Slashdot Mirror


Beyond Megapixels - Part II

TheTechLounge writes "This is Part II of a series of three editorial articles examining current digital photography hardware, as well as my views of what is to come. In this segment I will be focusing on build, size, weight and ergonomics of camera bodies, as well as the size, weight, function and versatility of the glass strapped to the front of it. If you haven't already, you may want to read Part I first."

14 of 135 comments (clear)

  1. Article is Wrong on Lenses by hanway · · Score: 5, Informative
    The article started out okay, although taking a whole page to say little more than "DSLRs are big", but on the second page I came to this statement as a justification for using smaller lenses in DSLRs:

    The sensor only receives the light that passes through the center of the lens, while the light on the outer region simply falls to the side of the sensor.

    That's fundamentally wrong. A light ray that falls on any part of the lens can be refracted to any point on the focal plane. What gets focused onto the sensor in the center of the focal plane is not just the light that passed through the center of the lens, but part of the light passing through the entire lens.

    The author is right that a range of smaller lenses would help reduce camera size, but with a smaller lens comes less light gathering ability and reduced ability to take advantage of depth of field when composing a photo, so smaller lenses would be a compromise in photo quality.

    1. Re:Article is Wrong on Lenses by dfghjk · · Score: 5, Informative

      The author also missed the point that DSLR's are an attempt to make cameras that best utilize existing lenses. He says over and over that the lenses aren't necessary for the camera but that's meaningless. The lenses come first and the bodies follow. This is the best they can do at the price point and time.

      He also stated that the 2/3" lens and DLSR lens at f/2.8 have the same light gathering ability. That's wrong. They have the same exposing rate ability but the SLR lens has greater light gathering ability because it exposes a larger image circle.

      Frankly, everyone should ignore these articles. The author doesn't know enough about the subject to do anything but damage.

    2. Re:Article is Wrong on Lenses by frdmfghtr · · Score: 4, Informative

      What may have been the point is that while a ray of light at the top of the lens may be focused on the focal plane, it doesn't land on the sensor.

      The sensor only receives the light that passes through the center of the lens, while the light on the outer region simply falls to the side of the sensor.

      This is actually true, due to the nature of focusing a round image from a round lens onto a rectangular sensor (the round plug into the square hole, if you will). Either outer parts of the circle will fall on the focal plane that is not convered by the sensor or the sensor will have areas not exposed to the image (circle fitting entirely inside the rectangle).

      --
      Government's idea of a balanced budget: take money from the right pocket to balance...oh who am I kidding?
    3. Re:Article is Wrong on Lenses by kfg · · Score: 5, Informative

      The author also missed the point that DSLR's are an attempt to make cameras that best utilize existing lenses.

      ". . . by utilizing the interchangeability of the lenses on a DSLR, you open yourself up to the use of dozens of lenses appropriate for all kinds of various uses and prices from around $60 up to, and in excess of, $8,000. For photographers switching from a film SLR to a DSLR of the same brand and mount, this means your investment in lenses does not go out the window."

      KFG

    4. Re:Article is Wrong on Lenses by Zocalo · · Score: 4, Informative
      This is actually true...

      Actually, it's kind of both. You need to think beyond those over simplified line drawings of a pinhole camera with a tree or vertical arrow as the "scene". A circular lens will create a circular image on the focal plane, which is where the sensor or film will hopefully lie, and since the sensor is usually rectangular part of that projection will indeed be "discarded".

      However, for a point in the center of the image, reflected light from that point source will almost certainly be striking the *entire* front element of the lens, and being refracted back onto the sensor where they ideally will focus on a single point again. Instead of a single line from a point on the subject to a point on the sensor, you need to think of two conical objects (yes, 3D) joined at their equal sized bases (the lens).

      The fact is that there are a lot of photographers that don't understand the finer points of optics, or need to for that matter, and are under the illusion that DSLRs are only utilising the superior glass in the center of their lenses. Given that many of them have only just grasped how the field of view crop actually works, and that it's not really the "zoom multiplier" marketing told them it was, I can't say I'm surprised. If anyone knows of a web page that explains this in laypersons terms, I'd certainly appreciate the URL though!

      --
      UNIX? They're not even circumcised! Savages!
    5. Re:Article is Wrong on Lenses by dfghjk · · Score: 5, Informative

      f/2.8 would be a specific rate of exposure per unit of imager area. It would expose at the same rate regardless of the size of the imager. The sensitivity of the imager is rated by ISO and is also a rating per unit area. In this way, cameras with identical ISO ratings and f-stops will require the same shutter speed regardless of imager size. That how you want it to work.

      Light gathering ability of the lens, however, is not per unit area. It means the total ability to gather light. Therefore a lens that covers a wider area at the same f-stop will have greater light gathering ability than its competitor. Whether that larger image circle is actually used is beside the point. The author was incorrect stating that the two lenses have identical light gathering ability. He would have been right had he said "exposure ability".

      Two cameras with different sensor size but identical ISO's and f-stops will require the same shutter speed for proper exposure, but the camera with the larger sensor requires more light to expose due to its larger imager area. Where does this extra light come from? Not from increased exposure time since the shutter speed is the same. It comes from the lens delivering more total light. This occurs because its lens actually has a larger physical aperture to achieve the same f-stop and the larger aperture allows more light through the lens. The "luminous flux" is unchanged, however, because it's spread over a larger area. How does the lens get away with this? The larger sensor area requires a longer focal length lens for equivalent perspective and f-stop = (focal length/physical aperture). It's all cleverly hidden in the math. Perhaps a little too cleverly.

    6. Re:Article is Wrong on Lenses by floateyedumpi · · Score: 4, Informative
      Actually, you're wrong. Imagine a set of mile markers set off in the distance you are imaging. Now imagine you've setup the same lens on a 35mm and DSLR, side by side, to capture the exact same field of view (same number of mile markers). To get the same field of view, you'll need to zoom the 35mm camera in. To get the same f-number, you'll need to open up its pupil aperture to 1.5x as big (it's collecting more light, as someone below points out, because it needs to preserve not total luminous input, but surface brightness per square mm of film). So we have our two cameras taking the exact same picture at the exact same exposure value, but one is zoomed in and has a larger aperture.

      In order to reach the film/sensor, a bundle of rays must pass through the pupil aperture, aka the "iris". Consider the bundle for the mile marker post #123 on the extreme right-hand side of the scene. For the 35mm camera, the aperture is bigger (e.g. 1.5x as big), hence the bundle of rays accepted from that milepost is bigger, hence the actual amount of glass it traverses is bigger. If the aperture and hence bundle of rays is too big, and the field angle at the edge of the sensor is too large, part of that bundle will be shadowed by the lens hood and you'll have vignetting: again not so much of a problem for DSLRs (but a real problem for wide-angle lenses)! This is a real physical effect.

      Yes, the front surface of the lens accepts light from mile post #123 over its entire diameter, but most of those photons are rejected by the aperture stop (i.e. they fall on the iris pupil and never makes it through). For a given FOV, the aperture diameter is *smaller* for DSLRs, simply because the focal length needs to be reduced.

      The bottom line is standard 35mm lenses are overdesigned for DSLR usage, accomodating a larger pupil aperture than needed for a given field of view. Less of the lens really is used, and luckily, all problems of chromatic and spherical aberration increase drastically with field angle: in this case sticking to the "center of the lens" should be thought of the angular sense, and really does improve the image quality. The flip side of this is that DSLRs will be more forgiving when it comes to lens quality, since they don't stress the off-axis performance as much.

      If you don't believe me, take a very wide angle lens on you 35mm film camera, open up the aperture as wide as possible, and take a highly-contrasting scene. Notice how the center of the image is sharper and more color-accurate? Get out your scissors and cut out the central 1/2 of that picture. You've just replicated what a DSLR at the same lens settings would have produced. Nicer looking, eh?

      By the way, the terminology "equivalent focal length" when applied to DSLR lenses is a complete misnomer: see this comment.

  2. Telecentric Lenses and Silicon by G4from128k · · Score: 5, Informative

    The article mentions the excessively large size of 35 mm lens for imaging on to small digital sensors, but misses the two additional problems with using film camera lenses with digital sensors.

    Standard film camera lens tend to transmit light from the subject to the sensor at the angle that it was received (similar to the way that a pinhole camera projects a bundle of rays from object space to image space). Silicon sensors suffer from two problems when light enters them at an angle. First, the high index of the material and coatings tends to reflect the angled light -- causing less light to enter the sensor and the image to have dark corners. Second, long wavelength light penetrates the sensor deeper than does short wavelength light. If the light enters at an angle, the red photons can angle down into the substrate and actualy register in pixels further out. The result is that the red and infrared portions of the image are misregistered, causing color fringing in the corners.

    The point is that the best lens for a digital camera will be different from the best lens for a film camera. A better lens design for digital cameras incorporates image-space telecentricity. Image-space telecentricity means that the light hitting the CCD is largely perpendicular to the sensor.

    --
    Two wrongs don't make a right, but three lefts do.
    1. Re:Telecentric Lenses and Silicon by jdg · · Score: 5, Informative

      I design lenses for a living.

      This statement is sort of correct. The real need for telecentricity is the limited acceptance angle of the lenslet arrays that are put on many sensors, particularly small pixel size consumer grade sensors. This is what causes the drop off in corner illumination with non-telecentric lenses in consumer grade digital cameras. Telecentricity is a real requirement, particularly in sensors which pixel sizes smaller than say 4um.

      Large pixel, higher end CCDs generally don't need the lenslet arrays because the fill factor on the pixel aperture is much larger, so there is not much of a problem with non-telecentric designs. There are no lenslets present to limit the acceptance angle. I have never seen reflections off the sensor be an actual problem. People also talk about ray bundles from one obliquity passing through the wrong filter on the color filter array in a non-telecentric design, but I have never seen this happen either.

      Longer wavelength light is generally eliminated by an infrared reject filter, so light rarely bleeds from pixel to pixel.

  3. dpreview.com by Seequeue · · Score: 4, Informative

    Go look at http://dpreview.com for as much detailed, objective information regarding digital cameras as you're likely to want.

  4. Re:35mm by sdr · · Score: 5, Informative
    There are two digital SLR cameras in market right now with full 35mm sensors - the Canon EOS-D1s and the Kodak DCS SLR/n. The Canon is an 11 MP camera and costs about $8000.00. The Kodak is a 14MP camera costing about $5000.00. The Canon produces the best quality pictures among digital SLRs. The Kodak is rather new. It is actually an update of an older camera using the same sensor called DCS 14n. This older model has been described as rather noisy (in comparison to the Canon at least) in all but the lowest ISO setting. It also produced bad colors in certain situations. The new model supposed to be better - but not in the same class as the Canon. The Kodak uses Nikon lenses and is based on a Nikon camera body. Kodak has also announced another body with the same sensors that take Canon lenses.

    So at least two full 35mm frame digital SLRs exist. None of them are cheap - and it it will quite possibly stay that way for some time.

    Leica has announced a digital back for their R series SLR cameras. This being Leica, it will possibly be rather expensive - not to mention the huge price tags for their lenses and film bodies. There has been some persistent rumors that Nikon is designing their next professional flagship SLR camera body (the successor to the film SLR model F5) as a camera that can take interchangable digital (and film) backs. Nikon flagship models are usually replaced every 8 years. If the pattern holds then they should come up with a new model (the F6) this year.

  5. Re:35mm by herko_cl · · Score: 4, Informative

    The Canon EOS 1Ds has a full 35mm frame, 11 Megapixel sensor. The Kodak DSC 14n accepts Nikkor lenses (for Nikon mounts) and has a full 35mm frame, 14 megapixel sensor (both links are reviews)
    The Kodak costs approx. $ 4000, and the Canon $ 7500. But at least 35mm, full frame sensors ARE here already. If you win the lottery, you can also buy medium format digital backs.

    --
    No .sig for you! ONE YEAR!
  6. Sample rules by jsse · · Score: 4, Informative

    Here is my little opinions on the subject. Not surprising I share the same view with the author.(My apology for the length and inaccurate technical details)

    Lens
    I agree with the author that the lens pays an important part of the overall quality, rather than no. of pixels. Generally speaking, lens with large aperture(F2.8>F4>F5.6>F8, etc.) can create better images. However, to compensate for the distortion near the edge, the larger the aperture, the bigger the lens size. You'd find digital camera with bigger lens(usually implies bigger aperture) cost more, regardless of no. of pixels.
    While it's true that camera with exchangeable lens is very desirable for photographers especially when you already has a good lens. However, I do not think the high price of those lens-exhangeable digital camera, especially Nikon D70, is justified(I'm a diehard Nikon film amatuer photographer myself). If you don't like those digital camera exchangeble lens, you may look at those already has good lens equipped, like Lumix DMC-LC1, which equipped with a F2.0 Vario-Summicron Lens, a legendary brand name for most film photograpers. (Mind you, some perfectists critize that the lensare not made in their original factory. Oh well.. :)

    Color
    The article touchs this topic very lightly, in fact most digital camera manufactuers avoid this. You can imagine different wave in light spectrum refract in different angle in each piece of lens. The problem is particular complicated when the lens group has more than one lens. That's why lens with more lens group is more expensive. This problem is called the chromatic abberation.

    Aspherical lens(glasses with uneven density) and coating could help solving this problem. You can see the color reflect from the surface of many professional lens are not white - usually redish or slight greenish. The less white light reflects from the lens' surface, the better the coating. (This is in fact one tip you can use in choosing a good digital camera)

    Light
    As implied in the word 'photographing', it's all about light(photo). The better the lighting condition, the better the images created - this is true for digital and film photographing. You can't control the light, but you can control how light enter the camera. Most digital camera owners would find that regardless of no. of pixels, the images quality drops drastically in low light condition.

    Guess what I'd say - yes, bigger(and high quality) lens invite more light thus create better images. What's so difficult to understand. XD

    Conclusion

    The quality of the lens outweights the no. of pixels. Well, in fact this is a most unwelcome answer, and people stop asking me for opinion on choosing digital camera, and go buy some fancy looking garblish. Luckily we've slashdot where I can find people still listening to me.....hello? HELLO???......

  7. Re:What is this article trying to say? by carlislematthew · · Score: 5, Informative
    It seems that Nikon are avoiding the 35mm size sensor more than anyone else. They've come out with these crippled "digital" (or whatever) lenses specifically for their DSLR range.

    Right now, Canon actually *has* a 35mm sensor DSLR (EOS-1Ds) - it's supposed to be awesome, as well as being awesomely expensive ($9,000ish I believe). From what I've read, the problem is the low yield on making the sensors themselves and also some fancy expensive anti-aliasing filter that goes in front of the sensor.

    Unfortunately, I don't think you can compare yield improvement of expensive 35mm 12MP sensors with yield improvement (and therefore cost reduction) on things like LCD flat panels. The reason is that consumers don't *need* image quality like the Canon EOS-1Ds provides. It's almost medium format quality and 99% of consumers used crappy tiny-lensed 35mm negative film for years, printed by shitty machines on 4x6 paper that fades.

    So if it *is* the case that 35mm sensors are the future for DSLRs, I do not believe we can expect the kind of quick generational reduction in cost that we're used to for other more "commodity" consumer items like LCD flat panels, PDAs, cell phones, and so on.