Low Element Count Lenses, 3D pop, and Micro-contrast -- My Theory

charjohncarter said:
In passing I've thought about this subject but in a very personal way: which of my lenses have 'pop.' The lenses that I have with 'pop' are: my Thorium Pentax Super Takumar f1.4 50mm, My Pentax 105mm 6x7 Thorium f2.4, and surprisingly my lens on my Balda Hapo 66e which I think is a three element. I have a few others but those stand out.

I think you are on to something here, this is my 6x7 lens which always looks better than anything else I have (3D):
Click to expand...

I strongly agree - the Pentax 105mm f2.4 is one of the most remarkable lenses I've used in terms or 'look' or 'special properties' or 'magic dust' or whatever you want to call it.
 
helenhill said:
Yum to 4 elements...ALL Elmars !
be it f3.5 or f2.8
they are so damn good, workhorses, Beautufully sharp, lovely oof
Click to expand...

This is to illustrate I agree with Helen. From my Elmar-M 50/2.8. I have three Elmar 50's now, so I must like them.
U51008I1529304130.SEQ.0.jpg

Leica M6, Elmar-M 50/2.8, Delta 100
John Mc
 
retinax said:
Fully agree with your points 2 and 3. Not so sure about 1. Why does less light in absolute terms have to equal less information? The information is "encoded" in the relative amounts of light. It doesn't decrease when you stop down or when a cloud obscures the sun. There is something to it though if one thinks about what happens to the light that doesn't get transmitted–it ends up bouncing around in the lens, creating flare. Flare lowers contrast in the shadows, so information is indeed lost. Now coatings are very, very good nowadays and I'm pretty certain that the high element count lenses still transmit more light than uncoated simple designs.

About the lens with T=f, that's just not possible. Can you see that there's glass in there? See, some light is reflected off of the lens surfaces, ergo not transmitted. Either it actually has a larger f-stop than indicated, in order to reach T 1.4, or there's another error in the test.
Click to expand...

I will concede that I'm a bit out of my element (see what I did there?) on this and not married to the notion. The simple answer is "dunno". However, it makes more sense perhaps if you don't think of "light as light", rather "light as a medium for information". "10" units of information is being captured on the front end of a lens. Only "8" units of information (which is "carried" by the "medium" of photons) is being passed out the other end reaching the sensor. The balance is being absorbed (for lack of a better word). However, it's not merely "light" being absorbed, it's the information that the light carries. Since modern lenses have high resolution specs, and all/most of the image detail information is making its way to the sensor, detail information is not affected by glass elements -- it passes all those elements easily. Therefore, what I propose is that it's "microcontrast" or that information being carried by the light which conveys perceptual dimensionality.

You are absolutely welcome to disagree. I tried to be careful in my word choice to hopefully say this is a "possibility" or "seems to maybe make sense" etc and not proclaim it as absolute fact. But because something may make sense up front doesn't make it true -- phlogiston theory made sense, turned out to be completely false. However, it perhaps merits consideration if you think of light as a medium that carries information.

If a lens has a 1.4 aperture value, and is collecting "1.4" in the front element but only 1.7 or 82% of that light makes its way to the sensor/film plane, the INFORMATION being carried by the light is being lost (I posit) along the way. If it is not "detal information" (which doesn't seem to be affected by element count) then what information is being lost? What remains? The information that conveys (or helps convey) a sense of dimensionality, "possibly", "perhaps".
 
Noserider said:
Yes a lovely bit of conversation where we all can learn something and have some fun.
That said to the OP
Well buddy if you certainly enjoy your AF-D Nikkor 50mm f/1.4 knock yourself out; however its widely known to be the least favorite of the Nikkor 50's and I've had them all (including several AF-(D) versions.
I would heartily endorse doing some reading of Erwin Puts his optical explanations you might find educational on how (mostly Leica) lens are made and the balancing act of optical design.

That being said; I very much do think that there *can be* a noticeable difference in my own preference of certain images made with lenses with *less* glass and truth be told I've been going thru an Elmar phase myself much like Ms. Hill.

From what I've gathered slower designs with fewer elements many be worth examining if this interests anyone. Zeiss Tessar 3.5 vs Planar 2.8 is an example.
Click to expand...

I do not doubt that Leica makes fabulous glass. However, I no longer shoot small format film (or APS-C or smaller digital), I shoot FF digital. If Leica made an "affordable to me" full frame digital camera body with up-to-date sensor specs that matched Sony and Nikon (instead of Canon), to attach that Leica glass to... I might very well shoot Leica. I shoot Nikon not because I'm a fanboy but rather by process of elimination. Nikon and Sony have the best senors in their cameras -- that much is verifiable fact. Others are lagging behind them and not by a little, especially Canon and Leica. Since I'm not a fan of Sony cameras (or mirrorless cameras in general), I shoot Nikon. Now to find a lens to slap on that camera body, so I'm confined to the Nikon universe, and nothing wrong with that given the backwards compatibility of the F-mount. For a 50? I insist on the 1.4 spec in A 50mm prime and do prefer autofocus --with a slight preference for screw drive for durability and smaller lens size. That rather narrows my choices.

As far as the 50/1.4 AF-D, I won't bore you with DXOmark scores. It still holds its own, does quite well, in fact, in all these metrics. It's a basic 50/1.4 with the same 7 element, 6 group classic design as its predecessors and used by the usual suspect lens makers with Nikon coating and color rendition, which I like. It's small, lightweight, affordable, focuses fast enough for me, and is a "true" f1.4, one of the few (perhaps only) 50/1.4's made by anyone that passes all the light/information gathered in the front element on to the sensor.

But ultimately? I like the way it renders images. I believe this is still a current production lens. There's a reason for that. As far as Mr Putts? Very knowledgegable. But is as biased as he is knowledgeable.
 
NickTrop said:
I will concede that I'm a bit out of my element (see what I did there?) on this and not married to the notion. The simple answer is "dunno". However, it makes more sense perhaps if you don't think of "light as light", rather "light as a medium for information". "10" units of information is being captured on the front end of a lens. Only "8" units of information (which is "carried" by the "medium" of photons) is being passed out the other end reaching the sensor. The balance is being absorbed (for lack of a better word). However, it's not merely "light" being absorbed, it's the information that the light carries. Since modern lenses have high resolution specs, and all/most of the image detail information is making its way to the sensor, detail information is not affected by glass elements -- it passes all those elements easily. Therefore, what I propose is that it's "microcontrast" or that information being carried by the light which conveys perceptual dimensionality.

You are absolutely welcome to disagree. I tried to be careful in my word choice to hopefully say this is a "possibility" or "seems to maybe make sense" etc and not proclaim it as absolute fact. But because something may make sense up front doesn't make it true -- phlogiston theory made sense, turned out to be completely false. However, it perhaps merits consideration if you think of light as a medium that carries information.

If a lens has a 1.4 aperture value, and is collecting "1.4" in the front element but only 1.7 or 82% of that light makes its way to the sensor/film plane, the INFORMATION being carried by the light is being lost (I posit) along the way. If it is not "detal information" (which doesn't seem to be affected by element count) then what information is being lost? What remains? The information that conveys (or helps convey) a sense of dimensionality, "possibly", "perhaps".
Click to expand...


michalewj has further up in this thread explained what kind of information light can carry. Now that get this all out of the realm of the esoteric, you should find out which part of the information might convey what you call sense of dimensionality. After you've established what that is so that other people can follow. The article you've linked to also fails to do that, so I think that's the real obstacle. I haven't seen any examples that can't be explained with other factors like lighting, composition, dof.
This discussion could be more productive if someone made some tests with different lenses side by side and posted the results. ATM we even disagree if the thing we're discussing here exists or not.
As Michael mentioned and I hinted at, the information that can get lost due to transmission being not perfect, as far as conventional wisdom goes, is some contrast and thus tonal separation and shadow detail which gets covered up by flare and can't be recovered as far as it falls under the noise floor. Everything beyond that would be new territory AFAIK.
 
shawn said:
...A big part of 3d pop will depend on lighting, subject distance vs background distance, focal length and aperture. Still, different lenses will react somewhat differently if all those variables are kept the same.

Shawn
Click to expand...

This seems to be the case pretty often. There's that razor-edge zone thing of hitting just the right combination of light, subject and lens settings that works so well.
 
gavinlg said:
I have to agree with the original post, the angry photographer, and Yannick Khong - even if I'm not entirely sure of the theory or technical explanation behind it. Not discounting any theory of course - the idea that more glass elements dulling the quality of light transmission sounds theoretically possible to me (not that I'm scientifically proficient in any way), and the theory that ironing out aberrations to such a high degree that the lens performance 'flattens' also seems very plausible to me.

What I do know is that through the years and my experience with lenses, I've realised that there is an era of lenses that tends to give me photographs that I strongly prefer over the more modern and highly corrected bunch of lenses. In general, these lenses tend to:

- Have a lower element count
- Use less or no aspheric or plastic elements (with exceptions - ie 35mm sumicron asph)
- Have original formulations based around the 70s

Another really obvious thing to me is that these differences are more obvious on film cameras rather than digital cameras. Another point is that I really tend to dislike the rendering properties of lenses like the sigma art series or Olympus Zuiko digital lenses that are extremely highly corrected and high element count. Not saying they make bad images, or that good images can't be made with these lenses, just that I strongly prefer how older, more simple lenses translate light, tone and color.
Click to expand...

I regard those lenses (Sigma Art, Oly Digital) as "fixed focal length zooms" (rather snarkily). The Sigmas are as big/heavy as zooms and have as many elements. And from what "some" have said (can't speak from experience) they render images with high detail by a rather "zoom-like" quality.

I agree I like the look of older glass to modern counterparts. I don't entirely agree that they look better on film than full frame.
 
Interesting discussion, although referencing Ken Wheeler of all people? really? (this from personal experience.)

Anyway...why would this theory be limited to primes? The Canon 70-200/2.8 IS L has plenty of '3D pop' and it has 23 (!) elements in 18 groups...

There are plenty of modern lenses that have both 3D and lots of elements...
 
I needed to clarify something here. You can't just compare apertures by dividing them to figure out light transmission. Apertures are actually based on area, which is proportional to the square of the diameter. From Wikipedia, the definition of an f/stop:

The f-number of an optical system (such as a camera lens) is the ratio of the system's focal length to the diameter of the entrance pupil.
Transmission = (aperture #1 squared / aperture #2 squared).

If you go through the math, f/1.7 has 41% less light transmission than f/1.4. f/1.7 is actually a 1/2 stop down from f/1.4.

1/3 of a stop is 26% difference in light transmission.
1/2 stop is 41% difference
2/3 stop is 59% difference.

Light transmission varies exponentially. The f/stop scale makes the exponential linear, as it's logarithmic (based on log 2).

To the poster who talks about pinhole photography:
Diffraction will rob you of the image information.

A big reason why modern lenses need so many elements is that digital sensors respond best when the light comes at them directly from a perpendicular angle. A large number of those elements are concerned with reducing the incidence angle of the light beams to the corners of the image. Film doesn't care about this. Thus, we film users have the glorious Zeiss Contax 35mm Biogon, with the big bulging lens element in the rear, which is only a few millimeters from the film surface. This lens, and many of the early CV LTM lenses meant for film, perform poorly on digital for that reason.
 
For one reason or other we all have our favorite lens; in my case it is the final product: the print. The 'pop' factor is something I like, and I care less about crazy edge to edge sharpness. I even like this lens: a Brownie Hawkeye Flash lens mounted on my 6x7:


Ektar 100 by John Carter, on Flickr

Not much 'pop' but some.
 
I'd love to see a side by side comparison between two 50mm 1.4 lenses, one with the celebrated 3D pop and one without, same camera to subject distance, f-stop, etc. I'd really like to see some evidence that 3D pop is a phenomenon separate from shallow depth of field and thus can't be predicted from the standard depth-of-field calculations.

I also wonder how many 3D pop seekers are rangefinder-only users. Depth of field is quite apparent using an SLR or mirrorless camera.
 
michaelwj said:
Just quickly, there is lot of misinformation and "tin foil hat" science going on here.

Your point 1;
Light carries information about intensity (number of photons), colour (wavelength of the photon), and polarisation (of each photon) and that is it. More light has more intensity (more photons), not more types of information.
Click to expand...

Couple things. Firstly, where did I say "types of information"? I effectively agreed with your characterization of "type of information" and we also agreed that light is a medium. What I am saying as that the informaiton being carried by the light is possibly being lost in high element count lenses where the aperture value is greater by than the t-stop value (such as the Sigma Art), and the additive effect of the loss of "this information" (not "new type of information") COULD BE causing a loss of perceived dimensionality in the image. Secondly, I never said that what I floated was "fact", I suggested it as a "possibility" and was careful with my word choice and reiterated this as "possiblity" on several subsequent posts. Therefore, no "tin foil hat" as nothing was stated as irrefutable fact. "UFOs HAVE landed" -- tin foul hat. "There is ample evidence, seems to me, to leave open the possiblity of UFO visitation" -- no tin foil hat. And thanks for the unnecessary line about critical thinking skills. Not necessary, mind you, but the it's always nice to reinforce them.

And, really, what fun would the internet and blogs like this be without its malaperts?
 
NickTrop said:
Couple things. Firstly, where did I say "types of information"? I effectively agreed with your characterization of "type of information" and we also agreed that light is a medium. What I am saying as that the informaiton being carried by the light is possibly being lost in high element count lenses where the aperture value is greater by than the t-stop value (such as the Sigma Art), and the additive effect of the loss of "this information" (not "new type of information") COULD BE causing a loss of perceived dimensionality in the image. Secondly, I never said that what I floated was "fact", I suggested it as a "possibility" and was careful with my word choice and reiterated this as "possiblity" on several subsequent posts. Therefore, no "tin foil hat" as nothing was stated as irrefutable fact. "UFOs HAVE landed" -- tin foul hat. "There is ample evidence, seems to me, to leave open the possiblity of UFO visitation" -- no tin foil hat. And thanks for the unnecessary line about critical thinking skills. Not necessary, mind you, but the it's always nice to reinforce them.

And, really, what fun would the internet and blogs like this be without its malaperts?
Click to expand...

Do you have examples of the Nikon 50mm that you have compared to the Sigma art lens which you claim has less 3D pop? You seem to be assuming that transmission properties of the lens are related to the out-of-focus areas of the image. I'm confident if you put a 10 stop ND filter on your Nikon, it would continue to exhibit the same depth-of-field properties, even though the light transmission has fallen to a small percentage of what it was before.

Are there any images already posted in this thread which you would consider to have 3D pop?
 
mich rassena said:
I'd love to see a side by side comparison between two 50mm 1.4 lenses, one with the celebrated 3D pop and one without, same camera to subject distance, f-stop, etc. I'd really like to see some evidence that 3D pop is a phenomenon separate from shallow depth of field and thus can't be predicted from the standard depth-of-field calculations.

I also wonder how many 3D pop seekers are rangefinder-only users. Depth of field is quite apparent using an SLR or mirrorless camera.
Click to expand...

Shallow depth of field can certainly play a part on this. Lighting is a huge part too. Take two outdoor shots with the same lens. For the second shot use a fill flash to boost your subject 1/2 or a stop above the rest of the scene. There will be way more 3d pop in the second picture, even though the lens (DOF, focus point, perspective and so on) is the same.

Shawn
 
I echo the comment made by Mich. Does anyone have two pictures taken under identical conditions except one was taken with a lens having 'pop' and the other with a lens not having 'pop'? It would be nice to see visually what is being talked about.
 
mich rassena said:
I'm confident if you put a 10 stop ND filter on your Nikon, it would continue to exhibit the same depth-of-field properties, even though the light transmission has fallen to a small percentage of what it was before.
Click to expand...

You are not getting what I'm saying. Here is a cut-n-past direct from DXOmark lens comparison table:

Light transmission
The photometric aperture, also known as “T-stop” (T = transmission), is the aperture of the lens corrected for its transmission loss (DXOmark's [?] definition.

Sigma 50mm f1.4 Art
Transmission [?] 1.7 TStop

Nikon 50mm f1.4 AF-D
Transmission [?] 1.4 TStop

This is how max aperture is calculated. Focal length/max aperture diameter.
__________

So -- divide the aperture value of both lenses by their t-stop value.
Nikon AF-D 1.4/1.4 = 1. Convert to percentage = 100%. (FWIW this is the only lens I've seen/am aware of with a 100% Max AP/T value...) Let's call this "ITV" for "information transmission value" that calculates the amount of information that makes its way from the front element to the sensor or film plane.

Now let's do 1 - ITV and call that "ILV" for information loss value. In the case of the old Nikkor, 1-1 = 0. 0 information is lost. You want the ITV to be as close to 100% as possible and the ILV to be as close to 0 as possible. In this regard, the Nikon is "perfect. 100% ILV, 0% ILV. ALL of the information gathered by the front element makes its way to the sensor whether I have an ND filter on the lens or not. Doesn't matter.

Not the Sigma Art:
1.4/1.7 = 82% ITV.
1-.82 = 18% ILV.

Now -- let's get your silliness regarding ND filters out of the way. It doesn't matter if I use an ND filter. If I slap on a ND filter and I lose two stops that two stop loss is consistent in the Nikon 50/1.4 AF-D. I can compensate by adjusting aperture or shutter speed. However, the ITV for the Nikon remains 100% with zero loss of information and a 0% ILV.

And if I put a 2X ND on the Sigma? Its ITV value will still be 82% and its ILV will be 18%.

These values remain constant no matter what kind of ND value you put on the FRONT of the lens (which would be the same as shooting in a 2 stop darker environment). In the case of the Sigma, those 13 elements are absobing 18% of the information. This lens WILL only and CAN only deliver 82% of information it collects in the front element, out its rear element to the sensor. And that is true regardless if you put an ND filter in front of the lens. 18% of that information is lost passing through the internal elements.

If detail information isn't lost -- this lens is off the charts in terms of sharpness, then what information -- 18% of it, IS being lost?

As far as samples, here is a link:

2016 DEPTH VS. FLAT LENS COMPARISONS
https://yannickkhong.com/blog/2015/11/12/depth-vs-flat-lens-quick-comparison
 
By the way -- if you Google "What lenses have the best microcontrast", Google delivers the following list:

Best Lenses for Micro contrast
Micro-Contrast, the biggest optical luxury of the world
Nikkor AF 50mm 1.4D.
Nikkor AF 35mm 2D.
Voigtlander SLII 58mm 1.4 Nokton.
Zeiss ZF2 35mm 2.0 Distagon.
Nikkor Ai-S 50mm 1.4.
Nikkor Series E 135mm 2.8.

I happen to own the first two lenses on this list, just by accident. And IMO they do deliver 3D pop. I "bet" -- (and I already did for the two I own -- the 50/1.4D and the 35/2D 100% and 93%, respectively) if you did my calculation for these lenses in the post above, you may possibly find they have high ITV values and low ILV values.
 
Also, the 50/1.4D would be a terrific cine lens if not for the click stops. This is because cine lenses "go by" t-stop not f-stop. However, because you only know the t-stop values of a cine lens and not the f-stop, you can't calculate its ITV and ILV. In the case of the Nikon fstop = tstop or, it has a 100% ITV.
 
You must log in or register to reply here.
Back
Top Bottom