"Best" 35mm focal length lens, for Nikon F mount

[brbo]

The digital sensor used in the D2x is notoriously sensitive to the quality of lenses that one uses with it, and readily shows any flaws in their performance.

In what way?

The only way I can imagine this statement would make any sense is that you are talking about edge/corner performance or D2x's lack of good in-camera corrections that later and better cameras employed. Many pre-digital lenses (especially wide angle) perform poorly on digital. But that is more about limitations of digital sensors in general (which modern lens designs successfully worked around).

 

[George Mann]

In what way?

The only way I can imagine this statement would make any sense is that you are talking about edge/corner performance or D2x's lack of good in-camera corrections that later and better cameras employed. Many pre-digital lenses (especially wide angle) perform poorly on digital. But that is more about limitations of digital sensors in general (which modern lens designs successfully worked around).

You just answered your own question. Current digital sensors indirectly gather the waveforms of light, leading to lense interface complications that film doesn't suffer from.

But as I said, the sensor in the D2x is particularly sensitive to the quality of lense used with it do to its unique design.

Note that I am not referring to corner or edge performance specifically.

 

[brbo]

Thanks, although I must admit that after your clarification my understanding of what you are saying is even less than before

 

[Film-Niko]

Although the D2x is the highest resolving 12mpx camera yet created, my comment isn't solely about its resolution capabilities.

You are completely ignoring the fact that the D2x with only 12 MP is a lower resolution camera. With a resolving power much much lower than that of mediocre lenses.
And because of that it is not at all a good tool for lens evaluations.
Being a crop-camera which excludes all critical parts of the image circle (exactly those parts of the picture where good lenses especially demonstrate their superiority to lower quality lenses) further contribute to the ineptness of this camera for proper lens tests.

The digital sensor used in the D2x is notoriously sensitive to the quality of lenses that one uses with it, and readily shows any flaws in their performance.

As a former D2x user I cannot support that claim at all. The D2x has no such special capabilities. Just in the opposite with that cam I've seen the same result I've seen with all other digital SLRs in the 12-20 MP range: Because of their lower resolution they are levelling differences in lens performance.
Differences, which were clearly visible on films like Provia, Velvia, TMX, HR-50 e.g. But not visible, or to a much lesser extent visible, on these DSLRs.

 

[Film-Niko]

But do note that there isn't a single small format photograpic medium that outresolves the multicoated Nikkor 50mm f2.

That's a bold claim.
As you are saying outresolve: You can only say that if you have made resolution tests which allows you to quantify your results:
Therefore I am very interested in your data:
- which lenses have you tested in comparison to the 2/50 Nikkor?
- on which films you have tested the lenses?
- which object contrast has the test chart you have used?
- which lens has got what resolution values?
- at which distance the tests have been done?
- which focus operation (e.g. focus bracketing) was used?

 

[George Mann]

You are completely ignoring the fact that the D2x with only 12 MP is a lower resolution camera. With a resolving power much much lower than that of mediocre lenses.

The resolution capabilities of the D2x greatly exceeds any other 12mpx models, meets or exceeds the resolve of all 16 mpx models but the Pentax K5II, and even manages to outresolve nearly all 24mpx models that I have tested in low light/high ISO.

But as I stated, its peculiar lense matching property is not about resolution!

As a former D2x user I cannot support that claim at all. The D2x has no such special capabilities. Just in the opposite with that cam I've seen the same result I've seen with all other digital SLRs in the 12-20 MP range: Because of their lower resolution they are levelling differences in lens performance.
Differences.

Until you have performed the tests that I have, you don't have a leg to stand on here.

which were clearly visible on films like Provia, Velvia, TMX, HR-50 e.g. But not visible, or to a much lesser extent visible, on these DSLRs.

You are contradicting known facts to support your flawed position.

 

[George Mann]

That's a bold claim.
As you are saying outresolve: You can only say that if you have made resolution tests which allows you to quantify your results:
Therefore I am very interested in your data:

The lense in question managed to resolve to the factory stated limit's of both Tech Pan and CMS 20II.

The subjects were complex, real world scenes plus a USAF resolution chart photographed at noon using a superior quality Toshiba Y48 contrast filter, and viewed with a laboratory grade Zeiss microscope.

 

[dynachrome]

Some of this discussion falls into the area of artistic taste. On the technical side, if you carry out a calculation to too many decimal places the end result is less meaningful. It may be exciting to know that a particular document film can resolve X lines but not even the best general purpose lens will provide 250 lines/mm. That's just not realistic. A Printing Nikkor could get you closer but only if it is used in its optimum magnification range. Improvements in lens sharpness have been incremental. It has been observed that the 85/1.8 AF Nikkor is sharper than the 105/2.5 AIS Nikkor. I have and like both. If you can't get sharp images with a 105/2.5 AIS, it's not the lens. The fact that some lenses which were designed for film are not perfect in the corners with non-film cameras is not really important to me. When we are talking about very slow films (under ISO 100), these are already not suitable for every kind of picture taking. Can someone see great improvement in lens quality by looking at 5X7" prints? Probably not. Again, if I know I will need to make a much larger print, I would rather use medium format film. The question of whether a certain recent lens design is better also has to do with what the lens is being used for. I can look up the specs for a car that has 500HP. That doesn't mean I need such a car or have any good use for it. There is a misconception concerning slower film making a lens look bad. Years ago, Modern Photography tested a variety of black & white films. The sharpest images were made Kodak 5069 - High Contrast Copy Film. Technical Pan was a close second but still not quite as sharp. If used properly, slower film can make the images look even better.

 

[Film-Niko]

The resolution capabilities of the D2x greatly exceeds any other 12mpx models, meets or exceeds the resolve of all 16 mpx models but the Pentax K5II, and even manages to outresolve nearly all 24mpx models that I have tested in low light/high ISO.

Oh, now it is getting even more weird. Claiming that the D2x surpasses in resolution cameras like the 16 MP D7000, the 24 MP D7200, or the professional 20 MP D500.......Sorry, but as someone who has used and tested them, I have to completely disagree.
You cannot change physics, and you cannot obtain the same or even more resolution (as you have claimed) from a 12 MP compared to a 20 or 24 MP cam. You will always get higher resolution with a D7200 and D500 compared to the D2x. If you don't trust me, go to one of the numerous Nikon forums and ask there, why almost all have upgraded from the D2x to higher resolution models. When you tell them the D2x is superior in resolution, and low light performance, just be prepared to get a lot of shaking the head or laughter......

But as I stated, its peculiar lense matching property is not about resolution!

The D2x has an AA filter, and that is leveling lens differences, making them less visible.
Compare that with DSLRs which are available both with and without AA filter, then you will see the difference. For example the D800 vs. D800E.
Sensors without AA filter are a bit more "lens-sensitive", but not those with AA filter like the D2x.

Until you have performed the tests that I have, you don't have a leg to stand on here.

Considering your statements, I have certainly done much more tests, and I have done them in a much better, technically correct way.
My test procedures are mainly based on the procedures published by Image Engineering, the leading European camera and lens test institute, and what Zeiss had published.

You are contradicting known facts to support your flawed position.

No, I am not doing that at all. I invite you to come here, and then I show you in person the results of all my tests. E.g. my results of the mentioned films, with their higher resolution compared to the specific DSLRs I mentioned.
And I recommend to you to read about sensor technology and the physical resolution limit of the Nyquist frequency. Then you'll immediately realize the mistakes in your statements.

 

[Film-Niko]

The lense in question managed to resolve to the factory stated limit's of both Tech Pan and CMS 20II.

Sorry, but that is not possible at all because of the laws of physics:
CMS 20 II has a resolution limit of 800 lp/mm at high contrast details. That is measured without (!!) a lens, in a direct copy process.
You cannot get 800 lp/mm when a lens is involved in the imaging chain:
1/system resolution = 1/lens resolution + 1/film resolution.

To get 800 lp/mm with a lens you therefore need a lens with a resolution power above (!!) the diffraction limit. But that does not exist of course, because the diffraction limit is a physical limit which cannot be surpassed.

The highest resolution values so far on film have been obtained by Zeiss with their ZM 25mm Biogon at f4 with 400 lp/mm on Spur Orthopan film (which is identical to Adox CMS, first version).
And lots of others (me included) have reached the diffraction limit at f5.6 with their lenses, which is about 250 lp/mm (also on CMS 20 II).

 

[George Mann]

Oh, now it is getting even more weird. Claiming that the D2x surpasses in resolution cameras like the 16 MP D7000, the 24 MP D7200, or the professional 20 MP D500.

It clearly outresolves the D7000 across the board. The other 2 however do clearly outperform it in good light.

You cannot change physics, and you cannot obtain the same or even more resolution (as you have claimed) from a 12 MP compared to a 20 or 24 MP cam.

It is not dependant on mere physics alone, and is much more complicated than you assume.

You will always get higher resolution with a D7200 and D500 compared to the D2x.

In most instances, yes.

If you don't trust me, go to one of the numerous Nikon forums and ask there, why almost all have upgraded from the D2x to higher resolution models.

Completely unrelated cause and effect.

When you tell them the D2x is superior in resolution, and low light performance, just be prepared to get a lot of shaking the head or laughter.

What imbeciles think is completely irrelevant.

The D2x has an AA filter, and that is leveling lens differences, making them less visible.

This is a source of great contention since it performs like it doesn't have one at all, and I have yet to find proof that it does.

E.g. my results of the mentioned films, with their higher resolution compared to the specific DSLRs I mentioned.

I am well aware of the technical capabilities of film.

And I recommend to you to read about sensor technology and the physical resolution limit of the Nyquist frequency. Then you'll immediately realize the mistakes in your statements.

I am actually one of the worlds formost experts on the subject, and I am here to tell you that the Nyquist theorem is rubbish!

 

[George Mann]

Sorry, but that is not possible at all because of the laws of physics:
CMS 20 II has a resolution limit of 800 lp/mm at high contrast details. That is measured without (!!) a lens, in a direct copy process.

Again, you are qouting things out of context to my claims. I am well aware of these limitations, and don't need to be schooled in them.

 

[Film-Niko]

Some of this discussion falls into the area of artistic taste. On the technical side, if you carry out a calculation to too many decimal places the end result is less meaningful. It may be exciting to know that a particular document film can resolve X lines but not even the best general purpose lens will provide 250 lines/mm. That's just not realistic.

Well, it is realistic when you are using the right lenses and the right film. See my post above.
Zeiss had published their test results on Spur Orthopan film (Agfa HDP microfilm, identical to Adox CMS 20, first version) with their ZM Biogon at f4 of 400 lp/mm. So they reached the diffraction limit at f4.
Here on photrio we've had several reports that members reached the diffraction limit at f5.6 (about 250 lp/mm) with high-quality lenses (me included).

Improvements in lens sharpness have been incremental.

No, they have been huge at open aperture, very big at one stop stopped down, and clearly visible more stopped down with the modern, highest quality lens designs.

It has been observed that the 85/1.8 AF Nikkor is sharper than the 105/2.5 AIS Nikkor. I have and like both. If you can't get sharp images with a 105/2.5 AIS, it's not the lens.

That is not the point. With a Nikkor 2/105 DC, 1.4/85 AF-S or a Zeiss Milvus 1.4/85, you get a significantly better performance in the range of f1.4/f2 to f5.6 compared to the old 2.5/105. That gives you more creative freedom, and the possibility to use slower, higher quality films in many situations.

When we are talking about very slow films (under ISO 100), these are already not suitable for every kind of picture taking.

The point is: With these improved modern lenses it is possible to use slower films with their better quality much more often, because it is not needed anymore to stop down the lens by 2-4 stops to get very good optical quality.

Can someone see great improvement in lens quality by looking at 5X7" prints? Probably not.

Defintely yes. As already explained, the difference in lens performance with the improved current lenses is clearly seen, especially at open aperture and one stop stopped down.

Again, if I know I will need to make a much larger print, I would rather use medium format film.

And with medium format film you often have to use higher speed film, which is significantly reducing the quality.
With the current improved f1.4 lenses, combined with films like Velvia, Provia, TMX, Panf+, Delta 100, Acros II, HR-50 you get an outstanding quality also with big(ger) enlargements. A quality with which you won't miss medium format very much, also having all the advantages of 35mm.

If used properly, slower film can make the images look even better.

Yes, of course. And with these modern, much improved lens designs you can use slower films much more often, in more situations.

 

[Film-Niko]

I am actually one of the worlds formost experts on the subject, and I am here to tell you that the Nyquist theorem is rubbish!

Just two questions:
1. If it is rubbish, why never ever results of resolution above the Nyquist frequency have been published?
To record one line-pair, you need at least two pixel: One for the black, the other for the white line.
It simply does not work with one pixel, or with 0.5 pixel.

2. The Nyquist frequency is approved by all experts in the scientific literature.
Please give me a link to your publications, in which you prove all others wrong.

 

[Film-Niko]

Again, you are qouting things out of context to my claims. I am well aware of these limitations, and don't need to be schooled in them.

No, I don't.
You made made these claims. And you have not given any proof at all to substantiate your claims.

 

[George Mann]

No, I don't.
You made made these claims. And you have not given any proof at all to substantiate your claims.

Nor do I intend to as these types of debates always prove to be a waste of time.

I have stated things as I see it, and you have as well so lets just leave it at that.

 

[George Mann]

2. The Nyquist frequency is approved by all experts in the scientific literature.
Please give me a link to your publications, in which you prove all others wrong.

Every real world test ever conducted will show you that they fall short of the Nyquist theorem.

 

[Angarian]

You are completely ignoring the fact that the D2x with only 12 MP is a lower resolution camera. With a resolving power much much lower than that of mediocre lenses.
And because of that it is not at all a good tool for lens evaluations.
Being a crop-camera which excludes all critical parts of the image circle (exactly those parts of the picture where good lenses especially demonstrate their superiority to lower quality lenses) further contribute to the ineptness of this camera for proper lens tests.



As a former D2x user I cannot support that claim at all. The D2x has no such special capabilities. Just in the opposite with that cam I've seen the same result I've seen with all other digital SLRs in the 12-20 MP range: Because of their lower resolution they are levelling differences in lens performance.
Differences, which were clearly visible on films like Provia, Velvia, TMX, HR-50 e.g. But not visible, or to a much lesser extent visible, on these DSLRs.

Yepp. From my own experiences and tests over all the years with the mentioned equipment I can completely agree. And as you have correctly added in your later post the AA filter also plays a significant role in levelling lens charcteristics. The Nikon D800E vs. D800 test - without AA filter vs. AA filter - demonstrates that indeed very obviously. I've done it together with a friend after the market introduction of both cameras.

And so far I have also got higher resolution values (at medium object contrast - important, as resolution is dependent on contrast) with T-Max 100, Delta 100, Acros, Velvia, Provia, E100 compared to D850.
CMS 20 II is of course even in another league!!

 

[Angarian]

Well, it is realistic when you are using the right lenses and the right film. See my post above.
Zeiss had published their test results on Spur Orthopan film (Agfa HDP microfilm, identical to Adox CMS 20, first version) with their ZM Biogon at f4 of 400 lp/mm. So they reached the diffraction limit at f4.
Here on photrio we've had several reports that members reached the diffraction limit at f5.6 (about 250 lp/mm) with high-quality lenses (me included).

You can put me in this club as well: I have also reached the diffraction limit of white light at f5.6 with my best lenses on Adox CMS 20 II. This film is absolutely incredible!

And I remember the Zeiss test as well: Was published at that time in their "Camera Lens News" publication. Breathtaking results!!

No, they have been huge at open aperture, very big at one stop stopped down, and clearly visible more stopped down with the modern, highest quality lens designs.

+1.

The point is: With these improved modern lenses it is possible to use slower films with their better quality much more often, because it is not needed anymore to stop down the lens by 2-4 stops to get very good optical quality.

After using more of the current premium lenses in the last years, I have also started using slower films more often. An additional very big performance boost when you are using Delta 100 instead of Delta 400 for example, or Fujicolor 100 instead of X-Tra 400.

Defintely yes. As already explained, the difference in lens performance with the improved current lenses is clearly seen, especially at open aperture and one stop stopped down.

Same results here as well.

 

[Angarian]

Just two questions:
1. If it is rubbish, why never ever results of resolution above the Nyquist frequency have been published?
To record one line-pair, you need at least two pixel: One for the black, the other for the white line.
It simply does not work with one pixel, or with 0.5 pixel.

2. The Nyquist frequency is approved by all experts in the scientific literature.

Correct.
The Nyquist frequency is a physical resolution upper limit of digital sensors, which cannot be surpassed at all. Physically impossible.
You will always get resolution values below that limit.
I have tested the resolution of all my digital cameras over the years, and also of some cameras of friends.
The real resolution has been about 10% below the NF with sensors without AA filter, and 15-20% below the NF with sensors with AA filter.

With film it is of course different, as there is no NF. Therefore often higher resolution values are possible with a certain object contrast and high(er) resolving low- and medium-speed films. And the better the lens, the higher the resolution.

 

[koraks]

these types of debates

From a moderator's perspective, and not specifically addressing @George Mann here, but merely taking his remark as a point to latch onto: it also appears to me that this is the phase on a debate where the discussants will have to "agree to disagree". Of course, a respectful and constructive further exchange of insights is welcomed, as always. But please, keep it civil, respectful and overall nice. Thanks all!

 

[BillBaileyImages]

Correct.
The Nyquist frequency is a physical resolution upper limit of digital sensors, which cannot be surpassed at all. Physically impossible.
You will always get resolution values below that limit.
I have tested the resolution of all my digital cameras over the years, and also of some cameras of friends.
The real resolution has been about 10% below the NF with sensors without AA filter, and 15-20% below the NF with sensors with AA filter.

With film it is of course different, as there is no NF. Therefore often higher resolution values are possible with a certain object contrast and high(er) resolving low- and medium-speed films. And the better the lens, the higher the resolution.

Several years ago I was part of a confidential consulting engagement to see how many line pairs per millimeter could be resolved. With the confidential, legal, restrictions, I am not able to share any more than cursory details. With my Nikkor 600mm lens on my F5, controlled lighting, special film shot at ISO 12 and processed with carefully-controlled chemistry, we were able to resolve 1,500 lp/mm--exactly what the client wanted to know. It would be of interest to see folks' outcomes with even higher resolution. On to >= 6,000 lp/mm?

 

[Angarian]

Several years ago I was part of a confidential consulting engagement to see how many line pairs per millimeter could be resolved. With the confidential, legal, restrictions, I am not able to share any more than cursory details. With my Nikkor 600mm lens on my F5, controlled lighting, special film shot at ISO 12 and processed with carefully-controlled chemistry, we were able to resolve 1,500 lp/mm--exactly what the client wanted to know. It would be of interest to see folks' outcomes with even higher resolution. On to >= 6,000 lp/mm?

Just to avoid any misunderstandings, as in Europe there is a differnce to US in using a point or comma within numbers:
Are you saying you have got one-thousand-five-hundred linepairs per millimeter on that film?

 

[BillBaileyImages]

That is correct. We were determining if a specific invention might be feasible for manufacturing. Thanks for the reminder about European and United States usage of commas and periods. This is indeed an international forum.

 

[benveniste]

My favorite prime 35mm is the 35mm f/1.4. However, I recently sold it and use (my only zoom lens for Nikon or Hasselblad V-series) the Nikkor 17-35 (NOT the 16-35!) It is the proverbial "brick"--solid, dependable, sharp, and almost bullet-proof!

For values of "dependable" which don't include the autofocus motor. I've had to replace mine twice.

I'll pass on the "my lens is sharper" exchange, thanks. The only 35mm prime I own is an early 35mm f/2.8 PC. Instead, I'll point y'all to this comparison:

Variance Measurement for 35mm SLR Lenses

Last week we posted optical bench MTF and copy-to-copy variation measurements for 50mm SLR lenses. We're going to continue that series this week with the same measurements for the 35mm lenses. This set should fill out the wide-to-standard range lenses and gives us a lot of interesting data as we...

www.lensrentals.com

I'll also point out that the Sigma f/1.4 is more than twice as heavy as the Nikon f/1.8G.