Nikon's Hot Mega Zoom.
When the D200 came to market in mid-December, 2005, another Nikon product hit the shelves on the same day. The Nikkor 18-200mm zoom lens with VRII technology on board.
Question Number One:
Its aperture ranges from f/3.5 to f/22 at the widest to f/5.6 to f/36 at the telephoto extreme, behaving just like a 27-300mm zoom would on a full-frame or 35mm film body. It's one of their DX series, so it was designed from the ground up to behave correctly on the smaller APS-C size image chip found in all Nikon DSLRs.
If such a lens were possible and/or practical for 35mm film bodies, you might not ever want one, since it would necessarily be 1.5 x 1.5 x 1.5 times larger. In layman's terms, it would expand by 50% in every dimension, and that would force its weight skyward by a factor of... um, lemme see... yes, 3.375X the current lens's 560g (1.23 lbs), so the film camera or full frame version of such a thing would be 1890g (4.61 lbs) and cost $2400 at least. Scaling price up isn't really the same as scaling weight, so the Big Version would probably set you back $5000 to $10,000. Not kidding, but I have digressed.
The lens is the debut recipient of Nikon's latest refinement in Vibration Reduction technology, VRII (vee-are-two, not vee-are-ee) and it is purported to add yet another stop worth of shutter speed steadiness to your fingers for hand-held shooting. Previous VR implementations were said to add three stops of longer shutter speed as it ironed out the vibrations your nerves, heart and native jumpiness always imparts to the camera when you hold it in your hands. So now you have presumably four stops of greater clarity.
To give you a performance edge, the lens has two kinds of VR; Normal and Active. Normal is like previous implementations and Active is for times in which you are bumping along in something that is bouncing. Use it for horseback shots, shooting from a vehicle or small plane, while walking, in a boat, during an earthquake or hurricane (or any FEMA-sanctioned disaster) or any time you are being tossed around along with the camera.
Normally, VR detects small ANGLE changes and compensates for them by moving elements (the reddish rectangle) off the optical axis in exact proportion to the upset perceived by micro sensors. It takes special actuators, a computer and some very clever programming to do this at all focal lengths, but we live in an age that sees routine miracles. The blue elements are aspheric and the yellow ones are ED glass. Both of those ideas are more expensive to implement.
But what if the whole photographer were being jostled? Sensors that read angle might be fooled by lateral and vertical displacement of the entire camera that weren't really the sort of twisting motion that hands usually infect a camera with.
With Active selected, the lens measures not just angular changes, but physical translations up, down, left, right, forward and back, then it factors those motions into the stability computation. Other VR lenses can be fooled by these motions, the 18-200 compensates for them.
Other 18-200mm lenses exist in the world, and I find it rather interesting that the cross section of all of them are similar and different at the same time. The Nikkor has 16 elements in 12 groups. Sigma's and Tamron's have 15 elements in 13 groups. Tamron's lists for the same and Sigma's is considerably less cast. Neither have a VR-like function. Just thought you'd like to know.
The Nikkor 18-200 is what I think of as being "High performance."
Here's Nikon's published MTF charts, (red is 10 lines/mm and blue is 30 lines/mm, both wide open and solid lines are measured along a line radiating from optical center while dotted lines indicate performance tangentially to radial lines). Compared to other lenses, this one is doing pretty well.
When shooting real-world images, the 18-200 is one slice lower in quality than my 55mm Micro Nikkor and at f/11 both tele and wide make very nicely defined, sharp images on the D200. Both ends of the zoom resolve very small details often under two pixels across. Camera processing always smooths the finest details, but it's not difficult to find examples at the nearly-one-pixel resolution. And since one pixel is the Nyquist Limit, there will be no further resolution.
This tiny leafy slice is 1:1 pixels and although it is small, it demonstrates detail, color trapping and a virtually complete lack of sharpening artifacts. Rollover it with your mouse to see the effect of adding Unsharp Masking in Photoshop at a setting of 90, 0.3, 0--a very mild, but just slightly noticeable common tweak. Missed it? Very subtle indeed.
At wide open, it's a tad softer, but who among us isn't?
I've noted in the past that some cameras tout a zoom ratio that was defined by the marketing department more than by the engineering staff. The Coolpix 5700's so-called "8":1 zoom turned out to be more like 7.3:1 (the engineers DID name it).
So I was curious about the 11.1:1 zoominess of it I read in the brochures and web pages. I tested it. The results were mind-blowing, but not in any way you might imagine.
Setting the camera up on a tripod, I arranged a yardstick to exactly span the width of the image along the viewfinder's center lines at full wide zoom. I'm one of those folks who really prefers to have the viewfinder lines switched on so I can do estimates of horizon, proportion, verticality, etc.
After a few tests, I had it filling the frame, dead level and observed the results on my computer. Since the viewfinder only shows 95-96% of the image, it took a little tweaking. Then I zoomed in to the center of the yardstick and kept shooting. Laying the zoom-in over the wide-zoom images, I shrank the zoom-in until it exactly fit over the same chunk of yardstick in the wide shot. You can do this in Photoshop with the Transform > Scale tool by entering numbers in the header bar data boxes, so the exact numerical amount of scale can be tracked.
At a scale reading of 17.5%, the closer image exactly fits (within reason) over the wide shot. B'b'but wait a sec! That isn't even CLOSE to being an 11:1 zoom! In fact, that's almost a perfect 6:1 zoom. What the @#%&$! is going on here?
One of the cool features of many zoom lenses these days is their internal focusing elements. You focus and the front element does nothing. It's all being handled internally. Polarizing filters love it. Nothing rotates but the focus ring itself. Still, when that 18-200 mm lens widens out, it gets mighty short, and when it telephotos in, it grows like bald face lie from Pinocchio. Internal element focus doesn't grow or shrink the lens the way it does with a prime, but it does alter the size of the image.
Well, hey. All focus operations alter the lens's focal length. The closer you focus, the longer the apparent focus. It's just physics. But internal focus designs do NOT have to obey the same set of rules. And this IS the case with the 18-200. As you focus on nearby subjects, the apparent focal length actually does the opposite of what you see with a prime lens. Primes get longer--more tele--as they focus closer. The 18-200 widens out.
Net problem: Zero. It's just the physics of this particular lens prescription. But it's good to remember that it becomes a 6:1 zoom lens at more macro shooting distances.
In the real world, the 18-200 is closer to the ideal expressed by the Marketing department. Keep in mind, the Marketing Department didn't start their careers as camera testers.
The image above is a small central chunk of a 100% scale image. The whole image is just over 600% wider than this. Rollover with your mouse to see how small we had to shrink the Tele Zoom shot to fit onto this background. Keep in mind that the picture extends several feet larger than the chunk you see here. The inverse of that reduction factor = the true zoom range.
By reducing the Zoom-In frame to 9.4% we find that the zoom range is really 10.64:1. Close enough to the 11:1 range most folks report it to be. But that Marketing Department! Sometimes I think they are constitutionally incapable of not fudging the numbers. They call it 11.1X. Geez, Nikon, you had me at 11:1.
Why does the center get sharper on rollover? C'mon. It's a 10MP image down-sampled to about 2.5 inches across. Of course it gets sharper.
If you can spring for this level of lens, I doubt you will be disappointed. At f/8 to f/16 it achieves quite good sharpness and the auto features are worth the extra investment. It's small, making it a good all-around choice for portraits, candids, sports and even nearly macro work inside a business card size frame. (Closest focus is 0.5m--20 inches--from the image plane.)
By not being limited to a small portion of the desirable zoom range, it stays glued to the camera most of the time, rarely coming off for interchange with other optics. That makes the dust exposure quotient go down.
I get letters. Whadda ya mean that zoom is anywhere NEAR as sharp as the renowned 55 Micro Nikkor? You nuts? It's not even in the Same League!!! (I cleaned that up.)
Okay. Let's see what reality is versus the Mythos of Photographos. No sense letting people believe in false gods, I always say. Well, not always. I do make time to say other things.
The test: A convenient distant vista at optimum f-stops for the 55mm Micro and the 18-200 set to the same framing, gauged through the viewfinder. As it happened, the file data reported the Zoom Nikkor as 3mm off target, and there was a small size discrepancy between the shots, but not enough to throw the test.
What we are really doing here is visually measuring the ability of various optics to resolve picture details on one specific imaging chip--the sensor in the D200 (this time).
You're looking at a 700 x 500 pixel crop out of the test images presented at 100% scale. This is roughly equivalent on most modern LCD screens, which have 90-100 pixels per inch of display, of peering at a piece of a 1 meter wide (40 inch) print. In other words, it's an extreme blow up-by anybody's standard. Not only that, but you're looking at it from screen-reading distance, probably in the under 0.5m (18 inch) zone of comfort. You probably don't look at 40-inch prints that close. Most images are comfortably digested by human eyes at around a viewing distance equal to their longest dimension--at minimum.
Now for Question Number One: Which lens shot the above image? Roll your mouse over the picture and you see the other lens in action. Question Number Two is this: How close to each other do they look in performance? Both are f/11, which seemed to be optimum for the 18-200. The Micro Nikkor 55 performed virtually identically from f/5.6 (its maximum f-stop is f/3.5) up to f/16 and the Zoom was slightly less sharp at f/5.6. The slightly larger image is the Micro Nikkor.
An interesting side note: the Micro Nikkor performs better in the corners than the zoom, of course.
There will doubtless be purists who complain that he Micro Nikkor is indeed a tad sharper, more detailed and perhaps more "transparent" than the 16-element zoom lens, and as far as that goes, they would be right.
But not one full pixel's worth of worry.
The distant power tower when viewed in extreme close-up shows well-resolved struts at nearly the Nyquist Limit (which says one pixel is all you get for maximum detail--notice the arrows on the 200% blow up, right) at various places along its height.
When a detail is in the one-pixel range and happens to land on an individual photosite of the imaging chip, the best it can do is make a detail one pixel wide or tall. An antialiasing filter over the sensor physically spreads this effect a little, to avoid harsh, stepped detail, but the majority of the light is concentrated in the target pixel.
On a print 40 inches wide, you might possibly be able to tell that one lens was delivering more detail in a static landscape like this, but for shots of your kids, Olympic athletes, evidence photos, catalog shots, excursions into photojournalism or pictures of a babe, the overwhelming versatility, range, stability and convenience of the 18-200 DX VRII wins the day, hands down.
My largest printer delivers 19-inch shots in the longest dimension of the paper, and on images this big, neither lens is obviously better than the other.
Viewing digital images at the one-pixel viewing level reminds me of the bumper sticker "If you can read this--you're too darn close!"
Numbers reveal different sorts of qualities than do these photographic appraisals. We like to pour over the numerical findings just as much as anyone else, so here is the most interesting site for reading about lenses in numerical, test-results form:
Be sure to compare this lens to others of similar range and/or prices.