1. Welcome to Mu-43.com—a friendly Micro 4/3 camera & photography discussion forum!

    If you are thinking of buying a camera or need help with your photos, you will find our forum members full of advice! Click here to join for free!

Crop Factor with ISO & Aperture: How Sony, Olympus, Panasonic, Canon, Nikon & Fuji Cheat You

Discussion in 'Back Room' started by Klorenzo, May 19, 2014.

  1. Cruzan80

    Cruzan80 Mu-43 All-Pro

    Aug 23, 2012
    Denver, Co
    Sean Rastsmith
    Yes, the larger the focal length the bigger the pupil at the same fstop. Hence my 45/75 example. Nothing to do with what it was designed for. The reason people dont mention it is that it is redundant. Any given focal length and fstop means the enterance pupil is x. If you change one, x changes. This is true no matter the sensor designed for. The reason a speedbooster doesn't work is twofold. One is that the image circle cant be compressed without vignetting. The other is that there is no room to add to the flange distance without losing infinity focus.

    The 12-35 second example is wrong. He will get the same FoV as a 24-70 lens while still having the same fast shutter speeds allowed by 2.8 at the same ISO. He is getting a 2.8 lens. Nothing impied the guy wanted the DoF of a 24-70 lens at 2.8, but this guy told us that is what he wanted to make a point.

    .The problem is the guy takes the thoretical noise to sensor ratio as given even though all sensors behave differently. A 16mp 35mm sensor would have each pixel being 4x the size of a 16mp 4/3 sensor, if built the exact same, or it would need to be a 64mp sensor to be perfectly upscaled. ISO works and has for a while. He is trying to compare s-n ratio in a theoretical sense, the taking that assumption, along with equivalent dof and applying to all of the rest of the things.

    Sent from my LG-P769 using Mu-43 mobile app
     
  2. GFFPhoto

    GFFPhoto Mu-43 All-Pro

    Feb 24, 2013
    I understand. I tried to explain that absolute pupil size is irrelevant taken out of context. But I guess I didn't explain it well enough. Seems obvious to me, but some people just can't wrap their head around it. If you find value in a big pupil, regardless of context, switch to full frame and use f4 lenses. You will have a much bigger hole than any full frame f1.2 :2thumbs:
     
  3. Klorenzo

    Klorenzo Mu-43 All-Pro

    Mar 10, 2014
    Lorenzo
    I think that the thing that bothers him is that the math is wrong when you do the focal length conversion without recalculating the aperture. This error leads to absurd statements like a "300mm f/2.8 equivalent" lens the size of a can of beer.

    He does not want a dof of 2.8, instead with the "correct" (in mathematical terms instead of exposure terms) aperture and the the "iso leveling" (see below) he gets the exact dof as expected.

    Yes, he's making a simplification, IMO acceptable considering the context. He choose to first "level the iso" according to the actual light that reaches the sensor. As a consequence he is forced to use different apertures that gives him the same brightness and dof. Then again, making the "full sensor size equivalence" he finds, with the correct math, an aperture that matches the other ones in terms of dof and brightness (at the "leveled" iso). It's just a different model.
    Is this wrong in any way? I do not think so. Is this confusing? Yes it is. Is this useful? Well, I do not fell like it's worth the effort to discover it.
     
  4. Cruzan80

    Cruzan80 Mu-43 All-Pro

    Aug 23, 2012
    Denver, Co
    Sean Rastsmith
    But the mathematics are wrong. Nobody is saying it is a 300mm lens, just the field of view is the same. He takes field of view, and aperture, which only function is to use for exposure, and then implies companies are deceiving people. The lens has a 2.8 aperture. Its field of view is the same as X. No lies there.


    But ISO was used and is still used with larger than 35mm sensors. So was ISO lying in the 70s between 35mm and MF? No, because the actual light reaching the sensor doesnt actually stand for what he is implying. All of the rest is based on his S-N ratio amd how smaller sensors need more light for it to be the same. Shoot a grey wall with ISO 200, and different sensors (something he doesnt show). The noise will not be 4x worse. Noise is not proportionally related to ISO (something he implies). All of his arguements are based on noise and DoF only to make equivalencies. Math is relative to this. Ever seen the equation where 1=0? When you know what you want to show, making the math appear to support you is easy.



    Sent from my LG-P769 using Mu-43 mobile app
     
    • Like Like x 2
  5. Klorenzo

    Klorenzo Mu-43 All-Pro

    Mar 10, 2014
    Lorenzo
    If you say 2.8 no lies. If you say 300mm equiv no lies. If you say "300mm 2.8 equiv" IMO it's on the border. Anyway as I said in the first post I did not like the "cheating" part.
    I think it would be a lot more confusing for anybody to say: "look, the Oly 45 is equivalent to a 90mm FF f/4 with iso fixed based on the area, but for exposure you have to use the actual value in both cases".


    He says ISO was good for film but may be outdated for digital. No lies from ISO. He does not imply IMO any linear correlation between ISO and noise and have no reason to do it anyway. He does not use noise for anything. He use picture brightness, not noise, to find the correct iso for a fixed shutter speed.
    The math he use is just to compute the aperture given the equivalent, in terms of angle of view, focal length. You get a number, no tricks here. You find it hard to give a meaning to this number? Me too. It only relates to dof equivalence I suppose.
     
  6. dhazeghi

    dhazeghi Mu-43 Hall of Famer

    Aug 6, 2010
    San Jose, CA
    Dara
    As I just explained, you can't use a 'SpeedBooster' to make use of it for the simple reason that it screws up the registration distance. That has nothing to do with the exit pupil. SpeedBooster is not some generic thing - it's a product (and a trademark) by Metabones and it requires at least 20mm of thickness to work with. This is is irrespective of the imaging circle of the lens. Moreover, the imaging circle of the lens is not defined by the maximum aperture as you imply above. A focal reducer lens (of which the SpeedBooster is an implementation) will work even if that full-frame lens is stopped down to the point that the physical aperture is as small (or smaller) than on a m4/3 lens.

    What is true is that a focal-reducer will generally not be very useful with m4/3 lenses and a m4/3-sized sensor, even with a shorter-flanged mount, because the imaging circle will not fully cover the sensor, but this is not a hard-and-fast rule - lenses may have imaging circles that are larger than the sensor they're designed for, and indeed several m4/3 lenses do.

    In any case, considering the needlessly inflammatory title and the general sloppiness of the author regarding basic definitions and terms, it's really hard to take anything he says seriously.
     
    • Like Like x 1
  7. Cruzan80

    Cruzan80 Mu-43 All-Pro

    Aug 23, 2012
    Denver, Co
    Sean Rastsmith
    But the issue is that no one (besides this guy) fixes ISO based on area. Neother do they say "300 2.8 equivalent" from the manufacturers. They only say "300mm equivalent".

    Then why if he doesnt care about noise is he sayong you need to do the equivalent squared? To make the sensor "receive the same amount of light", whivh he shows by how noisy the files are, and the Signal to noise ratio. At a given shutter and aperture, ISO is locked for a given brightness.

    The problem is the terms arent in the same units. The aperture is a function of the focal length of that lens. End of story. Comparing angle of view on different sensors doesnt change the inherent focal length. Therefore the aperture is solely a function of the lens. He doesnt say that it behaves the same with regard to the amount of DoF as a lens which is 2x the aperture on 35mm, he says it must be an actual 2x aperutre which isnt true. When he talks about "bokeh" (see how the window is more blurry here than there), he is confusing the effect the aperture has on DoF, vs the aperture itself. What he doesnt show is that if he cropped the same picture with the same lens by ~2x, then compared 35mm to m4/3 with the same aperture, it would be an essentially identical picture.

    What I can do on M4/3 is get more in focus, at the same relative noise level (up to ISO 1600) with the same shutter speed. The only part he got right is that if I wanted to get essentially the exact same shot with 35mm, I woud need to use 2x the aperture for the same DoF, meaning adjusting either the ISO or shutter by two stops. Which everyone already knew. Nothing changes regarding aperture or ISO between sensors.

    Sent from my LG-P769 using Mu-43 mobile app
     
  8. Klorenzo

    Klorenzo Mu-43 All-Pro

    Mar 10, 2014
    Lorenzo
    I do not imply anything about the maximum aperture and a few other things you wrote.

    25mm / 1.4 = 17.85
    50mm / 1.4 = 35.71

    You can repeat this for ANY aperture you choose. And you can theoretically ALWAYS build a "speedbooster like" product for any lens designed for any larger sensor: FF, MF, APS-C, etc. because this lenses project a bigger circle. And the "speedbooster like" adapter can take and use extra light. The point is: where does this extra light comes from?

    This to me means that when I set my adapted Nikon 50mm AI-s to f/5.6 and an Oly native 25mm to the same aperture I expect to find a bigger hole in the middle of the blades that control the aperture (I hope that in this way it's more clear, entrance pupil should be the exact term). I'll check later with a ruler. So the aperture on the 50 is as big as the 25 opened by two stops more. This does not imply anything, it's just the physical size of things, if I got this right.

    So he says: "wait, if you multiply the focal length by 2 AND KEEP THE SAME FRONT HOLE SIZE you get a different aperture". Yes, fine, correct. But why should I ever do such a thing? I think you can say that:
    1. it's obvious that, if one should ever repeat the calculation of the aperture, he should multiply the pupil size too. This IMO address his main argument.
    2. there is no reason to recompute the aperture at the equivalent focal length because the equivalent computation has meaning only regarding the angle of view. You should not use that number in any other context and purpose.
    3. is the fact that the rx100 entrance pupil at 2.8 has a smaller physical size then a 24-70 FF relevant and not enough obvious to the point of accusing the makers of cheating? I do not think so. The pupil of the 24-70 at 2.8 is probably bigger then the whole lens.
     
  9. Klorenzo

    Klorenzo Mu-43 All-Pro

    Mar 10, 2014
    Lorenzo
    Because it's the way you compute the area given a diagonal that it's two times the other one. It's geometry, there is no noise there. Two times the diagonal, four times the area, four times the light per surface. The part about noise is a general discussion to show that less light give more noise. Giving a comparable amount of light per squared inch you get a roughly comparable amount of noise. No demonstrations here. He says he will get into more details about noise in another video because it is not the point of this video.

    I pass on the part about what he does NOT say, does NOT show, etc.
     
  10. mattia

    mattia Mu-43 Hall of Famer

    May 3, 2012
    The Netherlands
    Short response:

    1. The physical characteristics of a lens are fixed. Aperture - the ratio of focal length to pupil diameter, is fixed. Physical property of the lens.

    2. The only equivalence that matters in terms of getting the desired framing and correct exposure is the angle of view and the F-stop (actually T stop, amount if light truly transmitted through the lens). The former depends on focal length and the interaction with sensor size (the camera is merely 'cropping' the full image circle). The later remains a physical constant. It's about letting the same amount of light in per exposable sensor surface; the fact the hole is physically smaller is totally irrelevant.

    2. The fact the RX100 has a small sensor means the size of the entrance pupil relative to the sensor size and relative to the focal length (actual, not equivalent) is identical to the hole size on the FF lens with the same aperture and (equivalent, not actual) focal length. And the amount of light hitting the sensor is the same per unit of surface area. The FF gathers more of it because it's bigger.

    3. Focal reducers/speed boosters change the physical characteristics of the lens and how light passes through it. And this delivers more actual light. Hence the equivalent conversion makes sense to use, because it affects exposure. It's basically a 'reverse crop' of the sensor.

    3. DoF just depends on sensor format, all else being equal. Sometimes shallow is nice (which is why I have an A7r), sometimes it really doesn't matter (in comes the E-M1) and sometimes I want a flexible camera that fits in my pocket, and the fact I can shoot at f/1.8 and still have a good amount of the scene in focus a a big advantage (my rx100).
     
  11. Klorenzo

    Klorenzo Mu-43 All-Pro

    Mar 10, 2014
    Lorenzo
    mattia, I agree with everything you wrote. And I'm sure Northrup does too because I never found he says anything different in this video (or in other videos for what that matters). And I also find nothing really new or complex here.

    The only point I think could be better phrased is the one about the speed booster: of course the speed booster does not "actively sucks" in more light (creating a small black hole! :)  ), obviously it just concentrate the available light on a smaller surface.

    So if you accept that I understood what you wrote and I agree with it you should ask yourself what exactly is the point that is stated in the video and that IMO is NOT contradicting anything that you wrote here.
    He just introduce a new concept the "equivalent aperture". Maybe it has a bad definition from the beginning, it's a stretch of another concept in a different concepts, it's confusing regarding exposure, etc. but I think there is not much more than this.
     
  12. lightmonkey

    lightmonkey Mu-43 Veteran

    479
    Dec 22, 2013
    speedbooster... just an additional convex element added to the raypath to the sensor. it 'concentrates' the light.

    like a magnifying glass.

    it moves the focal point forward, which shortens the FL. diameter of aperture (=opening) stays the same. F=FL/d... drop the FL and you get a numerically smaller F-value.

    intuitively, we're condensing more light in a smaller area... higher light intensity (per pixel)... thats why we're gaining "speed".


    if you want to 'boost' a m43 lens on an m43 body....sure you can design one.... you'll just reduce the projected image circle so the outer pixels are unexposed/underexposed. i.e. throwing away pixels (narrowing field of view !!!)
     
  13. Cruzan80

    Cruzan80 Mu-43 All-Pro

    Aug 23, 2012
    Denver, Co
    Sean Rastsmith
    I think the biggest issue I have is that he is claiming they are lying, and introducing equivalent aperture, whithout referencing why you would care, and that he seems fixated on ISO being crop squared. The overall amount of light hitting the sensor doesn't matter. No-one computes things this way. It is almost like he started by deciding that he wanted to go with equivalent aperture, and then worked backwards from there for the generation of his arguments. Then when he presents them the way he does, it follows naturally that if you buy into the first, the second is true and so on. But the problem is the first thing he mentions doesn't matter. Measuring the overall photons hitting the sensor is an asinine way of trying to work out equivalent shots. So if I don't agree with that, why would I change the ISO by the crop squared? Means that I still agree with "equivalent aperture" as it relates to DoF ONLY. Everyone already knew that. He then seems to think that he needs to justify why we should then change the ISO (changing shutter would be too obvious).

    Sounds like you understand the aperture as a function of the opening of a lens. That is why I suggested comparing two m4/3 lenses and a 35mm one. They will all follow the same rule, though the image circle is projected differently. A 25mm m4/3 has a smaller entrance pupil than a 50mm 35mm, lens which is smaller than a 75mm m4/3 lens when they are all set at the same aperture.

    ISO is a standard. The slide about how it would be called different things is wrong. Ever heard of the International Standards Organization (ISO)? ISO 9000 for manufacturing?
     
  14. mattia

    mattia Mu-43 Hall of Famer

    May 3, 2012
    The Netherlands
    My point is aperture equivalence is ridiculous. Most people have never shot full frame, so why should that be the standard? Depth of field control can be an artistic choice facilitated by focal length and sensor size and aperture selection, but aperture is simply a physical lens characteristic. Equivalent FOV just helps folks moving between sensor sizes get a feel for what they will be able to frame.

    Mostly I don't think there's anything dishonest going on anywhere. And I just don't see the point to the video. At all.
     
  15. bikerhiker

    bikerhiker Mu-43 Hall of Famer

    Dec 24, 2013
    Canada
    David
    Circle of confusion is a bunch of photographers sitting around discussion depth of field!

    Parameters of Depth Of Field. This is basic photography they teach at least in my school.

    1, Aperture
    2, Distance from the subject
    3, Angle Of View

    Despite a lot of instructors and experts who teach online, focal length has nothing to do with Depth Of Field! Na da..

    All lenses have the same depth of field. Now you would wonder, why then a wide angle lens has a greater depth of field than a telephoto lens. Well, it's because the smaller the objects (the tighter the edge) they are the more they are in focus and the larger the objects (the further apart the edge) the more out of focus they look to be. This is more apparent if anyone here has shot film. When you look at a contact sheet, everything look relatively sharp. But until you put it through the enlarger is when they don't look as sharp.

    Here's another interesting comparison. An 80mm lens on a 35mm full frame sensor is a telephoto lens. The same 80mm lens on a 21/4 film format is considered a standard lens and last but not least, the same 80mm lens on a 4x5 view camera is now a wide angle lens. What has changed?!? The angle of view and is dictated by the size of the capturing device. Nobody in the film days would bother to convert back and forth between 35mm and 4x5 view or 6x6 Hassy. They just shoot.

    Here's another misnomer.. A lot of people teach that as you stop down the aperture, you get more things in focus. That is not true, at least according to physics or unless you utilize the Schiemplug principle which is the basis of the Tilt Shift lens. According to physics, you only get 1 plane in focus and anything from that 1 focus plane is the increasing area of out of focus that gives you the illusion of sharpness and that denotes the use of your aperture and your target output resolution. You either open it up or stop it down to varying this area of out of focus. Let's use another example. How many photographs you need to take if you want 4 objects in focus? You will need 4 photographs!

    But here's the kicker. The blur effect of the out of focus area is greatly influenced by the target resolution; in many cases print. For ease of understanding, I will use line pairs as a measurement of resolution. Most modern lenses today can resolve 200 line pairs. 35mm film can resolve about 150 to 175 line pairs. So the lens has more resolution than film. The blur area will be less for a given aperture because if the out of focus area is 175 line pairs, then film will record it as 175 line pairs and this will give you the illusion of sharpness. Print has a resolution of 50 to 75 line pairs. If we are to print to film, what registers as 100 line pairs on film will be registered as 75 line pairs due to the lower resolution of film and therefore will have a greater area of sharpness based on this illusion, compared to the film media it was used to get prints from. Therefore even if you have used the same f-stop through, the blur effect will be less pronounced once you print to film from film with the sharpest lens. So the target resolution is key to determine the resulting DOF. The number of the aperture means nothing until we understand final output.

    Now with a digital camera, if the lens can capture 200 line pairs, sensor can capture 200 line pairs but the latest high end printers can print on the upwards of 300 line pairs. So now, what is the photographer is MORE CONCERNED about? A photo in focus or a photo in blur. I would worry about the photo in blur. Why, because the printer resolution is greater than the capturing device and the ramping of the blur cuts sooner on a full frame rather than on micro 4/3. In fact, if you listen to a lot of portrait photographers, their preferred choice of aperture is, in full frame, f/5.6. Not f/1.4, not f/2.8 and not even at f/4. It's f/5.6 and it's got to do with printing to those high end printers. In full frame, f/5.6 gives you a balance of out of focus blur and in focus perception. Using m4/3, an aperture of f/2.8 gives an equivalent blur effect on a high end digital printer as a full frame would @ f/5.6. In this case, you don't need to stop it down.

    The video tutorial assumes the person's final image output destination is a computer monitor and if that is so, only applies to that one output medium only and ONLY TO HIS MONITOR. Not everyone has a Retina display nor a high end display! But by him generalizing that this is the acceptable norm means he is as equally confused like all the photographers sitting around the circle of confusion.
     
  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.