1. Reminder: Please use our affiliate links for holiday shopping!

Lens resolution and sensor MPs

Discussion in 'This or That? (MFT only)' started by Speedliner, Sep 23, 2015.

  1. Speedliner

    Speedliner Mu-43 All-Pro

    Mar 2, 2015
    Southern NJ, USA
    Rob
    Last edited: Sep 23, 2015
  2. David A

    David A Mu-43 All-Pro

    Sep 30, 2011
    Brisbane, Australia
    No link so I don't know what you were referring to, but…

    If some lenses are sharper than others, and we know that they are because of personal experience and various test reports, then there has to be a point where you can see the difference and at that point you're seeing the maximum resolution of the less sharp lens. Sensors have resolution limits as well. The maximum resolution you can get from a given lens and sensor combination will always be determined by which has the lower resolution capability, lens or sensor.
     
  3. Speedliner

    Speedliner Mu-43 All-Pro

    Mar 2, 2015
    Southern NJ, USA
    Rob
    Oops. Updated the first post.
     
  4. Clint

    Clint Mu-43 All-Pro

    Apr 22, 2013
    San Diego area, CA
    Clint
    When the Nikon D800 was first announced there were many saying you needed top of the line lenses. Simply a myth.

    In fact what happens with the D800, the images are better from almost any lens than with a smaller resolution sensor and the same lens - and they can be some outstanding images! So larger resolutions (at least up to 36MP) pulls the best out of any lens. I'm sure there is some point at which a gain in IQ would be minimal - but for 36mp or less I would not worry about it.

    Of course if you want the best a sensor can deliver - then you probably need to start with the best lens available.
     
  5. eteless

    eteless Mu-43 All-Pro

    Jun 20, 2014
    Earlier this week I was projecting a picture of a car taken on microfilm on a wall, the image was about 4 meters tall and 6 wide. At this size you could see scratches on the paintwork and could tell that the screws holding the number plate on were Philips heads (5mm wide on the projection). The lens used to take the photo hardly matters, the lens projecting the image was an uncoated 3 element f2.5 lens made in the early 1950s. I have no clue how much information is actually stored in the microfilm, given it's about the size of a small postage stamp and there was still fine detail I don't think it really matters.

    The lens might be the limiting factor however I think that modern lenses are more than good enough even if the megapixels on m4/3 sensors were to quadruple.
     
    • Agree Agree x 1
  6. pellicle

    pellicle Mu-43 All-Pro

    Feb 10, 2010
    Southport, OzTrailEYa
    pellicle
    its true and an interesting subject. Its intuitively obvious if you examine the extremes. For example this little fella was made for C mount and is quite old, and even in the areas it covers does not have the resolution of the kit 14-45 lens

    lens.

    So if you look at this comparison the fine details in the pine trees are lost on the elgeet image:

    center.
    There comes a point where the ability to capture the image so far exceeds the ablity of the lens to project it. I see a strong relationship with audio recording where eventually your sampling frequency and bit depth are so far in advance of the mic that you're literally wasting recording data.

    I happen to like the Perceptual Megapixel concept developed by DxO ... its a digestable metric which seems to be more comprehendable than LP/MM or MTF charts. It rates lenses on bodies and you'll see that any given lens will perform differently on a different body. Indeed format of capture makes a big difference as if you are capturing on a smaller higher density sensor you'll need sharper to get the same sharpness over a larger less dens sensor.

    I tried to demystify it on my blog post here:
    http://cjeastwd.blogspot.com/2014/01/a-perceptual-mega-pixel-explained.html

    the premise of my exploration is that if you have 16 "perfect megapixels" then scaling that down - saving - scaling up will show loss. If you have 16 megapixels which equates to 7perceptual megapixels then the same process should not show loss.

    hope that helps you a bit ...
     
    • Like Like x 1
  7. tkbslc

    tkbslc Mu-43 Hall of Famer

    That article uses the dreaded Perceptual M-Pix, which is something DxOMark made up. The theory probably has some merit, but I can't take it seriously when it is based on made up metrics.

    The more relevant question is whether 20MP has more resolution than 16MP and if 16MP has more resolution than 12MP. If so, then it doesn't really matter what the "true" resolution is, we are still getting the extra you expect when you buy a camera with a new generation sensor.
     
  8. eteless

    eteless Mu-43 All-Pro

    Jun 20, 2014
    I'm okay with things that are made up as long as I know the theory behind them. DxO doesn't release any of the workings behind the theory, there is no peer review, you cannot double check or replicate their numbers.

    Can we at least all agree that the DxO One is the best camera ever made?
     
    • Funny Funny x 4
  9. pellicle

    pellicle Mu-43 All-Pro

    Feb 10, 2010
    Southport, OzTrailEYa
    pellicle
    yeah ... many people seem to be OK with Fahrenheit, which is a made up metric ( along with inches, feet, yards, acres, perches, grains which are essentially made up ...), as long as you know what the metric is doing and its intent then it does its job.

    fahrenheit_to_celcius.
     
    • Agree Agree x 2
  10. bikerhiker

    bikerhiker Mu-43 All-Pro

    Dec 24, 2013
    Canada
    David
    This is a typical fear tactic to get people to think they have to buy a better lens or simply to invest in good software to squeeze out the best the lens can do.

    There is no limit of what a lens can resolve. After all, what does a lens do? Pass light through to the sensor. Is there a limit to that? So yes.. It's a myth. What a Nikon D800, Sony A7RII or the Canon EOS 5DSR do is able to record every lens' weaknesses. The lens aberrations like LCA, CA, SA are recorded in detail. What tend to do well with high resolution sensor are lenses that have relative good micro-contrast, but they aren't the lenses that you would want to use for artistic flair.

    I think the issue here stems from the fact that people don't understand that lenses are made like paint brushes. Too many people these days are to fixated about the lens' capabilities to get a super sharp, finely detailed photos from an ultra high-rez 50MP sensor. What's important is that the basis for a high resolution sensor is to allow for larger files, which helps printing large. Any lens can do that. The question is; how a perfectionists are you?
     
  11. Klorenzo

    Klorenzo Mu-43 All-Pro

    Mar 10, 2014
    Lorenzo
    I get what you mean but in common terms this is just plain false. Usually you draw a small pattern on a paper, place a focuses lens in front of it and look at what appears on a screen on the opposite side of the lens. If you clearly see the same pattern then the lens has enough resolution, if you see a small blurry gray circle it does not. It's true that the details in the blurry circle are infinitely fine but they do not match the subject.
    If the blurry circle is smaller then a sensor photosite then the lens is good enough for that sensor, otherwise it's the opposite.
     
  12. Speedliner

    Speedliner Mu-43 All-Pro

    Mar 2, 2015
    Southern NJ, USA
    Rob
    I understand the skepticism around DXO scores. It seems that their rankings match our impressions reliably though...O75,1.8 one of the sharpest, O40-150R low on the list.

    I wish their were metrics for contrast and rendering. To me that's where the most noticeable differences occur.

    Using the 40-150R as an example...I think it's a great lens for the price that punches above its weight. But it doesn't compete in resolution to my 50-200. Color doesn't pop quite as much, contrast less defined and I don't know what to call or what parameters really make an image pop while still looking natural, but the R is flat in comparison to the 50-200, 45,1.8,etc...

    It seems likely that a mediocre lens limits the gains in sensor capability, but there would still be some gain.

    As always though the quality of the light, the dynamics of the setting matter most.
     
  13. bikerhiker

    bikerhiker Mu-43 All-Pro

    Dec 24, 2013
    Canada
    David
    What you are describing is the lens aberrations, because as light enters through the lens elements, it bends. What you are seeing is how well or poor the lens designer corrects for these aberrations. These aberrations always exist in all lenses; some more than others. These aberrations haven't posed a problems up until now, because the advent of the high resolution sensor and the Nyquist theorm of oversampling. These flaws, known as the lens aberrations are part of that detail and the cause of the blur you see as you try to resolve the line pairs on a standard resolution chart. Reduce the flaws, then you get a higher resolving lens which is what the Zeiss Otus and Milvus are. But to say that you can't simply use a lens not designed for 36MP or 50MP or you need to buy a software to do this because it's not going to give you great results is not simply true and is a classic fear tactic.
    Even Ansel Adams said this..
    There is nothing worse than a SHARP image of a FUZZY concept.

    Now turn it around.

    There is nothing worse than a FUZZY image of a SHARP concept.

    We know the latter is true. Vivian Maier and Diane Arbus are 2 photographers that do just that.
     
    Last edited: Sep 24, 2015
  14. Turbofrog

    Turbofrog Mu-43 Hall of Famer

    Mar 21, 2014
    Many people strongly dislike Fahrenheit, because it's silly and arbitrary.

    The best part of it is, that description of his methodology is a little bit off: "On Fahrenheit's original scale, the lower defining point was the lowest temperature to which he could reproducibly cool brine (defining 0 degrees), while the highest was that of the average human body temperature." In actuality, 100 degrees was decided to be his particular body temperature. And it turns out that he had a bit of a fever at the time. Which is why the average healthy human body temperature is 98.6 degrees, not 100 degrees.

    Like I said, silly and arbitrary.
     
    • Agree Agree x 1
  15. Klorenzo

    Klorenzo Mu-43 All-Pro

    Mar 10, 2014
    Lorenzo
    I think there is no doubt that a lens performance/resolving power does not change when the sensor change. And if the picture was good enough before is going to be good enough with the new 50 MP sensor.

    But it makes sense to expect better resolution from the new sensor and some lenses could not have enough resolving power for this. As you said not all lenses are created equal and there is a limit on how many details/lines they can resolve before the blur (caused by aberrations, sure, there was nothing new to me in what you wrote). So my personal conclusion is that for pictures where resolution is relevant you can get better results with a new 50 MP sensor and only the best lenses. Do you agree?

    If marketing people use this things to scare people or not, of course they do, does not change these things I think.

    BTW I do not know which magic software you are talking about.
     
  16. Turbofrog

    Turbofrog Mu-43 Hall of Famer

    Mar 21, 2014
    A big part of my skepticism of DXO's P-MP value is that all their lens scores for the GX7 are nearly half that of every other 16MP M4/3 sensor. Supposedly they even tried two samples after they saw the results.

    It's utter nonsense is what it is, or at least totally incompatible with anything that can be perceived by a real human's eye rather than a computer algorithm. If anything the GX7 is capable of sharper images than my GX1, due to the presence of the e-shutter.
     
  17. bikerhiker

    bikerhiker Mu-43 All-Pro

    Dec 24, 2013
    Canada
    David
    The article the OP tagged alludes to the lens correction software advertisement but I never disagree with you nor the article on the notion of resolution. What I would disagree is in the context in which how a lens should be graded. And that is, the important of its maximum capable resolving power. We know that lens designers design lenses to provide a certain look for a certain application in working with the compromises they face. Sometimes a very expensive bokeh friendly lens may not achieve the highest resolvency as a lens designed to cut air like a sharp Samurai blade.

    What I have found is this modern fetish with super sharp lenses, high resolution bodies and ultraprints and that they believed that is the gauge mark and the optimum plateau we must reach. It's funny that authors and web bloggers who promote these things always have viewers coming back asking; "What the **ck is the image about?!?". Sometimes, I even asked those questions.. Then you get these photographers firing back salvos of comments that basically say their viewers simply just don't get it or I'm so dumb that I can't interpret and they have to write an essay or a novel explaining. Perhaps they had a SHARP image; but it's a fuzzy concept they are unwilling to admit.

    I treat lenses as paint brushes. I use them when it deems necessary to create atmosphere, mood and ambience to incite an emotional response. Usually people get my message within a few minutes. I don't need to write an essay or a novel to let them know what's about. But then, I used a lens that's not so sharp. If I have used a really SHARP lens, maybe the message will not be as effectively communicated. Choose the lens appropriately for the task.

    This was what I was leading to. Sharp Zeiss Otus and Milvus are not the end all be all if you know what you're doing photography wise..
     
    Last edited: Sep 24, 2015
    • Agree Agree x 1
  18. Mellow

    Mellow Mu-43 All-Pro

    Aug 27, 2010
    Florida or Idaho
    Tom
    You want to explain this one to me? I'm pretty sure I can't resolve the tail number on a 747 flying at 50,000 feet with my 20mm f/1.7, regardless of the sensor. Nor can I see Jupiter's spot or Saturn's rings. So what gives?
     
    • Like Like x 1
  19. bikerhiker

    bikerhiker Mu-43 All-Pro

    Dec 24, 2013
    Canada
    David
    There are various ways to design a lens. If you search in this forum about bokeh that I wrote, you will find it. The limit is not the glass, BUT how the designer intended the lens to do. If you specifically want a lens to resolve purely on resolution, then you should get a lens to do so but it MAY NOT fair in doing other things other less capable lenses can do well -- bokeh, contrast and color/bw reproduction.

    It's like a car. There is no limit how fast you can design a car. As technology moves forward, car speed will increase. But what I found to be unfair is the gauge of which how you rate cars; that is ONLY in horsepower and nothing else. In a marketing standpoint, lots of horsepower means lots of horsepower and that sells. But what about a Toyota Prius which doesn't have a lot of muscles compared to say a super car like a Ferrari or a Lotus. If we simply refer to a car's best traits as in horsepower only, than a Prius is not a good car. Even on Germany's autobahn, a Prius is BAD and a Ferrari is good. This is simple black and white thinking, while neglecting the Prius purpose is not designed purely for performance, rather it's designed for fuel economy for which the purely performance muscle enthusiasts would ignore. Or you can design a Toyota Prius to go as fast as a Ferrari; but then it wouldn't deliver the best fuel economy would it.

    Judging a lens is not as easy as black and white, but unfortunately in the fetish world of high resolution sensors and Otus and Milvus lenses, the purely performance people think this is the ultimate goal everyone should reach. It's not.
     
  20. Mellow

    Mellow Mu-43 All-Pro

    Aug 27, 2010
    Florida or Idaho
    Tom
    I understand your point, but it's just plain inaccurate to say that "There is no limit to what a lens can resolve." Sure there is. You can't design a 25mm m43 lens to resolve the rings of Neptune on a m43 sensor any more than you can design a car to go the speed of light. What I think you're saying is that there are always design compromises, so that if you optimize for one thing you'll have to accept sacrifices in another. "Things" including size, weight, aberrations, cost, etc. We agree.

    The point of the thread, however, was that there are inherent limitations in resolution posed by both the sensor and lens. That's still true.
     
    • Agree Agree x 3