Personal View site logo
Make sure to join PV Telegram channel! Perfect to keep up with community on your smartphone.
Putting the FULL FRAME confusion to bed
  • I can't tell you how many producers have freaked out over the "crop factor" and thought they needed a 9mm lens to shoot an 18mm shot on the GH2 when using PL mount lenses. So I am here to try to explain this "full frame" nonsense as clearly as possible. (It is important to note that you will still get a little more crop from a 4/3 camera than you will from the RED or the 7D, but it's nominal.)

    Okay, for starters:

    CP.2s are not real cinema lenses. They are still picture lenses that have been geared with a clickless aperture ring and high end follow focus gears. They don't actually project a cinema sized image circle onto the sensor.

    image

    Most DSLR shooters are unaware that every digital cinema camera has a sensor that is ROUGHLY the size of a Super 35mm piece of motion picture film. This is true for the RED, Alexa, F65, Phantom, Genesis, Viper, 7D, GH2, FS100, etc.... The only aberrations are the 5D, 1DX, and (edit) the D800 which have a sensor the size of a 35mm piece of still film.

    image

    Most of us use Still Picture lenses (Nikon, Canon, Zeiss, Voightlander, Nokton etc.) on our DSLRs, and as a result, the smaller sensor crops the still frame image circle projected onto it by a factor of anywhere between 1.5x to 2x. Except on the 5D, 1DX, D800 which have a "Full Frame" sensor that takes advantage of the full projection circle from these lenses.

    BUT

    Cinema Lenses - those big PL mount suckers (Your Cooke, Angenieux, Panavision, Arri, Eagle, Zeiss, etc...) project a smaller image circle which is the size of a Super 35mm piece of motion picture film. Which is why they can't be used with the 5D, and why Zeiss created the CP.2s for DSLR shooters, namely, 5D shooters, who wanted a lens that would project the same "Full Frame" image circle as the ZE and ZF series, but machined to be used with a professional follow focus geared for Cinema Lenses, and would have a manual, clickless aperture ring - just like the big boys.

    SO

    Using Still Lenses - To get the same framing, the 5D will have to be placed closer to the scene than all the other cameras. (Like almost twice as close)

    But when dealing with Cinema lenses on a GH2 (not CP.2s for reasons discussed above), you will not get the same FOV as say a RED or Alexa, but it is nowhere near the drastic 2x crop factor that so many producers seem to think. It's closer to about a 1.25x crop. (So add 25% -- 50mm FOV becomes a 62.5mm FOV, a 14mm FOV becomes a 17.5mm FOV, 18mm FOV becomes a 22.5mm FOV, etc. -- To get this multiply the lens focal length by 1.25.)

    Ain't camera technology fun???

    sensor and lens comparison.png
    1920 x 1440 - 79K
    sensor.png
    1103 x 869 - 20K
  • 44 Replies sorted by
  • LOL, Panavision too busy with dinosaur and $ilverscreen taxonomy to note furry M43 mammals scurrying about?

  • Another handy little guide, from my friends over at Panavision.

    image

    sensor_compare.jpg
    600 x 514 - 40K
  • I stopped worrying about this long ago. I bought a director's viewfinder that had framing for academy 35mm and it matches pretty much perfectly with the Nikon lenses I have when on the GH2. Honestly I'd rather use the M4/3 sensor than a full frame. I used a 5Dmk2 on a shoot a couple weeks ago and while being nice, it was damn hard to focus it wide open, that razor thin DOF was killing me.

  • There is a series of shots on Vimeo from the Portland test of 'medium' priced PL-mount lenses (20K to 40K U$ for a full set!) which shows clearly that visually the DOF is quite constant when you have the same framing. Only the perspective changes (of course). See the portrait in part 2.

    BTW, I second that producers with cine experience will always judge FOV referring to 35mm film, not photography. To them you can easily explain that the FOV for the same lens is just a bit narrower on the GH2.

    Beware of producers who refer to full-frame photography. They probably don't have a clue about other aspects of cine work either!

  • @Shian I have great respect for your knowledge and your desire to clarify these factors for PV readers, and it certainly is not my intention to add any more confusion to an admittedly confusing issue. That having been said, in my view you are the one that is adding confusion to it at the moment.

    To address your statement that "when keeping the framing the same (equivalent FOV) there is zero difference in DOF between any lens at any focal length when used at the same aperture", that simply is not true. See my edited-in note in my last post about the differential DOFs of a GH2 with 25mm f/0.95, 5D with 50mm f/0.95, or MF with 114mm f/0.95. All of these may be set to the same f/0.95 aperture, and will have a roughly equivalent FOV, but they produce dramatically different depths of field. To simplify this further, put all of these lenses on a GH2, and crop the 25 and 50 to match the FOV of the 114mm. This will again produce vastly different depths of field.

    Backing up and zooming in does not "keep framing the same", nor does it produce an "equivalent FOV". Zooming in by definition reduces FOV... that's precisely what "zooming" is. Moreover, the distance of the subject has no bearing on FOV; FOV is dependent only on focal length and the dimensions of the sensor (and any cropping that may be done in-camera or in post-production).

    If by "equivalent FOV" you mean only that the subject in focus takes up an equivalent area of the image, and you're talking about moving away from the subject to maintain equivalent framing in that regard, then it gets even more complicated, and you certainly can't count on the same aperture producing the same DOF. Looking at that equation I posted above, you'll see that the relationship between the focal length and distance from subject in deriving DOF is more complicated than that. In particular, the relationship of [2 * focal length^2 * distance to subject (distance to subject - focal length)] means that focal length is squared, while distance to subject is multiplied by itself minus the focal length. This basically means that as focal length increases and distance increases (in other words, stepping back and zooming in), the ratio of the numerator to denominator decreases, meaning a narrower DOF. It's subtle, because distance is usually measured in meters while focal length is measured in millimeters, but it is happening.

    In any case, all of this presupposes a subject that's at least a few feet from the lens. Another reason that your proposed rule of thumb fails is that DOF equations go haywire with any object closer than a certain distance. Here it gets even more complicated, but Wikipedia covers it somewhat coherently: http://en.wikipedia.org/wiki/Depth_of_field#Close-up . This is relevant, because at a certain point a wide-angle lens very close to a small object will in fact produce a shallower DOF than a telephoto lens further away set to the same f-stop.

    TL;DR: depth of field is a highly complicated optical phenomenon. If you don't want to deal with complicated equations, fine -- I personally think just experimenting with lenses and getting a feel for how DOF relates to aperture, focal length, and distance from subject is a much better way of understanding it on a practical level anyway.

  • @Sangye apples and oranges.

    When keeping the framing the same (equivalent FOV) there is zero difference in DOF between any lens at any focal length when used at the same aperture. I was trying to bust the myth that I hear nearly everyday. "Just back up, zoom in and reduce the depth of field..."

    Beyond that, sigh yes, focal length is a factor. So, if you have any other extraneous facts, please, by all means, add some more confusion to the issue (and thus defeat the purpose of this thread, which was to try and clear up the confusion regarding lenses and crop factor, not create more confusion.)

    I guess I'm just going to have to shoot a tutorial over at Panavision and allow the real experts to clearly explain everything, since I don't seem to be making any headway in that department.

  • @Shian regarding the relationship between focal length and DOF, I'm afraid you're mistaken. The equation for calculating is as follows:

    Depth of Field = [2 * focal length^2 * distance to subject (distance to subject - focal length)] / [(focal length^4 / circle of confusion criterion^2 * f-stop^2) - (distance to subject - focal length)^2]

    More here: http://www.naturescapes.net/102004/ps1004.htm

    Clearly, depth of field is mathematically dependent on focal length, distance to subject, "circle of confusion criterion", and f-stop.

    Here's a chart of the DOFs of a 25mm, 60mm, and 100mm lens all set to f/8: http://www.naturescapes.net/102004/Fig4.gif

    All of this should be obvious, though. If you compare how a 200mm f/2.8 lens renders an object and its backdrop to an equivalent crop from a 20mm f/2.8 lens, it's clear that DOF is highly dependent on focal length, well beyond what could possibly be explained by this "optical illusion" that the site you linked references.

    After all, the f-stop of a lens is the ratio of the focal length of that lens to the diameter of the entrance pupil. A 200mm f/2.8 lens has a much larger diameter of entrance pupil than that of a 20mm f/2.8 lens. The only thing that a 20mm f/2.8 and 200mm f/2.8 have in common is that they allow a similar amount of light through to the sensor... which is not at all the only factor relevant to establishing DOF.

    Edit: more directly on topic, this is why a GH2 with a 25mm f/0.95 Nokton can't touch the DOF of a 5D with a 50mm f/0.95 Noktor, let alone a medium format camera with a 114mm f/0.95 Perkin Elmer, despite all having the same max f-stop and roughly the same field of view. It has nothing to do with the sensor sizes of the cameras, only that to have equivalent fields of view, these cameras need lenses with significantly different focal lengths, which in turn have significantly different DOFs.

  • @christianhubbard face palm No, I won't. A 14mm would be more than sufficient. But 18mm is still plenty wide on a 4/3 sensor.

    Both shot with Generic 35mm Cinema lenses one an 18mm on a Red MX, the 2nd a 14mm on the GH2 (or the setting I've used to determine GH2 fov which has been pretty damn accurate. (sorry iPhone not on tripod, but you can still tell I've got approximately the same framing if not a bit wider)

    image

    image

    image.jpg
    480 x 320 - 47K
    photo.JPG
    480 x 320 - 48K
  • For those who use the iphone or other gadgets, there's a great app called "pcam", that does a perfect job of this math. We used to carry around the ASC manual before as a reference. Great topic. Thanks

  • but the FOV does change. And for that, shian is wrong to his producer. He will need a 9mm to get an 18mm FOV

  • @Meierhans

    50mm lens has 50mm focal length. It's the optical characteristic of the lens, and it doesn't change whether you use it on FF, Super35, or M43.

  • @Sangye @shian Thx mates, this was really messing up my understanding of focal lenght. ;)

    So to whats the exact crop factor between super 35 and oversized gh2 mft sensor? Ruffly 1:1.25 ???

  • If you have a phone that will run it, or an iPad, get Artemis Viewfinder, if you go full screen, and snap pictures using different sensors, lens brands, etc... you'll see pretty quickly the difference.

    What Sangye is saying is true the 50mm designation is the measurement between the front element of the lens to the focal point where all the rays of light coming through that element converge. On a 50mm lens it is 50mm from front element to the point of convergence.

    DETERMINE_HOW_THE_FOCAL_LENGTH_OF_A_MAGNIFYING_LENS_AFFECTS_ITS_MAGNIFYING_POWER_01.gif
    438 x 284 - 8K
  • @Meierhans: No, 50mm is always 50mm. Crop factor is a useful tool, because it lets you standardize focal lengths. Although 50mm looks very different on a 5D than it does on a GH2, if you know the crop factor you can equivalate the two. 50mm focal length * 2x crop factor = 100mm, meaning that a 50mm lens on a GH2 will have roughly the same perspective as a 100mm lens looks on a 5D. Full frame is definitely an arbitrary term for an arbitrary sensor size, but it is useful for those who have a background in 35mm still photography.

  • @oscillian, this is kinda the 35mm adapter principle all over again isn't it? Project an image from the still lens onto the much smaller sensor while preserving the same image characteristics.

  • It would look like crap or cost a fortune. Have a look at the adapters with a lens for Canon EOS that adapt lenses with a shorter flange distance to make them focus to infinity – serious image degradation!

  • @oscillian I've been wondering why this has not happened yet, and can only assume that it must be far more complicated than I am imagining. I assume if it were possible to "shrink" the image circle, that the effective angle of view would widen back out as well, which would be very useful, actually. Worth spending quite a bit of money on, in fact.

  • So the first company that comes up with a m43 adapter for still frame lenses with optics that shrink the projected image circle will get my money!

  • Here's a practical way to look at crop factor and DOF:

    "Crop Factor" is a term created for still photographers who were used to shooting on 35mm still film, to explain the relative size of the then-new-fangled digital sensors. Crop factor made it easier for them to calculate what to expect from the new digital cameras with various sensor sizes. This is also where the term "Full Frame" comes from. Still photography.

    But 35mm still cameras are not movie cameras. 35mm still cameras run film horizontally, and 35mm cinema cameras run film vertically. The imaging area of 35mm still film is almost twice the size of 35mm cine. #5mm still film is called "65mm" or "Vista Vision" in the cine world, and is considered "large format" like IMAX.

    If you want to compare cine cameras, don't use 35mm Crop Factor as a guide. Instead, use Super35. It is most like cinema 35mm film. Some people call this CineCrop.

    So the closer your camera's sensor is to Super35, the more the AOV and DOF will resemble 35mm movie film.

    A word about DOF: Yes, DOF is a function of the lens. For practical purposes, here's how sensor size affects it. If you select a shot with a particular AOV on a large-sensor camera, and set up the same shot on a small sensor camera, the lens and camera position required to match the shot will result in the small sensor camera showing greater depth of field. The DOF of the lens doesn't change, but the adjustments you make to create the same shot result in different DOF.

  • Well since the D800 was singled out earlier - let's not forget in full HD recording there is a slight crop factor:

    FX.png
    584 x 589 - 159K
  • @stonebat "Two lenses having same focal length might cover different image circles. The difference in image circles makes them having different FOV." So basicly you are telling me that two lenses with the same focal lenght - like one 50mm canon designed for FF and a 50 mm lens nativly designed for MFT will have a different FOV when used on the same cam (preferably mft in this case)? So the "output angle" inside the cam of these lenses are different, and the lens made for full frame would really get cropped on mft, resulting in overall longer focal length? If I want to know what FOV I will see I would have to know not only the focal lenght of a lens, but also what sensor it was made for. Now this really starts to confuse me...

  • @Vitaliy_Kiselev

    Lol. I'd love to, but too bad idk how to mess w/ that shit. But that is what this site needs :-) Someone make it happen!

  • @Macalincag

    Btw, if you want to help, you can look at page source and make simple JS to redirect search to site specific google. I'll insert it in the page. :-)