Strodav wrote:
I have seen some very, very good shots using kit lenses lately, so have been asking the same questions. So, I set up a still scene with a lot of texture, color, an X-rite Passport and took some shots with a D850 at ISO 100, f8, 1/20s, 50mm, tripod using a Nikon FX 50mm f1.4, Tamron FX 24-70mm f2.8, Nikon DX 18-140mm f3.5-5.6, Tamron DX 18-400mm f3.5-6.3. Viewing 1:1, I could not see a difference in sharpness between the 50mm, 24-70mm, and 18-140mm. The Nikon 18-140mm kit lens was a hair less sharp, but you really had to go looking for it. Yep, the nifty-fifty opens up to f1.4, is smaller, is lighter, which are very useful things, but the 24-70mm is more flexible in setting up the shot and has VR. Also, very useful things. I am going to sell the kit lens.
I have seen some very, very good shots using kit l... (
show quote)
Just curious: how did you view the image files?
The eye tends to mistake acuity (sharp edges) for resolution (fine detail). As you enlarge an image,
the presence (or absence) of fine detail becomes obvious.
It's hard to tell how much fine detail is in a scene, and usually it varies at different parts of the image.
And there is no way to quantitize it. That's why lens test charts with resolution targets were invented.
Lens testing on digital cameras is greatly hindered because there is no standardized RAW file format,
and because diigital cameras can do a lot of processing in the camera (even to the point of producing a
JPEG file with lossy compreession). On a film camea, you can just look at the negative with a loupe.
RAW require processing before they can be viewed. Sensor geometry (typically hexagonal) has
to be converted to rectilinear in order to be displayed or printed. Colors have to be demosaiced.
There are various algorithms that give different results. So it is very hard to make sure that
you are comparing apples with apples.
Worse, most computer monitors are quite modest resolution. So it's necessary to magnify the image
at sseveral locations, including the center and several points on the periphery.
Also, LCD/LED monitors are low contast. A contrasty lens and one that reduces contrast by three
stops may look the same.
It's really quite difficult to test lenses. Correct focus is essential, AF may not be accurate enough,
so without optical focusing aids, you have to magnfiy the liveview and check. Again, it was
much simpler on an SLR (with ground glass, microprisms and a split ring).
The first step with any lens is to find its sharpest aperture. But on a zoom, that can be different for
different focal settings.
All camera lenss have some Petzval field curvature. So they will be sharper in the center than in the
periphery. Unfortunately, two bad aberrations -- coma and oblique astigmatism -- have the same
symptom. To rule them out, you need to try focusing on the periphery to see if it becomes sharp.
(All lenses will have some oblique astigmatism.)
Good lens test charts are expensive. If you want to be able to use a reasonably sized chart at a reasonably
short distance, then the printing has to be very sharp. For this reason, printing a DIY is difficult. I use the
Century Precision Optics test chart kit made by Schneider: one large chart and five small ones.
Basically, in includes resolution (similar but not identical to USAF 1951) and focus targets (Siemens star).
It's available from Hot Rod Cameras in Hollywood:
https://hotrodcameras.com/camera-department/century-precision-optics-lens-complete-test-chart-kit/Geometric distortion does not cause unsharpness. For some types of shots it doesn't matter, for others
(e.g, architecture) it matters very much. Using a grid target (or a piece of pegboard) will show distortion.
So it's not enough to "this lens is good", "this lens is bad"--good for what? A lens with a lot of spherical
aberration and Petzval field curvature is perfect for portraits--but only for portaits. But a lens with a lot of
chromatic aberration is only good as a paperweight.
Contrast and flare are a separate issue--but very important. Both highlights in the image and
outside the
angle of view can cause flare. Zoom lenses are extremely susceptable to flare (both the visible kind and
the insidious kind). A zoom may perform well in the studio but lose all it's contrast when you take it outside.
MTF graphs do not test with light sources outside the angle-of-view, so the low spacial frequency numbers
are not a good measure of the contrast you will get outdoors.
Far from being "easy to use", zooms are a source of a get deal of complexity and many pitfalls. Even chosing
a lens hood is a problem--for optimal protection at all focal settings, it really needs to be adjustable.
If you want image quality, far from being a "convenience lens" a zoom is a pain-in-the-ass.
I own half a dozen zooms (mostly picked up used), but frankly I can't remember which zoom is sharp at
which focal setting, or what the sharpest aperture is for each zoom and focal setting, or how much contrast
each zoom loses. Nor do I own the best lens hood for each one. So I just don't use them unless I have to
(i.e., I need to rapidly change the focal length, or or for someson I can't carry additional lenses). That
hardly ever happens.
The design of any zoom lens is a compromise. Why compromise if you don't have to?