Author Topic: 50 mm contest, part 2 : simulated night  (Read 3651 times)

Airy

  • NG Supporter
  • **
  • Posts: 2590
    • My pics repository
50 mm contest, part 2 : simulated night
« on: August 04, 2015, 01:24:04 »
Second test : simulated night. Computer screen with starfield; backlit computer keyboard; crystal vase lit by distant diffuse source. Camera: Df.

50/1.4 AIS and 50/1.8 AIS _____________________

At f/2, there is less blooming with the 50/1.4. Highlights are outlined in both cases, but they are smoother with the 50/1.4. Highlight shapes are however distinctly heptagonal with the 50/1.4, but still nearly round with the 50/1.8. Colors if the 50/1.8 AIS are more neutral; with the 50/1.4, there is a pink tint. Keyboard letters are surrounded by a slight haze with the 50/1.8.
At f/2.8, the 50/1.4 still bleeds less. Highlights of the 50/1.8 still appear bigger.

50/1.4 AIS and 50/1.2 AIS ________________________

At f/1.4, both yield the same color. The highlights, with the 50/1.2, are slightly smaller and have softer edges, compared to the 50/1.4 that shows distinct green edges behind the focus plane. In front, the 50/1.4 highlights have distinctly blue edges, which is much less the case with the 50/1.2. This result is also partly due to the 50/1.2 forward field curvature, resulting in corner highlights to be more in focus.
At f/2, the highlights of the 50/1.2 are similarly smaller, and the 9-blade diaphragme contributes to softening the edges.
At f/2.8, same observations. Overall, the 50/1.2 is slightly better, especially at 1.4.

50/1.2 AIS and 50/1.2 AI ___________________________

At f/1.2, they are astonishingly identical. I expected very similar results, not near-identity. What about stopping down, where the 9- vs. 7-blade diaphragms might make a difference?
At f/1.4, a bit surprizingly, the 7-blade diaphragm seems rounder. The 9-blade one seems slightly irregular, and less round.
At f/2, the AI version displays some more bokeh outlining (still much less than the 50/1.4), and the heptagons are apparent now. The AIS version seems to bleed a bit less in the extreme corners. Foreground blurs are slightly less mushy with the AIS. The AIS may also exhibit slightly different colors (more magenta).
At f/2.8, same observations, with the heptagons tending to appear bigger than the enneagons. So far, the 50/1.2 AIS is the best among MF Nikkors.

50/1.2 AIS and 50/1.4 SC ________________________________

I'd expect to have the best and the worst here. Not so.
At f/1.4, the backlit keys are definitely sharper with the non-AI lens, but away from the focussed keys, the blur increases much faster that with the AIS. In other words, the SC has a thinner "night DOF". This was unexpected (not completely - in fact, it is about performance degradation towards the periphery of the image). The shape of the background highlights (enneagon vs. circle) and the stronger outlining with the SC are no surprises, the latter also being quite visible by daytime.
At f/2 and f/2.8, the old SC still has a slight sharpness advantage, and the same steeper blur increase away from in-focus zones.

50/1.4 SC and Zeiss 50/2 ________________

At f/2, the lower CA and sharper center of the Zeiss are immediately visible, as is the color (leaning more to the yellow, but very balanced). The bleeding in the extreme corners is definitely less, which may partly be due to the considerable vignetting of the Zeiss.
At f/2.8, same observations. The rounded blades of the Zeiss also contribute to preserve the Zeiss advantage.

Zeiss 50/2 and 50/1.8 G ___________________________

At f/2, the Zeiss retains its considerable advantage in terms of sharpness (in the focussed zones), but the Nikkor has an astonishingly low CA, about as good as the Zeiss. The backlit keys unfortunately do not look sharp with the 50/1.8 G  because of "bleeding". Here, the 50/1.4 SC is better at f/2! the oldtimer deserves an applause.
At f/2.8, the Nikkor 50/1.8 G bleeding is over, and both perform in a similar way, the remaining difference being the color cast (magenta for the Nikkor, yellow for the Zeiss).

What about the Voigtländers ? ________________________

Comparison with the Voigtländer 58/1.4 is made difficult by the FL differences. The 50/1.4 SC seems sharper in the focussed zones (bleeding, again, is the issue for the Voigtländer). Also, the Voigtländer sports slightly more decent blur circle outlines. At f/2 and f/2.8, the sharpness advantage of the SC recedes and the blur circles become similar, except that they are now heptagons with the SC, while they remain circles with the Voigtländer.
As for the 40/2: funny guy. At f/2, sharpness of the in-focus keys is as good as with the SC, but coma seems to be higher. Above all, the blur circles are outlined heptagons in the case of the SC, but atolls with a coral reef and a central island with the 40/2. Call that a "special effect". Color neutrality of the 40/2 is however excellent. At f/2.8, the 40/2 becomes much less extravagant, and blur circles are now near perfect, with a nice round shape and very discrete atolls.

Conclusion : Zeiss is pretty good; 50/1.8 G is the runner-up for color, low CA while 50/1.4 SC is the runner-up for center sharpness (with fast degradation towards the edges though), and 50/1.2 AIS for overall image quality (sharpness, bokeh) at very wide apertures.
Airy Magnien

Erik Lund

  • Global Moderator
  • **
  • Posts: 6480
  • Copenhagen
    • ErikLund.com
Re: 50 mm contest, part 2 : simulated night
« Reply #1 on: August 04, 2015, 09:23:34 »
Thanks for the effort!

50/1.2 AIS and 50/1.2 AI ___________________________

At f/1.2, they are astonishingly identical. I expected very similar results, not near-identity. What about stopping down, where the 9- vs. 7-blade diaphragms might make a difference?
At f/1.4, a bit surprizingly, the 7-blade diaphragm seems rounder. The 9-blade one seems slightly irregular, and less round.


50mm 1.2 Ai should have 6 blades...

Also; This is why Roland added the info about rounded aperture blades on AI lenses on his site with Nikkor lens info.
Erik Lund

Frank Fremerey

  • engineering art
  • NG Supporter
  • **
  • Posts: 12334
  • Bonn, Germany
Re: 50 mm contest, part 2 : simulated night
« Reply #2 on: August 04, 2015, 10:09:30 »
The files promised, enjoy comparsion, names are speaking (both on the D600):

http://ZENTRALKRAFT.COM/ed20_50ai_OBF_0567_v1.jpg

http://ZENTRALKRAFT.COM/ed18_50g_OBF_0571_v1.jpg

I feel both are very much in the same league with a tendecy towards the 1.8/50G

You are out there. You and your camera. You can shoot or not shoot as you please. Discover the world, Your world. Show it to us. Or we might never see it.

Me: https://youpic.com/photographer/frankfremerey/

Frank Fremerey

  • engineering art
  • NG Supporter
  • **
  • Posts: 12334
  • Bonn, Germany
You are out there. You and your camera. You can shoot or not shoot as you please. Discover the world, Your world. Show it to us. Or we might never see it.

Me: https://youpic.com/photographer/frankfremerey/

Airy

  • NG Supporter
  • **
  • Posts: 2590
    • My pics repository
Re: 50 mm contest, part 2 : simulated night
« Reply #4 on: August 04, 2015, 11:59:44 »
Thanks for the effort!

50/1.2 AIS and 50/1.2 AI ___________________________

At f/1.2, they are astonishingly identical. I expected very similar results, not near-identity. What about stopping down, where the 9- vs. 7-blade diaphragms might make a difference?
At f/1.4, a bit surprizingly, the 7-blade diaphragm seems rounder. The 9-blade one seems slightly irregular, and less round.


50mm 1.2 Ai should have 6 blades...

Also; This is why Roland added the info about rounded aperture blades on AI lenses on his site with Nikkor lens info.

I just checked Roland Vink's files. All f/1.2 lenses have either 7 or 9 blades. The 50/1.2 AI exists only in the 7 straight blades version. My copy bears the ser. No. 191682 and confirms Roland's publication on that point.
Airy Magnien

Erik Lund

  • Global Moderator
  • **
  • Posts: 6480
  • Copenhagen
    • ErikLund.com
Re: 50 mm contest, part 2 : simulated night
« Reply #5 on: August 04, 2015, 12:16:23 »
Could only count to 6 this morning...  :o sorry for the confusion!
Erik Lund

Roland Vink

  • NG Supporter
  • **
  • Posts: 1523
  • Nikon Nerd from New Zealand
    • Nikon Database
Re: 50 mm contest, part 2 : simulated night
« Reply #6 on: August 04, 2015, 22:04:13 »
50mm 1.2 Ai should have 6 blades...

Also; This is why Roland added the info about rounded aperture blades on AI lenses on his site with Nikkor lens info.
The  AI 50mm 1:2 has 6 aperture blades, the AI 50mm 1:1.2 has 7, the AIS 50mm 1:1.2 has 9. If anyone can send me a picture showing the AI 50/1.2 aperture blades are curved, I'll update my site.

Airy

  • NG Supporter
  • **
  • Posts: 2590
    • My pics repository
Re: 50 mm contest, part 2 : simulated night
« Reply #7 on: August 04, 2015, 22:17:30 »
I'd like to satisfy your request, but the more I look at these, the straighter they are.
Airy Magnien

John Geerts

  • NG Supporter
  • **
  • Posts: 9119
  • Photojournalist in Tilburg, Netherlands
    • Tilburgers
Re: 50 mm contest, part 2 : simulated night
« Reply #8 on: August 04, 2015, 22:26:43 »
I think Roland means the Nikkor 50mm F/2.0 which has six curved blades.

Roland Vink

  • NG Supporter
  • **
  • Posts: 1523
  • Nikon Nerd from New Zealand
    • Nikon Database
Re: 50 mm contest, part 2 : simulated night
« Reply #9 on: August 04, 2015, 23:10:44 »
It's true the aperture blades of the AI 50/2 are curved (I just checked). Very few Nikkors have dead straight aperture blades (the 35/2.8 PC comes to mind) so you could say that most lenses have "curved" aperture blades. But in most cases they are not curved in a way that makes the aperture opening pleasantly rounded. Many of them have a slight S-shape curve, so they nicely fit around the aperture when fully open, but have tips which straighten out or even curve back slightly when closed down. When stopped down this gives a polygonal opening with slightly wavy sides, becoming straighter as the blades close down more. Also in some models the aperture blades are a bit short so at wider apertures they don't overlap fully, and the rounded tips intrude into the shape of the opening giving a saw-tooth effect.

Among the lenses with curved aperture blades which do produce a nicely rounded (but not fully circular) opening are the gaussian 105/2.5, 28/2, 58/1.2 Noct (all excluding AIS versions) and later pre-AI 135/3.5 versions. There may be some others I missed.