(Effective) pixels?

Discussion in 'Digital Photography' started by Gregory L. Hansen, Dec 13, 2005.

  1. When a camera is advertised as having, say, 5 MP (effective), what does it
    mean for them to be (effective), and why is (effective) in parentheses?
     
    Gregory L. Hansen, Dec 13, 2005
    #1
    1. Advertisements

  2. Gregory L. Hansen

    default Guest

    That would be how many pixels are in the final image. The sensor has more
    pixels that this. Most sensors have some pixels that are optically shielded
    (under the metal frame) that are used to measure the dark current for that
    row for correcting the black level. There may be some that are partially
    lit from being very close to the frame. These are also cropped off as they
    have imaging problems. Finally the sensor may have a few more imaging
    pixels than are used if the aspect ratio is not exactly what the
    manufacturer wanted or for use in "de-bayerizing" the edges. These might be
    cropped out also to produce the image size that they wanted and not have
    colour distortions at the edge of the picture if they want to use an
    algorithm that doesn't handle edges correctly.

    All of this adds up to an image that has fewer pixels than the sensor. You
    are probably not missing more than a couple of hundred thousand pixels
    compared to the full sensor array count. If your image included everything,
    there would be a border around the edge that starts with soft focus (from
    the frame edge and diffraction), the progresses into a light band (but no
    image), fading to black at the outside edges. So they crop off the bad
    parts and give you the effective size so you don't wonder why your 5.2 mp
    sensor only gives 5.0 mp images.
     
    default, Dec 13, 2005
    #2
    1. Advertisements

  3. The sensor has *more* pixels? Good for technology! I had assumed that
    (effective) pixels was marketing speak for fewer pixels on the sensor but
    making it up somehow in the image.

     
    Gregory L. Hansen, Dec 13, 2005
    #3
  4. It is marketing speak in some cases. A prime example is the Bell & Howell
    camera ads seen in some magazines. Those extra pixels are interpolated.
     
    Charles Schuler, Dec 13, 2005
    #4
  5. Darn, I was hoping for a simple answer.

    What about Canon's new camera with the full-frame sensor for $3000? That
    seems like too high a price tag for stupid marketing speak.
     
    Gregory L. Hansen, Dec 14, 2005
    #5
  6. Gregory L. Hansen

    Skip M Guest

    That isn't marketing speak, it is a full 12.8mp, 35mm sized (24mmx26mm
    approx) sensor.
     
    Skip M, Dec 14, 2005
    #6
  7. Each photosite (sensor) on a Bayer sensor determines what color it's
    supposed to be from the surrounding sensors. For instance, if the red
    sensor detects it's fully activated and the surrounding blue and green
    sensors are too, then the red one knows it's supposed to be white. If the
    blue ones are inactive (only red and green active), then it knows it should
    be yellow...and so on.
    Well, the photosites around the perimeter of the whole sensor don't have
    surrounding pixels from which to determine their color. They have to be
    discarded because of that. They're only used so that the inner pixels can
    determine their color. So, for a sensor, say, 2400 x 3600 there's 12,000
    photosites that don't contribute to the final image.
     
    William Oertell, Dec 14, 2005
    #7
  8. In the early days of digital cameras, manufacturers sometimes quoted the
    total number of pixels fabricated on the sensor, including ones that did
    not appear in the final output image (for whatever reason). So you
    would get an image with 1.2 million pixels from a "1.3 megapixel"
    camera.

    Then someone established the standard that "effective pixels" are ones
    that actually contribute to the image. So, for most cameras, the number
    of effective pixels is a fair way to compare them.

    Dave
     
    Dave Martindale, Dec 14, 2005
    #8
  9. Gregory L. Hansen

    Tony Guest

    Tony, Dec 14, 2005
    #9
  10. Gregory L. Hansen

    cjcampbell Guest

    Fuji has been known to do that, too.
     
    cjcampbell, Dec 14, 2005
    #10
  11. Gregory L. Hansen

    Ron Hunter Guest

    There are two reasons for using the 'effective pixels' spec. One is
    that many cameras have more actual pixels from the sensor than they
    actually use and keep for the picture size selected. Mine, for
    instance, will capture photos in either 3:2 or 4:3 aspect ratio, and
    delivers slightly less than 4 mp in the 3:2 mode, and slightly more in
    the 4:3 mode. This is an advantage.

    The other reason for using 'effective pixels' is a marketing ploy to
    make a camera seem better than it is by 'interpolating' extra pixels to
    make the image SEEM better than it is. You should always try to
    distinguish between these two approaches, especially in an inexpensive
    camera.
     
    Ron Hunter, Dec 14, 2005
    #11
  12. Gregory L. Hansen

    Ron Hunter Guest

    In the case of the Canon camera, I am sure it is the first reason I
    mentioned, rather than interpolation. I doubt any Canon camera goes the
    interpolation route for advertising purposes.
     
    Ron Hunter, Dec 14, 2005
    #12
  13. Gregory L. Hansen

    spudnuty Guest

    Fuji has been known to do that, too.
    Like in a F700 I just worked on. It's advertised as a 6 MP camera but
    if you read the fine print it's actually 2 3 MP sensors working
    together in an attempt to improve dynamic range. The actual resolution
    looked to me like that of a 3 MP camera.
     
    spudnuty, Dec 14, 2005
    #13
  14. Gregory L. Hansen

    Skip M Guest

    Hardly.
     
    Skip M, Dec 14, 2005
    #14
  15. How can they be distinguished? I don't think I've seen that data printed
    on packaging.
     
    Gregory L. Hansen, Dec 15, 2005
    #15
  16. Y'all could just read the specs for a sensor.

    http://www.sony.net/Products/SC-HP/cx_news/vol42/pdf/icx624series.pdf

    And note that there are four pixel counts. And then Google around to find
    out what they mean. The last I checked, there were lots of articles out
    there.

    David J. Littleboy
    Tokyo, Japan
     
    David J. Littleboy, Dec 15, 2005
    #16
  17. Gregory L. Hansen

    cjcampbell Guest

    Fuji actually brags about their interpolation scheme in their marketing
    material and it is written all over the box. And with good reason.
    Fuji's point is that wide dynamic range is better than a large number
    of pixels. For many purposes it is. Doubling the number of pixels does
    not double your resolution; it only increases it by about 25%.
    Increasing dynamic range, though, makes the picture appear much sharper
    than it really is, without the telltale white outlines of sharpening
    algorithms.

    A lot of guys don't like Fuji's approach, partly because it undermines
    the "more pixels is better" marketing of everybody else. Fuji's real
    trouble is their cameras have a history of being unreasonably fragile
    and some of them also have a lot of digital noise. It would be
    interesting to see what a manufacturer of more robust cameras could do
    with Fuji's technology.
     
    cjcampbell, Dec 16, 2005
    #17
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.