Z 24-70 2.8 vs Z 50mm 1.8

[Abitconfused]

I purchased a small color target (2014 DGK Color Tools) from B&H that I like quite a lot. I was tickled plum silly that the colors on the chart were identical to the colors on my eight year old ASUS monitor (98% sRGB). Then I ran some test shots with my Z 7II using the Z 24-70 2.8 and Z 50 1.8. I saw NO discernible difference in resolution at the 50mm mark.

I suspect both lenses are bumping their heads against the 47 megapixel sensor. But I don't know for sure.

[Manfred M]

I have no idea as to what you think you have accomplished. The colour reproduction of a modern lens, from the same manufacturer should be extremely consistent; this is something that they strive for. That is not one of the factors that normally plays a significant role in colour accuracy.

The light source used (spectral makeup and distribution) are the main issues here. Metamerism also plays a significant role, but that is more of a real world problem and not one we would expect to see in a test chart image.

The final comment is that the human visual system is quite suspect when it comes to reading colours / colour accuracy. This is one place where I put on my "techie" hat and rely on software tools to get a colour reading. Seeing no discernible differences is what one would expect, given the vagrancies of human vision.

[xpatUSA]

..... Then I ran some test shots with my Z 7II using the Z 24-70 2.8 and Z 50 1.8. I saw NO discernible difference in resolution at the 50mm mark.

Ed,

How did you measure the resolution? eyeball? software?



Here's Bart van der Wolf's image for testing lens resolution:

http://kronometric.org/phot/iq/res/JTF144_noCenter.png

Google 'Siemens Star' and 'slant edge testing' ...

I suspect both lenses are bumping their heads against the 47 megapixel sensor. But I don't know for sure.

My research using QuickMTF shows that high-MP sensors are quite soft at the pixel level compared to say older Sigmas ... for example, my 3.4MP SD9 beats my 20.3MP Panasonic G9 quite handsomely!

[DanK]

Ed,

How did you measure the resolution? eyeball? software?





My research using QuickMTF shows that high-MP sensors are quite soft at the pixel level compared to say older Sigmas ... for example, my 3.4MP SD9 beats my 20.3MP Panasonic G9 quite handsomely!

Interesting. But wouldn’t lp/mm or some such be more germane?


Sent from my iPhone using Tapatalk

[xpatUSA]

Interesting. But wouldn’t lp/mm or some such be more germane?


Sent from my iPhone using Tapatalk

Sorry, Dan, I don't fully understand. Germane in what way?

[xpatUSA]

I purchased a small color target (2014 DGK Color Tools) from B&H that I like quite a lot.

Ed,

A caveat: DGK Color Tools' stuff is OK for learning the basics but not so hot for serious work.

I bought one of their bigger targets for combined resolution and color testing: https://www.amazon.com/DGK-Color-Too...523036-5179626

I found that the target was printed cheaply by the half-tone (?) method and the dots showed up in my image ... grump ...

[DanK]

Sorry, Dan, I don't fully understand. Germane in what way?

In terms of visible differences in sharpness. To compare sharpness, one would want the same magnification. If you compare a sensor with a very large pixel pitch to one with a very small one at the pixel level, you are looking at very different levels of magnification.

I'd expect a high-resultion sensor to provide more, not less, signs of differences in resolution, measured in something lke lp/mm.

[xpatUSA]

In terms of visible differences in sharpness. To compare sharpness, one would want the same magnification. If you compare a sensor with a very large pixel pitch to one with a very small one at the pixel level, you are looking at very different levels of magnification.

The way I measure sharpness accounts for the magnification.

I'd expect a high-resolution sensor to provide more, not less, signs of differences in resolution, measured in something like lp/mm.

Yes, see MTF50 measured in c/mm which is lp/mm. Big pixels at right:

[pschlute]

I suspect both lenses are bumping their heads against the 47 megapixel sensor. But I don't know for sure.

How large are you printing that this makes a difference ?

[Abitconfused]

It occurs to me that the Nikon Z7II and Z lenses have taken the state of the art past evaluations derived from charts. After all, the 24-70 was made to be optically superior which can be interpreted in a variety of ways. Lens aberrations seem nonexistent. The Z system often shocks me not only in resolution and color but in contrast and a certain depth. Perhaps a photo to illustrate the idea...




Z7II & 24-70 2.8 @ 1/400s, f/7.1, ISO 100, 70mm

[Abitconfused]

Still trying to accomplish...



Z7II & 24-70 2.8, 1/80s, f/7.1, ISO 100, 24mm

[xpatUSA]

It occurs to me that the Nikon Z7II and Z lenses have taken the state of the art past evaluations derived from charts. .....

Ed, if that is your view, I have nothing further to offer.

[Abitconfused]

If everything could be measured then the universe would be a drum rather than a symphony.

[xpatUSA]

If everything could be measured then the universe would be a drum rather than a symphony.

So will you at least be answering my question:

Quote "Ed, How did you measure the resolution?" endquote.

[Abitconfused]

I examined the lines at the top and bottom of the chart being as objective as possible given circumstance that created as similar a perspective as possible. However, I think a "real world" test may be more revealing.

[Manfred M]

I examined the lines at the top and bottom of the chart being as objective as possible given circumstance that created as similar a perspective as possible. However, I think a "real world" test may be more revealing.

The tests are done in lab conditions when everything is set to be as close to optimal as can be. That is one reason I don't pay a lot of attention to said tests.

The two prints I had in a gallery (34" x 44" size) for a show were taken with my rather long in the tooth Nikon D810. I used the Nikkor 28 - 300mm lens that is known to have some strange distortion and is not being super sharp. I shot hand-held at middle focal lengths and apertures and shutter speed was not astoundingly high either. Everyone who saw them (including the pixel peepers) complimented me on how sharp the images were. I am in no rush to upgrade because I can still get some really technically great images. I suspect had I shot the Z7 with the current Nikkor 70-200m Z mount lens, people might have been able to tell the difference in a side by side comparison.

[xpatUSA]

The tests are done in lab conditions when everything is set to be as close to optimal as can be. That is one reason I don't pay a lot of attention to said tests.

The two prints I had in a gallery (34" x 44" size) for a show were taken with my rather long in the tooth Nikon D810. I used the Nikkor 28 - 300mm lens that is known to have some strange distortion and is not being super sharp. I shot hand-held at middle focal lengths and apertures and shutter speed was not astoundingly high either. Everyone who saw them (including the pixel peepers) complimented me on how sharp the images were. I am in no rush to upgrade because I can still get some really technically great images. I suspect had I shot the Z7 with the current Nikkor 70-200m Z mount lens, people might have been able to tell the difference in a side by side comparison.

I guess I sit somewhere between LensRentals and hand-held shots of shrubbery and rolling hills.

I shoot the slant-edge target that lives on my wall; then bring the image into QuickMTF which gives me numbers and an MTF curve within seconds. The curve can be saved as a PNG:

[Manfred M]

Okay Ted - here is the MTF for the lens I used....





Not pretty (in theory)...

[Abitconfused]

I see the truth of it.

[xpatUSA]

Okay Ted - here is the MTF for the lens I used....





Not pretty (in theory)...

Thanks Manfred. Here's MTFs for some older 50mm lenses that cover many apertures and spatial frequencies - not just two:



http://www.bobatkins.com/photography/technical/mtf/sqf1.html

See the link for what the x-axis 'S' means.

Looks like we might need 3D graphs to cover enough variables, eh?

I like Cicala's 3D plots in his blog but they do tend to hurt the brain ...