Home
SearchSearch MemberlistMemberlist RegisterRegister ProfileProfile Log in to check your private messagesLog in to check your private messages Log inLog in

The flattening of modern lenses or the death of 3D pop
View previous topic :: View next topic  


PostPosted: Mon Oct 09, 2017 5:17 pm    Post subject: Reply with quote

I have been following K. Wheeler's YT channel for a while, and if I understand him correctly, it seems that the '3D Pop' is not equivalent to the DoF tricks, but what he calls 'microcontrast', that is the abillity to distinguish very subtle tonal gradation, important for b&w shots.

I am not knowledgeable enough to either dismiss or follow his views on the effect of many glass elements in the lens; he backs it up with light physics theories I don't quite understand. However, I heard from other sources that many glass-to-air surfaces cause contrast to deteriorate.

He also likes to emphasize the low number of glass elements as the main optical advantage of a prime lens over a zoom. This doesn't convince me on its own; as I see it, a single focal length lens is just easier to correct for aberrations (hence requiring less elements) than a multi-focal, which is destined to be 'jack of all trades' by design. Also, the more glass elements, the heavier the lens. This can be a significant factor, but it's not an optical property obviously.

Anyway, putting theories aside, his vintage lens recommendations are good.


PostPosted: Mon Oct 09, 2017 6:21 pm    Post subject: Reply with quote

Ray Parkhurst wrote:
All the images I've seen with "3D Pop" show "quick contrast reduction" as someone said above. What I've noticed about a couple of lenses I own is that they don't focus "linearly". They are lenses which are optimized for a particular magnification range (I am a macro photographer) and within a narrow range around that optimum magnification the lenses show a flatter focus vs magnification than outside the range.

Magnification translates to working distance, so that objects a certain working distance from such an optimized lens have this quality of stable focus. Lenses with "close range correction" like the 55/3.5 or 55/2.8 Micro Nikkors have moving elements to help to optimize the lens to improve performance at higher magnification. These moving elements improve performance at close range, but what is their effect on objects farther away? From what I've seen, they show good 3D Pop when used in their optimum range of around 0.5:1 magnification. Depth of field is getting shallow here, and the fall-off seems very quick, perhaps due to the CRC corrections.

Another lens with this quality is the Zeiss 74mm S-Planar. It is not a full-range macro lens, but one specifically optimized at 1:1. So its "CRC" is designed-in to the lens formula. It has the strongest nonlinearity of focus of any lens I own.

Applying this principle to a portrait lens, which is often where I see the most 3D Pop going on, if the lens is optimized for best performance at middle distances, and allowed to fall-off a bit especially toward infinity, wouldn't that create 3D Pop?

I'm also curious about the Nikon "defocus control" lenses. These seem to have a variable control to the focus fall-off. Do they show better 3D Pop? I have no experience with them but the principle seems to be consistent.


Im reminded of Telecentric properties. How these vary between lenses and between different focus distances and magnifications?


PostPosted: Tue Oct 10, 2017 1:13 pm    Post subject: Reply with quote

BarneyL wrote:
However, I heard from other sources that many glass-to-air surfaces cause contrast to deteriorate.


The only rule you need to know is that the maximum contrast of a lens system cannot be greater than or, or even equal to the sum of the contrast of all the individual elements... which is basically what the quote above really means. I believe some lenses replace the air-spaces with another medium, that may maintain higher MTF for a particular application (like underwater photography).

The resolution of the image (and also the contrast) is also limited by the pixel pitch/resolution of the digital sensor or the film (although, I'm not sure that the highest resolution and contrast technical emulsions were ever paired with a lens capable of outresolving them, except perhaps in the research lab). I suppose most enlarging/printing lenses of apo- design would outresolve all consumer photo films - and that's a good thing to understand here. Despite their resolution, they are limited by the source data, and also by the photopaper below. However the printing lens, by being designed with a higher theoretical MTF (resolution and contrast), preserve more of the original information/data/signal because the sampling frequency is higher. That is the concept of the Nyquist Frequency limit.

When the contrast of a pair of lines (detail) drops below 50%, we no longer perceive it as being particularly 'sharp' even though the detail is technically resolved. On a digital camera, it can be easily understood that not all photons/wavelets will hit precisely along the lines of the pixel sites on the image sensor.
Therefore one could hypothesise that there is a point where more lens elements and air spaces cannot push resolution and contrast - MTF - any higher. That doesn't mean that all other aberrations have been satisfactorily corrected though, so one may continue to add elements at the expense of maximum MTF. We've been told though, that essentially anything is possible so long as there is a budget for it. Therefore such a statement that lower lenses leave more contrast doesn't hold true in reality.

Eventually you have to set physical limits like weight, opto-mechanical alignment and stability, size and length, cost etc. The more you expand the system, the greater the % of error. It may not be that expensive even to design and manufacture a (to be exaggerating) 20 element, 50mm F0.7 that is the world's most-corrected lens at all focus distances, but can you assemble it anywhere but in a computer simulation to perform at its theoretical limits? No. At the end of the day you may as well say, "the limit of c is c" (the limit of the speed of light is the limit of the speed of light).

BarneyL wrote:

He also likes to emphasize the low number of glass elements as the main optical advantage of a prime lens over a zoom. This doesn't convince me on its own; as I see it, a single focal length lens is just easier to correct for aberrations (hence requiring less elements) than a multi-focal, which is destined to be 'jack of all trades' by design. Also, the more glass elements, the heavier the lens. This can be a significant factor, but it's not an optical property obviously.



To the contrary, I've read that many zoom lenses are technically better-corrected than their prime lens counterparts, simply because they must be in order to perform and be saleable. They cover such a wide range of focal lengths and focus distances for each focal length, each presenting its own set of challenges. If zoom lenses were *only* corrected as well as a prime lens - a single focal length -, then they would obviously be under-corrected and perform quite terribly. Of course, terrible prime lenses also exist. Wink

I forgot to mention that you cannot record at a resolution greater than the wavelength or frequency of the light/data, either. Hence, electron microscopes.