, that's a very interesting point about the use of screen captures versus actual cropped images. Do you have any idea whether "good" monitors tend to agree with each other more than uncalibrated, low-gamut monitors? It must be true that a monitor with a low color gamut will compress color gradients. I will try from now on not to use screen grabs.

About comparing the PN105A and RN85 at slightly smaller apertures, I almost did that, but the coinimaging.com tests on my copy of the PN105A showed that resolution was highest at f/2.8. And we know that these superb lenses converge on each other as we stop then down, because diffraction becomes the main issue and that is an equal-opportunity robber of resolution.My main concern was to see whether the faster apertures of the RN85 were able to take advantage of their lower diffraction. Seems not, though they did stunningly well for such fast apertures.

Yes, most lenses out there are frogs. But a little research can improve the odds. It was thrilling to find the 14-element Scanner-Nikkor was nearly the equal of the PN105A at less than 1/10 the cost! And here are still whole classes of lenses that no one seems to have tried.

As I mentioned on another thread, on my last trip to the US I tried to bring a large pile of "frogs" back to Ecuador with me, but when I tried to check in at the airport, the airline told me there was an un-announced rule of no overweight luggage allowed, so I had to leave 40 pounds of frog- or Princess-lenses and other things back there. One of those is a large heavy lens I got for $6 on eBay (I was the only bidder) which looks like an ultra-fast Myutron machine-vision lens. I have high hopes for that. Most of the others will probably be frogs, or might be princesses only in monochrome light... Stay tuned to find out.

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

Lou Jost wrote:Chris, that's a very interesting point about the use of screen captures versus actual cropped images. Do you have any idea whether "good" monitors tend to agree with each other more than uncalibrated, low-gamut monitors? It must be true that a monitor with a low color gamut will compress color gradients.
I suspect it's actually the other way around.

The reason is that "screen capture" programs don't actually capture what the monitor displays. Instead, they retrieve numeric pixel values from the graphics card. Monitors whose gamuts are larger than sRGB can only show that by working from an internal representation whose gamut is also larger than sRGB. That means the internal representation has to be compressed (compared to sRGB), so that it can be expanded again by the monitor. Think of looking at histograms in Photoshop -- the "same" image whose profile is ProPhoto or Adobe RGB will have a histogram that is a lot more compressed toward the center than one whose profile is sRGB. So, if your image was originally sRGB, then I would expect the pixel values to be altered more when using a wide gamut monitor.

Of course some or all of this effect could conceivably be un-done by the screen capture program, based on its understanding of the monitor profile.

There's definitely a lot to go wrong on the display-and-capture path these days!

I just now ran what I intended to be a simple experiment: displayed an image with Photoshop at 100%, grabbed a screenshot with Snagit, saved the screenshot to PNG (to avoid compression issues), then pulled the PNG file back into Photoshop and using copy/paste, layered it onto the original image for comparison.

I was expecting some change in the histogram.

However, what I actually got was a lot of change in the histogram, combined with a bit of change in the geometry! When I flashed to compare the two images, I realized that the top halves of the images were staying nicely aligned while the bottom halves were bouncing up and down. On closer examination, I see that the file size is 1024x671, but the captured and re-saved image is 1024x672, with one row of pixels duplicated (or nearly so) in the middle of the image. I suspect this has something to do with my unusual setting of 175% for Windows scale factor on side-by-side 4K displays. Is the glitch in capture or in display? I don't know yet, but it's not the sort of symptom that I would expect to see in any case. More investigation required...

Chris and Lou, you might find it interesting to try running the display/capture/save/compare cycle on your own systems and see what turns up. I plan to post my own results to a separate thread (after some more investigation), but perhaps you folks can beat me to it.

--Rik

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

rjlittlefield wrote:Is the glitch in capture or in display? I don't know yet, but it's not the sort of symptom that I would expect to see in any case.
It's definitely in display. A synthetic slant edge, in an image sized 1024x671, shows an easily visible step at 100% display in Photoshop CC 19.0, when my Windows 7 is set for 150%, 175%, or 200%. Simultaneously the image size on screen is reported by Snagit (12.4.1 Build 3036) as 1024x672 -- one pixel too many vertically. This is using NVIDIA GeForce GTX 660 Ti card, with current drivers. But at Windows settings of 100% and 125%, the step is not visible and the displayed image is reported as 1024x671, the correct size. On my wife's computer, which is Windows 10 with a different graphics card but the same model of 4K monitors, also set at Windows 175%, no step is visible but the image window is only 1023x670 -- one pixel smaller in each dimension.

How bizarre!

Fortunately I have high hopes that the answer is contained in what appears to be a user review of some manual that I have been reading recently: The Hitchhiker's Guide To The Galaxy. Or if not there, then surely in Alice in Wonderland.

--Rik

mjkzz
Posts: 1681
Joined: Wed Jul 01, 2015 3:38 pm
Location: California/Shenzhen
Contact:

Post by mjkzz »

not sure how the image get displayed on a "preview" monitor, if it is via HDMI, then there are a few issues.

first being the output quality of HDMI on the camera. not all camera can output higher chroma subsampling or raw signal. for those camera with 4:2:0 HDMI output, I do not think that is enough fidelity, maybe we need at least 4:2:2, but only recent camera can do so, like the Sony A7III, Panasonic GH5, as far as I know.

second being how the image is converted to digital image so it can be displayed on a COMPUTER monitor, if digitizer is not good enough or can not stream raw data back to computer to be displayed, it is a problem. I have a USB 3.1 based digitizer, the output is in MP4 format and the quality is not good. I also have the Blackmagic Intensity Pro 4K, the quality seems to be OK, but I have not checked it as image, only as video.

third of all, not all camera can output clean screen via HDMI and output resolution is low.

Lou Jost
Posts: 5942
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Re: Clash of the Titans: Printing Nikkor vs Repro Nikkor f/1.0

Post by Lou Jost »

Resurrecting this thread, I now realize that I made a fundamental error in this test. At f/1.0 and f/1.4 (and possibly even at f/2.0), the sensor filter glass thickness will seriously degrade the image of a lens not designed to expect glass in that part of the optical path. This would not change the relative rankings of two tested lenses at f/2.8, but it probably accounts for some of the degradation of the f/1.0 lens at faster apertures.

I now have a sensor with no filter, so maybe I will re-run this test when I have time.

RobertOToole
Posts: 2627
Joined: Thu Jan 17, 2013 9:34 pm
Location: United States
Contact:

Re: Clash of the Titans: Printing Nikkor vs Repro Nikkor f/1.0

Post by RobertOToole »

Nice timing Lou. I had planned to add some new RN (repro-Nikkor) info to my Printing-Nikkor 1x Ulitimate Edition test as I was able to dig up the patent for the 100mm f/2.8 RN and it supports the info I had posted previously here:

https://www.closeuphotography.com/print ... -95mm-test

__________________________________

Other Optical Printing Lenses

There are other optical printing lenses available on used market. The Nikon Repro-Nikkor (RN) 100 mm f/ 2.8, and Kodak Printing Ektar (KPE) 103mm f/2.8 are two that quickly come to mind. You should be aware that both the RN and KPE lenses were used in ACME and Oxberry optical printers but these use a simple symmetrical double Gauss design, similar to an enlarging lens, without any low dispersion or esoteric glass. The KPE is a modified Double Gauss 7 element, 3 group optical design, the RN is a 12 element element design. If you are interested in performance, I would leave the RN and KPE to collectors since they are offered at asking prices higher than a Printing-NIKKOR.

Some of the RN lenses are extremely rare, the Repro-Nikkor 85mm f / 1.0 for example, was produced starting 1968 with a production run of no more of 200 units total. According to the writer Marco Cavina, the Nikon RN is the forerunner and the basis for the design of the later new and improved Scanner-Nikkor ED and Printing-NIKKOR lenses.

__________________________________

The patent confirms the above idea that no exotic low-dispersion glass was used in the 100mm f/2.8 RN design.

Below is the lens element layout of the RN100 from the 1968 patent photocopy (slightly crooked copy) on file at the patent office. This is before there was even a Nikon USA around. All Nikon RNs were sold through an importer in NYC that was later bought out by Nikon Japan to start Nikon USA.



__RN100.jpg

RobertOToole
Posts: 2627
Joined: Thu Jan 17, 2013 9:34 pm
Location: United States
Contact:

Re: Clash of the Titans: Printing Nikkor vs Repro Nikkor f/1.0

Post by RobertOToole »

Can't resist adding a little more interesting RN info, this is not even on my site yet!

1977 Price list

PN 95161
Repro Nikkor 85 f:1
$2262.00 1-2 units ($11,483.68 adjusted for 2023 dollars)
$1441 for 25+ units

For a nice comp

PN 95168
Printing Nikkor 105 f:2.8
$2468.00 1-2 units
$1573 for 25+ units

Lou Jost
Posts: 5942
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Re: Clash of the Titans: Printing Nikkor vs Repro Nikkor f/1.0

Post by Lou Jost »

Thanks Robert for the added details. I think the RN85 is really interesting. I should add an observation that I made at the time of these tests; I recall being worried about the sensor glass, so I used a Mitutoyo 20x objective to examine the aerial image produced by the lens. That aerial image was actually sharper at f/2 than at f/2.8. So this lens is actually too good for today's technology.

Post Reply Previous topicNext topic