Nikkor 20/2.8 AFD on D200 & F90X

[bruno]

As said I didn't do much shooting until now with the 20mm on the D200... anyhow,

I just took a shot done at 100 iso f5.6 and done some pixel peeping on it.

 

Here the full frame, being 500 pixels wide, it looks obviously good.<div></div>

[bruno]

Now just a crop from the center area.<div></div>

[bruno]

The center crop looks good enough to me. A crop from the far right border, instead, shows that it is really soft.

[bruno]

...ops, here:<div></div>

[bruno]

The crops are 100%.

 

Now I took a shot done with the 20mm on the F90X last summer. It is a negative scan. My neg scan usually look worse than my prints, but my flatbed scanner is not as good as my Minolta neg scanner... so I opted for a neg scan.<div></div>

[bruno]

And of course let's take a 100% crop from the center area and the border... the neg scan is approximately 9 megapixels.<div></div>

[bruno]

and a crop from the border...<div></div>

[bruno]

After I did all this work I just realized that the comparison is done on very different kind of images... so I took a similar one from the D200. This is again a bit unscientific... the first D200 image was shot at 100 ISO. The F90X photo is taken on TriX film.

 

The following image is taken on D200 with ISO 800 and a time of 1/40 at f2.8... don't imagine an amazing sharpness.<div></div>

[bruno]

Of course the small one looks good... so here the crops:<div></div>

[bruno]

And from the border<div></div>

[bruno]

Well. These portraits show much more the kind of shooting I do than the first picture.

 

If I look the fabric of the pants of the guy on film and the shirt of the guy on digital, I wouldn't say that I'm loosing much detail... high ISO noise is being much more nasty...

 

...I think I'll keep the lens and maybe add the 12-24/4 (Nikon or Tokina) to have a wider reach... although, being both f4 I judge them as a sub-optimal solution anyhow. This pushes me to get the Tokina instead of the Nikon for half the price and sell it if Nikon ever comes with a 16/2.8 DX prime. The 20 will serve me well for lower light work.

[Ed_Ingold]

This looks like a focus, shutter speed and cropping problem, not to mention an apples-to-oranges comparison with film, rather than a lens issue.

 

From the center crop, it appears you are focused on the building (the blinds are sharp, the person is not. In the side crop, the person is about the same distance and through a glass partition (or is it plastic). Considering the low light, slow speed and f/5.6, I suspect you are shooting at 1/30 to 1/50 second - marginal considering the crop factor. finally, you are taking a tiny portion of the image, compared to the N90 shot showing the texture of the flannel shirt. Furthermore, the B&W film image is grossly oversharpened.

 

There's also a psychological component to this evaluation. From a composition point of view, it's hard to make use of a wide lens in a meaningful way - to compose to the edges of the frame. In a picture like your DSLR example, there's no compositional focal point. Consequently, you tend to over-examine the details, which are much smaller in this view.

 

Just my $0.02.

[bruno]

So...

 

the first pic is not ISO 100 but 400. I did a mistake there, the speed is 1/90th, enough to have sharp images. If you see the pole, it is just as unsharp as the lady behind the glass.

 

I cannot remember if the focus is on the building or the guy, but at f5.6 they are both well within the range of hyperfocal distance for the lens.

 

I wanted to compare how the 20mm performed on film vs digital, since this was the matter of another post. I can't understand what you find wrong in this.

 

The crop factor won't change the circle of confusion, so I cannot see the problem there... the files of f90x and d200 are more or less of the same size (say less than 10% difference)... the bw image from film is not sharpened, the digital is, but a tiny amount, noise is much a bigger issue.

[Ed_Ingold]

Perhaps I wasn't clear. You weren't focused on the pole, the leaner or the lady, you were focused on the building in the background, which is sharp.

 

The final B&W example is sharpened, or printed at very high contrast (having the same effect) - there are halos around the pronounced grain in the image. The grain itself tends to exaggerate edge contrast. The digital image is smooth by comparison. Hence, and apples-to-oranges comparison.

 

I do not see noise of any merit in the digital image. Certainly someone who seems to like B&W grain (myself included) would have nothing to complain about in this regard. Most digital naysayers, having never worked with large or even medium format film, would tend to cite a "plastic" appearance for lack of noise/grain.

 

Finally, you introduce the spectre of "circle of confusion", which is irrelevant to your example, and erroneous in your interpretation. Depth of focus (or better, depth-of-out of focus) depends on the end-to-end magnification. Meanwhile, out of focus is ... out of focus.

 

In short, a lens like the 20/2.8 will behave as well on a DSLR as with film (been there, done that). Your examples do not contradict this observation, and in fact, are totally without merit to this end.

[bruno]

Edward, you were clear, but try to listen as well.

 

1. The lady is at a distance from me that is between the boy and the building.

 

2. The hyperfocal distance of the lens at f5.6 ranges from 2 meters to infinity. The boy is at least 4 meters from me. Maybe the building is sharper, but the lady is around 6 if not more.

 

3. I didn't make any comment on noise, except that it works AGAINST detail. Just as grain.

 

PS. I made just a few test shots in my kitchen, having the barcode of a bottle on the center or at the edge of the frame. The difference is dramatic.

 

I will not post the results because you will claim that they are meaningless anyhow.

[leland_bolleter]

One more issue. The 3rd picture in this thread, which may not be apparent, looks to me as to be through glass. That being the case... I'll stop there.

[wj_lee]

On your corner crop. The women isn't very sharp however the horizontal lines are pretty sharp which suggests the lens in focused slightly behind it.

[bruno]

That part through glass was an unhappy choice.

 

The pole is not taken through glass... just above the lady there are some writings that are etched on the back side of the glass and are just as fuzzy... you can compare the detail of the pole on which the guy is standing against the pole holding the glass where the lady sits.

 

Same applies for the brick building in the upper right corner.

[bruno]

Ok... I give up & go to bed... do some more shooting on the weekend and hope to prove myself wrong :)

[wj_lee]

Bjorn Rorslett said 20mm afd seem to have slight curvature of the field. Where as AIS version doesn't. If you are happy manual focusing may be try an AIS version.

[ofey_kalakar]

What about showing a crop of the sign as the center focus point and the brown building in

the upper right corner as a corner crop from the first picture? These shouldn't show any

artifacts due to light dispersion through glass or plastc as might be apparent in the crops

you chose for your comparison.

[bruno]

Well, here another crop. To me it looks pretty ugly.<div></div>

[bruno]

And another unscientific experiment done using a bottle of wine and shooting first at f2.8.<div></div>

[bruno]

then at f5.6<div></div>

[bruno]

then centering the bottle<div></div>