ISO and actual sensitivity in DSLR's (D70, *istD, 20D, S3...)

Discussion in 'Digital Photography' started by Alan Browne, Mar 25, 2005.

  1. Alan Browne

    Alan Browne Guest

    We've seen the occasional postings about the ISO setting and the actual
    sensitivity.

    I picked up Casseur D'Images (No. 271, March 2005) and among other
    things they did tests on a variety of DSLR's and one ZLR. (p. 169)

    They rounded the results when close to the standard 1/3. But where a
    little less clear cut, they put a +/- to indicate not quite in the 1/3 zone.

    They describe, in punishing detail, the test method, references, math,
    etc. Unfortunately, the 7D is not part of the grouping, but I'll take
    sollace in the A200 results. Minolta have long been known for their
    consistency in metering and exposure.

    [There is also an article on the S3 and I'll summarize tomorrow]

    Cheers,
    Alan.

    D70:
    Setting: 200 400 800 1600
    Actual: 160 320 640 1250


    A200 (Minolta ZLR):
    Setting: 50 100 200 400 800
    Actual: 50+ 100 200 400 800


    *istD:
    Setting: 200 400 800 1600 3200
    Actual: 250 500 1250- 2000 4000-


    20D:
    Setting: 100 200 400 800 1600 3200
    Actual: 125 250 500 1000 2000 4000


    S3:
    Setting: 100 160 200 400 800 1600
    Actual: 80 160 160+ 320 640+ 1250


    1D Mk II
    Setting: (L)50 100 200 400 800 1600 (H) 3200
    Actual 64 160 320 640 1250 2500 4000
     
    Alan Browne, Mar 25, 2005
    #1
    1. Advertisements

  2. Alan Browne

    Alan Browne Guest

    Chasseur D'Images
     
    Alan Browne, Mar 25, 2005
    #2
    1. Advertisements

  3. Alan Browne

    Pete D Guest

    And as has been said before, it does not matter one tiny little jot as long
    as the photos come out right.
     
    Pete D, Mar 25, 2005
    #3
  4. The table shouldn't be interpreted as good or bad, but informational.
    Built-in exposure meters can't be used for all conditions.
     
    Kevin McMurtrie, Mar 25, 2005
    #4
  5. You are not of the old school I see :)

    It does matter.
    - if you use an external meter.
    - when comparing the sensitivity of camera systems.
    - when using external flash.
    - etc

    But - if you just take pictures and like them - then
    the technicalities behind the making of the photo is
    of course uninteresting.

    But - it does matter - even if you don't care :)


    /Roland
     
    Roland Karlsson, Mar 25, 2005
    #5
  6. Alan Browne

    Owamanga Guest

    Okay, but whether it be by table or experience, if you find your
    camera to consistently meter under or over, you'd just adjust the
    exposure accordingly wouldn't you?

    ....annoying if you switch between brands all the time, but for a
    single DSLR body owner, no biggie.

    And the report (or Alan) missed another significant angle here:

    The D70 they used, is it the same as my D70? What's the consistency of
    metering within each model like? If it's bad for any particular model,
    their table becomes fairly irrelevant.
     
    Owamanga, Mar 25, 2005
    #6
  7. Alan Browne

    Alan Browne Guest

    For those who use an incident meter or seperate spot meters (including
    myself), such information is useful.
     
    Alan Browne, Mar 25, 2005
    #7
  8. Alan Browne

    Alan Browne Guest

    I do that.
    One of the roles of this NG.
    I do that too.
    er, this is an equipment group. This *is* one of the things we discuss
    here.
     
    Alan Browne, Mar 25, 2005
    #8
  9. Alan Browne

    Alan Browne Guest

    They don't mention if they're using samples of one or more.

    If one were talking about lens variations (sharpness) I would agree.
    But metering in electronic cameras and shutter speeds have become quite
    precise and consistent over the past 10+ years.
     
    Alan Browne, Mar 25, 2005
    #9
  10. Not really. You have to calibrate your meter against the histograms that
    result in your camera.

    And I suspect that the results are simply random, since presumably they used
    matrix/evaluative metering which is, in principle, completely random and
    unpredictable (assuming it does what they say it does<g>).

    David J. Littleboy
    Tokyo, Japan
     
    David J. Littleboy, Mar 25, 2005
    #10
  11. Alan Browne

    Owamanga Guest

    So, the database of 30,000 images is a big lie, it's just a
    pseudo-random number generator.

    That actually explains a lot.

    I've done some tests with the D70, if my daughter (caucasian), pulls
    her face to make her eyes to look more asian, the matrix metering is a
    lot more accurate. I guess she suddenly matches some of the faces in
    the database.

    <g>
     
    Owamanga, Mar 25, 2005
    #11
  12. What does the sensitivity of a sensor have to do with the built-in lightmeter
    of a camera?
     
    Philip Homburg, Mar 25, 2005
    #12
  13. Alan Browne

    Owamanga Guest

    Are you talking histograms here? One assumes the manufacturer is
    completely aware of the exact sensitivity of the sensor, and so takes
    this into account when drawing the histogram.

    So, compare external meter reading's suggestion of exposure values for
    a given ISO to the resulting histogram at those same exposure values
    to see if an adjustment to the meter's ISO setting is required to make
    it 'accurate' for that sensor. Personally, I wouldn't trust my
    interpretation of the histogram to not introduce yet another error
    factor using this method.
     
    Owamanga, Mar 25, 2005
    #13
  14. Alan - I just tried to be nice to Pete.

    Personally I fully agree with you. I just acknowledged
    that not everyone do. And that is OK IMHO.


    /Roland
     
    Roland Karlsson, Mar 25, 2005
    #14
  15. Alan Browne

    Alan Browne Guest

    Both groups are equipment related. I apologize for my use of the
    singular above and trust that the flogging, while justly deserved, will
    be administered mercifully.
     
    Alan Browne, Mar 25, 2005
    #15
  16. Alan Browne

    Alan Browne Guest

    Appolgies, I thought you were replying to me. My news reader does a bad
    job of organizing refreshes.

    Cheers,
    Alan
     
    Alan Browne, Mar 25, 2005
    #16
  17. Alan Browne

    McLeod Guest

    And if you have time to incident meter, you probably have time to look
    at the histogram and adjust accordingly, as well.
     
    McLeod, Mar 25, 2005
    #17
  18. I think you can also show a histogram if you are completely unaware of
    the effective sensitivity of the sensor.

    The histogram is supposed to reflect the distribution of the RGB values
    on an image. How the image was obtained is not revelevant for the histogram
    (though the color space might be if you want to show brightness)
     
    Philip Homburg, Mar 25, 2005
    #18
  19. Alan Browne

    Alan Browne Guest

    Quicker to set strobes with an incident flash meter. Though I am
    developing a worrisome, guilt inducing chimping habit when outdoors with
    the D7.

    Cheers,
    Alan.
     
    Alan Browne, Mar 25, 2005
    #19
  20. Alan Browne

    Jan Böhme Guest

    No need for correction. It was much more fun as it read first.

    Jan Böhme
    Korrekta personuppgifter är att betrakta som journalistik.
    Felaktigheter utgör naturligtvis skönlitteratur.
     
    Jan Böhme, Mar 25, 2005
    #20
    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.