brian_caldwell

The 17-35/2.8 may be your best bet because it has exceptional close-up capability, and has far less distortion than any Nikkor prime lens in the 24-35mm range. I often use this lens for closeup shots. The zoom feature may also be extremely useful in your application.

"And wouldn't the Series 1 90mm work well with the PN-11?"

 

If you focus the Series 1 90mm down to 1:2 and then add extension to get to 1:1 you wind up with plenty of coma when used wide open. However, this coma pretty much disappears if you stop down to f/5.6 or slower.

 

The 1:1 adapter that comes with the lens is essentially an extension tube with some optics to correct the coma I mentioned above. Its not accurate to call it a teleconverter.

If you can manage to remove the rear element of a 50mm f/1.8 and re-install it backwards you will be rewarded with a TON of astigmatism and over-corrected spherical aberration and field curvature. Even stopping down to f/16 won't produce a sharp image, except in the center of the field. Color correction will remain fairly good, however

 

You should be able to find a fungus-ridden sample for about $10 to try out this experiment.

"Why the non IF version is better then the IF?"

 

The non-IF version is a very simple design that is extremely well corrected for distant subjects. If my information is correct it also uses two large ED elements in the front group to achieve essentially perfect axial color correction. The sample I have is certainly a very high performance lens.

 

I don't know nearly as much about the IF version, but David is correct that it only has a single ED element, in this case one of the two large positive powered elements near the front. This alone would make it much more difficult to achieve the same degree of color correction as the non-IF. The need to incorporate an effective internal focusing group would place additional strain on the design.

 

I suspect that the ED non-IF version was more expensive to make and less popular due to the bulky focus movement, and this explains why it is now a rare lens.

"There is a 300/4.5 ED that is a non-IF. Bjorn Rorslett rates this one very highly."

 

The 300/4.5 ED non-IF is a great lens. Unfortunately, now that Bjorn Rorslett has found one and accurately described its imaging quality its likely to become even harder to find one! You have to be very diligent in sifting through ebay to find this lens because it is nearly identical in appearance to the non-ED non-IF version.

The 50-300 ED is a great lens, especially near the center of the image field. I prefer it to my 80-400 VR whenever I can use a tripod. Don't bother with the really old non-ED version unless you want a soft-focus look to your images.

"It will depend on the physical size of the aperture, so you can't really describe it as a f-number."

 

The effect of diffraction on image quality depends on the f-number, and not the actual aperture size. I think you may be confusing spatial resolution at the image plane with angular resolution in object space.

The 135 f/2 AIS has a little coma at closest focus, but this is pretty much a non-issue for reduced format D-series cameras. The close-focus coma disappears if you stop down to f/4 in any case. For distant subjects the 135/2 is very good wide open, although I think the 85/1.4 has a slight edge in performance. The 85/1.4 is certainly better closeup since it uses a floating group to correct coma at all magnifications.

"Notice that PC Nikkor shots stitched together have literally invisible boundaries as long as you make sure your tripod keeps the camera perfectly still when you rotate the knob."

 

Strictly speaking, you need to keep the lens stationary, not the camera. The parallax that results from keeping the camera stationary while rotating the lens is fairly small if the subject matter is reasonably far away.

"The tower in the corrected version looks abnormally short."

 

I think the corrected version may not be corrected correctly. Panorama Tools together with a good frontend like PTAssembler can give flawless results on images like this. The LensDoc plugin for Photoshop is also good, but I think not as accurate as Panorama Tools.

There is probably a difference in the helicoid design, but this probably won't matter much with a 50mm lens. If it were a 200/4 the difference would be significant, in favor of Ai-S due to the fact that long-throw AI helicoids suffer from focus creep in certain situations.

"I scanned some cross sections and kept everything in-scale so you can look at the elements."

 

Albert:

I can't help wondering about the accuracy of these drawings. For example, the front negative element of the Series E design looks impossibly thin to me. No designer I know of would submit a design with that sort of center thickness to diameter ratio.

"The high/low (RI) glass combination is used for color corrections, right?"

 

Color correction depends on differences in optical power and dispersion. An achromat is formed from a negative lens with low power and high dispersion together with a positive lens with high power and low dispersion.

 

The use of high index crown glass (low dispersion) in positive elements and low index flint glass (high dispersion) in negative elements helps to flatten the field, and this is how the Tessar lens achieves its correction. High index glass also helps to correct other aberrations, especially spherical aberration, because the lens radii can be flatter for a given optical power.

 

Unfortunately, high index crown glass tends to have secondary dispersion properties in the opposite direction from ED glass, which means it is not generally useful in long lenses which need partial or full apochromatic correction. So, when secondary color becomes the dominant aberration you have to bite the bullet and use low-index ED glass. This is taken to extremes with UV-vis apochromats which can only use fluorite and fused silica, both of which are very low index materials.

"High refractive index glass should allow more transmission. But any given "ED"lens has only one or few of the elements in its construction with such glass. They allow the designers to use thicker elements so that better corrections can be made without sacrificing light transmission. "

 

ED glass has a low refractive index and a low but abnormal dispersion. It is used mainly to reduce the secondary longitudinal color error in telephoto lenses and telescope objectives. It is also widely used in high power microscope objectives along with fluorite for the same reason. The low refractive index is almost always a problem to be overcome rather than a benefit.

If you analyze the patent prescription for the 55mm f/2.8 Micro Nikkor (U.S.Patent 4,260,223 example #1) you will find that it is fully diffraction limited (Strehl ratio > 0.9) at f/2.8 on-axis for monochromatic blue light (470nm). Under these conditions the lens will just barely resolve about 700 line pairs per millimeter.

 

Ask a silly question, get a silly answer . . . .

A well designed and built double-Gauss is inherently superior to a well-designed and built Tessar. Period.

"To be honest, I can't say if there are any true "on-film" differences from version to version"

 

The newer pancacke style 50/1.8 AIS has a completely different optical prescription than the older long-barreled AI/AIS versions. This can be seen by a slight difference in distortion, and by a dramatic difference in flare characteristics. The new version is plagued by the infamous central hot spot problem, while the old versions are not: http://caldwellphotographic.com/50oldnew.jpg

Even if the convertor were perfect, which it is certainly not, it would magnify the aberrations of your 80-200 zoom by a factor of three. This alone will undoubtedly cause a noticeable loss of clarity, especially if you use the lens wide open.

"Diffraction stars" refer to the pointed rays eminating from the street lights in my 60 second and 15 second photographs. They arise from the corners of the aperture blade intersections forming a point of low radius around which light rays bend.

 

The diffraction spikes actually arise from the flat sides of the aperture, and not the vertices. Each flat surface in the aperture produces a pair of rays at a bright highlight oriented perpendicular to the surface. This perpendicular orientation is shown clearly in your first example photograph where ghost images of the aperture stop conveniently show up in the dark sky area.

 

Apertures with an even number of sides effectively produce two sets of rays that overlap exactly, hence the number of rays is equal to the number of sides. Apertures with an odd number os sides have non-overlapping ray patterns, and hence the number of rays is twice the number of sides. Curving the sides of the aperture polygon dramatically reduces the ray intensity, and making the aperture circular eliminates the rays altogether.

"According to David Ruether, the 17-35 is great except at 28 mm, at which the Ai-S prime beats it handily. It's clearly the weak spot of the zoom."

 

A peculiar observation, completely at odds with mine and several others.

"Has anyone actually compared how AIS 28/2.8 is compared to AFS 17-35/2.8? I've always wanted to get the AIS 28/2.8, but never got around to it. Now that I have the zoom, I can't get myself to spend money on the prime. I wonder if the lens is worth it. Actually, the weight reduction alone might make it worth the cost..."

 

If the 24/2.8, 24/2.0, 28/3.5PC and 28/2.0 are any indication (and I suspect that they are), then the 28/2.8 will have much greater distortion and color fringing than the zoom. Like you, I have a hard time spending any more money on wide angle primes, none of which can match the 17-35 zoom.

The marginal ray angle at f/1.4 is about 21 degrees. This is near the limit of angular acceptance for many sensors, and as a result the outer parts of an *on-axis* f/1.4 ray bundle will suffer from falloff. Sort of like stopping down the lens with an iris made of radially graduated neutral density material. The net result is that the exposure is somewhat less than you would otherwise expect.

 

If the marginal ray angle is steep enough you might start to get crosstalk between pixels, which would lead to color problems. However, I doubt this is too much of a problem with most sensors at f/1.4.

 

If you were to use a *really* fast lens, like f/1 or faster, then the problems might start to get objectionable.

 

I only notice minor artifacts when shooting with f/1.2 lenses wide open on my Nikon D1x. The main artifact is that defocused highights aren't round even on-axis due to the fact that the acceptance angle is different in the horizontal and vertical directions.

"Early zoom lenses mostly had front element group focusing and this prevented them from focusing very close. The front helicoid could be only so long. Examples of this design are the 80-200mm f/3.5 Konica Hexanon and the 80-200mm f/4.5 Nikkor. "

 

Front group focusing is still extremely common in zoom lenses since it automatically maintains the focus distance regardless of the focal length setting. Virtually all of the compact negative-positive zooms used in digicams these days use front group focusing, for example. Close focus is not necessarily compromised with this type of focusing, either. The Vivitar 90-180 macro zoom and the more recent Nikon 70-180 macro both use front group focusing, and both are famous for their close-focus performance.

"I wonder if there is any "optical quality" diffence between 50 1.8 AIS (early) and the current 50 1.8 AIS. Hope Dave and others can throw some light."

 

The more recent AIS is useless IMO for landscape work due to its pronounced central hot spot at small apertures. I use my new version for reverse coupling and as a loupe, and not much else. Unfortunately, the optical deficiencies of the newer AIS seem to have been retained in the AF versions, altough I don't have first-hand knowledge of this. Stick with the old AI/AIS version.

Get the 50/1.2. Its very similar in performance to the 50/1.4 at equal apertures, and is more versatile since its faster. It also makes a logical pairing with a 50/1.8 or, better, a 55/2.8.