Micro-contrast?

Discussion in 'Native Lenses' started by oldracer, Jul 8, 2016.

  1. oldracer

    oldracer Mu-43 All-Pro

    Oct 1, 2010
    USA
    I occasionally see this term and I find it baffling. To me, contrast is a ratio of two light intensity measurements and I see no reason that contrast would have anything to do with size. (Whatever it is that "micro" means in this context.)

    Is this just an OWT or can someone give me a physical explanation of why there would be something called "micro-contrast" that is any different than plain old contrast?

    Please spare me the "I can't explain but I know it when I see it." type responses. I don't believe in magical stereo speaker wires either. There is no magic. Only physics.
     
  2. hazwing

    hazwing Mu-43 All-Pro

    Nov 25, 2012
    Australia
    *tongue in cheek*
    It's the magical property that makes photos look amazing, and justifies the price of spending big $$$ for expensive lenses.
     
    • Agree Agree x 1
    • Winner Winner x 1
  3. PakkyT

    PakkyT Mu-43 Top Veteran

    767
    Jun 20, 2015
    New England
    I believe it is a bullcrap term that means "contrast". But if you put a "micro" in front of it, then it sounds like your contrast's crap don't stink like us common folk's "macro contrast".

    Edited to add: sounds like a "I'm a true photographer" way of describing resolving power of a lens???
     
  4. Replytoken

    Replytoken Mu-43 Hall of Famer

    May 7, 2012
    Puget Sound
    Ken
    • Like Like x 2
    • Informative Informative x 1
  5. oldracer

    oldracer Mu-43 All-Pro

    Oct 1, 2010
    USA
    Thanks. That helps to explain the term, for sure. I would still be interested in the physics, though, since if a lens can produce the full range of pixel values from pure white to dead black from a suitable sensor it seems like it also has to also be able to produce all of the values in between. I don't think the mapping curve would have sharp nonlinearities in it.

    It's easy see that a sensor readout circuit might not be able to produce all the values (think 4-bit A/D conversion) but I don't see what characteristics of a lens could produce the problem.
     
  6. magicaxeman

    magicaxeman Mu-43 Regular

    76
    Feb 27, 2016
    Essex UK
    Ian
    As described in the above LL link, Micro-contrast is not a BS term but refers to the lens having good enough contrast that it can differentiate between small area's with similar contrast, this gives better detail and hence the term micro, rather than just contrast.

    The difference between a lens with good micro-contrast and one with average performance is quite perceptible & tangible.
    Even more so if your a B&W shooter.

    I used to shoot Fuji all the time, now Fuji are well known for their sharp and contrasty lenses, especially the primes and the premium zooms.
    Yet the difference when using for instance the Zeiss 50/2 Planar T ZM, a lens renowned for its sharpness and micro-contrast is quite marked.

    You wil also see sometimes a reference to how a lens "draws" this refers to the way the lens renders details within the image, but you don't have to worry about the technicalities of how a lens produces a certain image, just whether or not you like the images the lens produces.
    A good way to determine this is to do a thorough search for images using your camera with the prospective lens, looking especially for straight out of camera images with no editing apart from resizing for the web, that way you can see if you like what the combination produces and then decide whether or not to try the lens yourself.
     
    • Like Like x 2
  7. MJL

    MJL Mu-43 Regular

    199
    Feb 24, 2016
    Katikati, New Zealand
    Marinus
    I have not yet read a clear explanation of this but imho it has to do with understanding how light behaves. First understand a MTF graph. ( What is a lens MTF chart and how do I read it | NIKKOR lens technology from Nikon )
    If the sagital and meridional lines are on top of each other then there is expectation that the bokeh has the potential to be good. MTF HS043E.

    Unfortunately this MTF chart does imho not tell the whole story - light has a different wavelength at the red end of the spectrum than from the blue end which means that on a pure spherical ground lens the red and blue light does not get to a pinpoint at the same spot.
    In the above MTF chart of the Panasonic 42.5mm F1.7 we see that the lines are nicely following each other up to approx 8mm from the centre and then they diverge. This means that light does not arrive at a pinpoint outside the 8mm circle and (as far as my understanding goes) leads to color fringing and the bokeh no longer being a perfect circle at the edge. It is also suggestive that the "bokeh" circle does not have a nice gaussian circle that fades away from the centre outwards, in other words the centre should be the brightest and then the further it gets away from the centre the fainter the lightspot gets without a sharp defined edge.

    Designers have come up with lenses that have been ground in a special shape to overcome this but a number of compromises have to be made in order to obtain a happy medium for both the long and short wavelengths. The better they succeed the more the red and blue colors come at the same spot to a pinpoint.

    Getting to the question of "micro contrast" - the whole of the lens contributes to producing an image. The larger a lens the more light gathering and the higher the resolving power gets. But if there are anomalies in the shape of the bokeh and/or the red and blue light do not pinpoint at the same spot there might be a slight diffusion of colors, much like a "haze". Therefor the lens may be sharp yet the colors are not clearly separated from each other.

    PS The above is my understanding, if someone has a better explanation then I look forward to hear it, for all I know I might be totally wrong (and have a faulty memory).

    edit: once spherical sensors are introduced then pure spherical lenses can be used again which will result in higher resultions and better color seperation. I believe this is what already happens at the large telescopes in space.

    Peace to all.
     
  8. MJL

    MJL Mu-43 Regular

    199
    Feb 24, 2016
    Katikati, New Zealand
    Marinus
    The Olympus 75mm F1.8 is highly regarded. Just compare the MTF chart of that, there is no sharp "knee" when it gets towards the edge of the picture suggestion a smooth behaviour.

    olympus-75mm-f1-8-mtf-chart.
     
  9. kwalsh

    kwalsh Mu-43 Top Veteran

    775
    Mar 3, 2012
    Baltimore, MD
    Contrast has *everything* to do with size. The amount of contrast the lens can render depends on the spatial scale of that contrast. That is exactly what an MTF curve shows you - how much contrast you get at each spatial scale. In this case I'm talking of MTF curves that show contrast vs. spatial frequency. Lens manufactures often show another MTF curve in which they show the contrast at a given spatial frequency across the field of the lens - which is also very useful but doesn't really tell you much of anything directly about micro-contrast and resolution.

    For instance most decent lenses have quite high contrast for large spatial scales (a big black area next to a big white area) - say at least 90%. Most people define "resolution" to mean the spatial scale at which the contrast drops to 50%. Micro-contrast is a loose term referring to what happens between the contrast at very large spatial scales and the scale at which it drops to 50%. A lens with "good" micro-contrast will maintain decent contrast out to smaller and smaller spatial scales before dropping steeply towards 50%. A lens with "poor" micro-contrast may slowly roll off from 90% at large spatial scales slowly decreasing towards 50%. Keep in mind "steep" and "slow" are very relative terms as the MTF of any optical system plotted in linear scale looks pretty shallow to begin with.

    Some would argue that "micro-contrast" is all that matters and should just be referred to as "contrast" but the reality is that they really are different things technically. "Macro-contrast" is typically related to large scale low contrast caused by internal reflections that spread light all over the image (sometimes called veiling flare). "Micro-contrast" is related more directly to optical aberrations which affect the point spread function of the lens. Something like spherical aberration would be a classic aberration which spreads light over a nearby region and thus doesn't affect "macro-contrast" at all but is very detrimental to "micro-contrast". Coma and astigmatism also greatly impact "micro-contrast" as you move to the edges of the field.

    And as far as terms go "resolution" is closely related to "micro-contrast" since both relate to the MTF at relatively high spatial frequencies. The typical definition of "resolution" is the spatial frequency at which contrast drops to 50%. All MTF curves are relatively shallow and so naturally where the 50% point is affects where higher contrast lower frequency points are as well in a general sense. A high resolving lens is more likely to also have better micro-contrast than a low resolving lens. But that isn't universally true as the shape of the curve matters. An extreme example would be a mirror lens which for equivalent resolution to a refracting lens would have lower "micro-contrast". In other words if say a mirror lens and refracting lens both have MTF of 50% at 80 lp/mm and thus the same "resolution" we might find that at 30 lp/mm the MTF of the mirror lens is only 70% while the MTF of the refractor is 85%.

    Well before physics is math ;) In this case it helps to have a good understanding of transfer functions across frequency. That's the crux of the difference between "macro-contrast", "micro-contrast" and "resolution".

    So roughly "macro-contrast" relates to veiling flare - how much contrast the lens produces for very large patches of light and dark. "Resolution" refers to what scale the contrast drops to 50% (which visually is very low contrast). "Micro-contrast" is describing what is happening between those two points.

    Note also that all of these things can be altered to a degree in post processing. "Macro-contrast" is not surprisingly controlled by "contrast" sliders in editing software. "Resolution" is in fact adjusted with "sharpening" since it amplifies small spatial frequencies. For that matter plain old "contrast" sliders also technically increase resolution since they increase contrast for all frequencies which is why when visually evaluating lenses it is important to have the same contrast settings. "Micro-contrast" is most akin to the "clarity" slider which is essentially sharpening with a larger radius. It is amplifying medium scale frequencies the most.
     
    • Like Like x 2
    • Winner Winner x 1
    • Informative Informative x 1
  10. kwalsh

    kwalsh Mu-43 Top Veteran

    775
    Mar 3, 2012
    Baltimore, MD
    As an example, both shot in identical light a few minutes apart at F/5.6, processed identically after matching small exposure and white balance differences from the lenses:

    i-Shtq4pP.

    Notice that the contrast on the frame of windows is identical. Also notice there really aren't any more details visible in one compared to the other. However, the upper shot from the 42.5/1.2 is showing significantly more contrast in the marble texture and patterns despite the contrast on the larger scale window frame being the same as the other lens. That, roughly speaking, is "micro-contrast".

    As a warning, be careful if you try to compare the window blinds - since the focal lengths are slightly different the camera was moved a few feet between the shots so the magnification would be identical but that also means the window reflection is slightly different between the two (in the lower one you can see more of a tree only the tip of which is visible in the upper shot).
     
    • Like Like x 1
  11. AlanU

    AlanU Mu-43 Veteran

    484
    May 2, 2012
    If I had to describe micro contrast it's almost like the image it hit with Adobe's "unsharp mask" straight out of camera.

    I looked up quickly the description of micro contrast and found this:

    Micro-Contrast, the biggest optical luxury of the world

    As far as my Canon lenses are concerned my mk2 version zooms wideopen are equally as sharp as my prime lenses. Canon boasts a "nano coating" new SubWavelength Structure Coating (SWC) technology.

    I can confidentally shoot with a Canon zoom (mk2) and achieve images I normally get get when shooting primes....minus the shallow dof.
     
  12. pdk42

    pdk42 One of the "Eh?" team

    Jan 11, 2013
    Leamington Spa, UK
    Hmmmm - I'm no optical expert, but the problem of focusing multiple wavelengths onto the same plane is solved not by aspherical lenses but by clever use of combinations of glasses of different refractive indices. Happy to be proved wrong :)
     
    Last edited: Jul 9, 2016
  13. TransientEye

    TransientEye Mu-43 Regular

    138
    Sep 18, 2013
    Barcelona
    Mark
    It is a real thing. Very loosely, you can think of the "clarity" adjustment in Lightroom etc as a way of adjusting the micro-contrast in an image.

    Micro contrast is not just a property of a lens - it is a property of the complete imaging system, ie lens, sensor, processing, etc.

    My E-M5 shows noticeably worse micro contrast than the older GF1. In part this may be because the higher resolution is stressing the lens more, but I suspect the real culprit is cross-talk between adjacent pixels on the sensor as it remains even when the lens is stopped down. This is a common design problem for high density CMOS sensors, and it is one of the things that could improve with newer implementations.

    What ever the cause, the difference in micro-contrast between my mu4/3 and full-frame systems is fortunately small enough that with minor PP adjustments any difference is largely irrelevant.
     
    • Like Like x 1
    • Informative Informative x 1
  14. MJL

    MJL Mu-43 Regular

    199
    Feb 24, 2016
    Katikati, New Zealand
    Marinus
    In the past it was only the selection of glass that they could work with due to the difficulty of making a-spherical lenses (and the cost of making them a-spherical was prohibative). I do not know about consumer photo equipment lenses but I have come across professional optical stuff which is non-spherical. Even spectacles these days can have some strange non-spherical surfaces custom made to order at a very reasonable price.

    According to this website: 32. What is an "aspheric element"? there are aspherical lenses used in wide and zoom lenses.
     
  15. drd1135

    drd1135 Zen Snapshooter

    Mar 17, 2011
    Southwest Virginia
    Steve
    Nice work. My "what does it look like" questions are mostly answered.
     
  16. Turbofrog

    Turbofrog Mu-43 Hall of Famer

    Mar 21, 2014
    As I see it, Contrast is the ratio of values across the frame, whereas microcontrast is the ratio of values in neighbouring pixels or small areas.

    Put simply, the difference between the regular Contrast slider in Lightroom compared to the Clarity (Microcontrast) slider. They do very different things to an image.

    Macrocontrast crushes detail in lights and darks, while microcontrast reveals detail.
     
  17. oldracer

    oldracer Mu-43 All-Pro

    Oct 1, 2010
    USA
    Thanks guys. This is becoming a lot clearer. I can't respond to everyone, but kwalsh's explanation seemed most helpful to me.

    OK. I understand much more clearly what micro-contrast is, but I do not understand how a lens can affect it other than by what you call veiling, which is clearly a function of the lens and its internal coatings. But the lens doesn't even know that the light it is bending is going be organized into an image when it reaches the sensor, much less what the luminance values in the image are going to be. So ... "transfer functions across frequency?" Do you mean the frequency of the light? Lower frequency red vs higher frequency blue? (I understand chromatic aberration BTW and achromatic and apochromatic lens concepts. I don't think that it what you are talking about.) If not the frequency of the light, what? The lens knows nothing about the image IMO.

    Said another way: "Clarity" and indeed all of Post is basically messing with transfer functions. But again, I don't see how a lens can do that beyond "veiling."
     
  18. Turbofrog

    Turbofrog Mu-43 Hall of Famer

    Mar 21, 2014
    I don't either, but that's because optics are still largely a black box to me, and I'm not that smart. There are a ton of incredibly complicated relationships happening between all the different elements and the aberrations they are trying to correct.

    I still don't fundamentally understand how all the different coatings, materials, and optical designs intended to correct CA and astigmatism and coma and all the rest all work together to produce what we call resolution and contrast. But it's clear that they do. It doesn't seem strange to me that given all that complexity that different designs could enhance the property that is variously referred to as Microcontrast or Clarity or Structure or whatever other word we like to use to qualify a slightly abstract property that has a clear visual impact.

    (By the way, I find it interesting that my practical working definition of Microcontrast is essentially the opposite of everyone's favourite overconfident Internet pixel-peeping windbag, Yannick Khong - his conclusions are often literally the opposite of mine, despite us looking at the exact same images. Maybe his monitor is out of calibration.)
     
  19. AlanU

    AlanU Mu-43 Veteran

    484
    May 2, 2012
    I'm an enthusiast in the audio world. I've spent a fair share of $$$ on cabling and AC filtering.

    What I often find is that regardless of what profession you are in be it an accountant, lawyer, pharmacist, artist/photographer or starbuck's barista we all have opinions on audio and audio quality. A person who own's a $349 Pioneer receiver with a pair of consumer grade speakers may possibly hear little to no difference in "better quality cabling". On the other hand a person with a modest $20-30,000 hifi system will have substantially more resolution in hearing more details.

    "Better quality cabling" meaning sufficient AWG to provide a good means of analog signal transfer. Better cabling isn't necessarily an arm and a leg $$$$. Differences can be heard between 16awg to 10awg if the length of the cabling is too long.

    I think understanding technical terms "micro contrast" is fine. I would be certain many photographer's with great artistic skills cannot describe micro contrast nor do they care. Bottom line is that if it's visually seen in their final "work" that is more than enough to satisfy a professional photographer and his/her clients.

    My previous PL25 f/1.4 was very sharp but it did not exhibit micro contrast that I know exists in other formats. I guess if I owned a M43 sensor with no AA filter it would change the look of the sharpness/contrast.

    If I speak of my Canon gear the new gen 16-35 f/4IS has a different look with incredible micro contrast over my older 16-35 f/2.8L mk2 (older gen mk2 lens with no nano coatings). Virtually all of my FF gear is the newest gen mk2 lenses to obtain this micro contrast. The benefits is that the nano coated Canon zoom lenses provide sharpness equal to top tier prime lenses. This is unbelievable and I appreciate this new technology.
     
  20. kwalsh

    kwalsh Mu-43 Top Veteran

    775
    Mar 3, 2012
    Baltimore, MD
    "Veiling" is light spreading pretty much everywhere in the image from internal reflections. Very large scales affecting global contrast. As you say, pretty easy to understand and visualize.

    Resolution and micro-contrast are instead dependent on essentially "defocused" light surrounding a detail. This defocused light doesn't spread over the entire image, just very nearby the bright detail. As a result it doesn't affect the contrast of large scale details at all but it does lower the contrast of very fine details.

    Lenses do not image a perfect point of light as a perfect point on the sensor. Even a "perfect" lens would spread out the light a little bit due to diffraction. Real optics have other aberrations that cause a point to appear not just as a disk but to also have additional light spread a bit further away.

    Take a look at this helpful image from Wikipedia (you may need to click to see at full size to observe the described effects):

    1000px-Illustration_of_the_optical_transfer_function_and_its_relation_to_image_quality..svg.

    On the left is the MTF, that is the contrast vs spatial frequency. In the middle is the "point-spread-function" or PSF, that is what the optics image given a perfect point of light at the input. There is a one-to-one correspondence between MTF and PSF, you can compute the MTF from the PSF using a Fourier transform.

    Note that both PSFs have a central bright spot of about the same size. Note that both MTFs have some response out to about 500 c/mm. Note that if you look at the center of the resolution star you can see fine detail at the center of both. It is the diameter of this central bright spot that is limiting the "resolution".

    Now the bottom PSF has some additional light surrounding the central bright spot. This light is by no means "everywhere" - it isn't "veiling" the whole image but rather just a small neighborhood around the point of light. It is this ring that causes the MTF to drop so much between 200-300 and it is the reason most of the resolution star looks low contrast and fuzzy even though we can still see resolution at the very center of the star.

    The above is a synthetic example and is very strong for illustrative purposes. But it is very illustrative of how MTF and PSF are related and how they impact resolution and microcontrast in an image.

    Now then, typical optical aberrations produce PSFs much like the bottom one in the illustration above. Not all of them are perfectly symmetric like the one above but they all have the same effect of spreading additional light outside the central disk.

    Again from Wikipedia a bunch of PSFs illustrating spherical aberration:

    Spherical-aberration-disk.

    Note that the middle nine PSFs shown here all have a central disk of just about the same size but that there are widely varying amounts of additional light spread outside of that central disk. These PSFs would produce roughly the same limiting resolution (they have the same size central disk) but vastly different micro-contrast. Again very dramatic synthetic examples for illustration, the real world is not nearly so clean.

    In summary, the MTF curve is contrast vs. spatial frequency. That can be difficult to visualize, both the impact on the image as well as the mechanism in the optics that would produce it. The PSF is directly related to the MTF by the Fourier transform and it might be easier to visualize the effects and causes.

    The "resolution" of the lens is usually meant to be at what spatial frequency the MTF has dropped to a low value. In PSF space it is related to the diameter of the central disk of the PSF.

    The "micro-contrast" of the lens is more loosely defined but would be what is happening to the MTF curve at high spatial frequencies but before the MTF curve has dropped to very low values. In PSF space low micro-contrast would be the result of additional light outside the central disk. Optical aberrations (e.g. the Seidel aberrations) are the root cause.

    The "macro-contrast" of the lens would be how high the MTF gets on the very far left of the curve at the lowest spatial frequencies. As you mention, tends to be limited by internal reflections.

    Oops, sorry that was confusing. Everything I described would be true for monochromatic light, not talking about color/wavelength/frequency of the light. When I said "frequency" I meant spatial frequencies (so for example line pairs per mm). My saying "transfer functions across frequency" was confusingly redundant because usually "transfer function" already means the response of a system plotted with frequency.

    Another visualization of MTF and spatial frequencies is here:

    Understanding resolution and MTF

    Now one thing to realize is that anything that lowers contrast also lowers color saturation. And this saturation reduction is also dependent on the spatial scale of the color contrast in the image. So a "contrasty" lens also produces more saturated colors and a lens with good microcontrast will also better show small color variations on small scales. So even without thinking about chromatic aberration (assume the optics perform identically for all wavelengths) the MTF of a lens has some affect on how color appears.

    Exactly, post processing is just applying a transfer function to the image data and most of the functions we apply (sharpening, clarity) are something akin to the inverse of the transfer function of the optics. The same way the point spread function of a lens effectively attenuates high spatial frequencies the sharpening "kernel" used in post processing is amplifying those high spatial frequencies. A sharpening kernel is exactly like a PSF but of a different shape.
     
    • Informative Informative x 6
    • Winner Winner x 1