Comparing the bokeh of 3 Nikkor 105mm (and a 58mm Noct)

[Airy]

Now concerning the field curvature issue : first pic, shot at 5.6 with the 105/2.5 AI. Given the forward field curvature, you'd expect the top right corner to be significantly unsharp. Well, it is hard to judge given the subject, but that corner looks reasonably sharp.
Pics : full, center, top right, bottom left. All with Df.

[Airy]

Second series, with the 105/2.5 AIS, one year and a half earlier. Same details as above ; aperture was f/8.

[Airy]

Footnote : the processing (using LR) left default sharpness settings. Nowadays, I'd rather use Capture NX-D and higher sharpening settings (base setting = neutral, contrast +1, sharpness +7 which is +/- the limit before getting halos)

[Tristin]

Everything in this image that needs to be in focus is on a reatively slim focal plane, it would have to be some extreme curvature to exhibit here.  This is not an image the f/2.5 will struggle with.  The f/2.5 struggles in having objects in focus across a large focal plane.

Excuse the image hijack, but this is an example of the type of image the f/2.5 struggles with.  It was shot with the f/1.8 a f/11 I believe.  Had the f/2.5 been used, the building in the foreground certainly would have been oof still in the corners. 

I pulled this image from my instagram, so ignore any artifacts or whatnot.

[Airy]

Not sure how field curvature would reveal itself at f/11. The following is a comparative example of a shot taken at f/2.8, at a 2m distance.
First : overall pic
Second : right edge, 105/1.8; the book with the black cover (approx. in the same plane as the center book, on which the focus was set) is about sharp ; the writing on the yellow egg is fuzzy.
Third : right edge, 105/2.5 AIS; the writing on the egg is sharp, despite being approx. 4cm in front of the book backs; the writing on the black book is fuzzy.
Fourth : stopping down at f/8, and showing the right edge taken with the 1.8 (left) and the 2.5 (right). The effect of field curvature is +/- annihilated or, more precisely, the reference plane is in focus, while the egg is sharp (a bit better with the 2.5, no wonder).

[PeterN]

Aguinaldo, thank you for the comparitive shots (also with other lenses). I am very impressed by your website and photos.
Airy, thank you for adding your comparisons as well.
Very helpful indeed!

[atpaula]

Aguinaldo, thank you for the comparitive shots (also with other lenses). I am very impressed by your website and photos.
Airy, thank you for adding your comparisons as well.
Very helpful indeed!

My pleasure Peter.
I'm glad you like my work.

[Roland Vink]

The field curvature of the 105/2.5 at close range (typical portrait distances) may be one of the reasons why this lens works so well for portraiture, as it helps to produce a softer, rounder background rendition.

[Roland Vink]

Here the announced evidence concerning bokeh. Shooting distance is about 1.5m. Camera : D800.
Second pic : 2.5 AI and 2.5 AIS at f/4. This is the most I could get in terms of bokeh difference

This picture clearly shows the effect that the shape of the aperture has on bokeh, influencing the shape of de-focused highlights. However, since the lenses are optically the same, the correction for aberrations (especially spherical) is identical, so the smoothness of background and foreground rendition is the same.

Many reviewers make the mistake of assuming a lens will have good bokeh because it has rounded aperture blades. This is only partly correct, a lens may have a perfectly round opening but still have awful bokeh - harsh, nervous, double-lining - since the rendition of backgrounds is due to optical corrections, not the shape of the opening.

[Akira]

What Roland says.  The problem with the Ai"s" Nikkors in general is that the aperture blades make very harsh, gear-like septagons or nonagons when stopped down by around one stop which make the out-focused point light sources ugly.  This was clearly presented by the images posted by Airy in reply #14.  And this is a bit annoying because stopping down by one stop is a preferable way to render the focused area sharper by keeping the shallow DOF.

[Tristin]

The curvature "plunge" I noted on the copy of the Ai f/2.5 manifested distinctly in the extreme corners, and not the borders much.  As I crudely drew on these examples.  With the example of the buildings, f/11 was the widest aperture I was able to use without the lower right corner going oof.  Which happens to be right before diffraction becomes visible on both the f/1.8 and f/2.5.  Had I used the f/2.5 I almost certainly would have had to resort to f/16, at least, to save that corner and subsequently inviting diffraction.  The curvature on the borders, as you focused on, was not as extreme as the corners and I did not find these areas problematic in the f/2.5

I don't think the curvature plunge in the corners itself adds to portraiture, or anything really.  It just may force one to stop down further in some instances.  Overly nit-picky to most I am sure, and not at all the use people pick up the f/2.5 for.  I just happen to be striving for a very minimal and flexible set of primes.  I am interested to see any tests on this to deduce if it was the copy I tried or not.

For the future, I need to be more clear and give examples early on so these conversations can be spared plenty of fat.

[Airy]

Tristin, forward field curvature should actually help getting the bottom right corner in focus, in your quoted example.

Whether forward field curvature helps achieving better (central) subject isolation is an interesting question. I mostly saw such statements in relation with wide angles, by the way, where such effects tend to be more apparent (see: Zeiss 28/2). In the present case, I am a bit skeptical as you are.

Akira, Roland, I think that the main problem with "bad bokeh" is double edging, itself related to outlining of OOF points. This is the primary issue. The secondary issue is, outlined OOF points will in addition make the polygon shape of the aperture all the more apparent. This is where the 105/2.5 AI fares better. But all versions tend to outline the OOF points, i.e. bokeh can be harsh, depending on the background.

I'll try a systematic comparison - shooting reflections on beads of various sizes at various distances lit by one or several point light sources should help exacerbate the problem (the dreaded jewellry product shots, so to say, but at more usual distances).

[Airy]

The field curvature of the f/2.5 limits it's use as a landscape lense heavily and its poor handling of strong light sources in front of the lens (compared to the f/1.8 ) makes it a far worse lens for many uses.  I would not trust the f/2.5 to shoot live music, for example, as washed out images from stage lights would be the course for the night.

Finally I got an opportunity to shoot jazz musicians, and I had the 105/1.8 in the bag (together with the Voigt 58/1.4, unfortunately not the Noct Nikkor).

It seems that you were right about the qualities of the 105/1.8 with light sources in the field. I'll open a thread when the processing is finished.