Are traditional exposure calculations meaningful?

Discussion in 'Open Discussion' started by WestOkid, May 18, 2014.

  1. WestOkid

    WestOkid TalkEmount All-Pro

    Jan 25, 2014
    New Jersey, USA
    Gary
    Interesting article/video. It discusses the relationship between exposure calculations and sensor size. It attacks some traditional measurements like ISO as being obsolete in the digital age and questions the use of it by manufacturers. Should there be new standards geared to our digital age? I would be curious to hear some opinions.

    http://petapixel.com/2014/05/18/are...y-not-calculating-the-sensor-size-into-specs/

    Sent from my XT1053 using Tapatalk
     
    • Like Like x 1
  2. christian

    christian TalkEmount Veteran

    447
    Apr 12, 2014
    Boston MA
    Well, it's sort of a mathematical debate. And it seems a bit biased: "it’s well worth noting that he specifically calls out Panasonic, Sony, Olympus and Fuji as the main perpetrators, while giving a bit of praise to Canon, Nikon and Sigma for being the honest “Good Guys.”

    He forgot that Sony is the supplier for Nikon's sensors...


    Envoyé de mon iPhone à l'aide de TalkEmount
     
  3. Poki

    Poki TalkEmount Hall of Famer

    Aug 30, 2011
    Austria
    ISO is, mathematically and physically, a quite good measurement for light sensitivity of photosensitive surfaces, no matter if they are digital or analog. With digital, the signal amplification brings in some new variables, of course. So the "problem" is that ISO 100 in a Sony camera is not the same sensitivity as ISO 100 in a Canon camera. Having that said, the difference never really goes above 30%, and there are reasons why these differences exist.

    Some reasons are stupid, like showing ISO 100 when it's actually at ISO 80. That way, manufacturers get a 20% better result in most low light tests (DxO corrects for these variances). Some reasons are good, too. Like in Sony SLT cameras, where the mirror eats about a third of a stop. So in order to compensate, Sony made the numerical ISO 100 to an actual ISO 130.

    None of this makes ISO obsolete in any way. And everybody knows what ISO is, how it works and how to handle it. Introducing a new measurement system, which could theoretically correct for manufacturer and amplification differences, would spread more chaos and confusion than it would get rid of.
     
    • Like Like x 3
  4. WestOkid

    WestOkid TalkEmount All-Pro

    Jan 25, 2014
    New Jersey, USA
    Gary
    I agree that would be the case initially. How long? I can't say, but you yourself point out a few differences in how manufacturers treat ISO. How about a supplement instead of a replacement? Wouldn't it be nice to have something of an universal equalizer? It would be nice to be able to look at the ISO for any given camera and have an idea of what type of signal-to-noise you will get.



    Sent from my XT1053 using Tapatalk
     
  5. Poki

    Poki TalkEmount Hall of Famer

    Aug 30, 2011
    Austria
    I don't think it's necessary. A difference of 30% is small enough so that you can compare exposure values between different systems, and the difference does not introduce any problems while shooting because the in-camera metering systems, either in an automatic mode or via the histogram, are accurate enough.
     
  6. MAubrey

    MAubrey TalkEmount Top Veteran

    DxO 'corrects' for it, but not really. Their definition of ISO diverges from the actual international standard. What they're measuring is essentially photon 'well-depth,' which has nothing to do with ISO as a defined standard.
     
  7. Bill

    Bill TalkEmount Veteran

    339
    Oct 22, 2012
    Brisbane, Australia
    Bill
    Sure, crop factors are more complex than just the multipliers (1.5, 2.0, etc). But it's not possible to attach a 40 minute video to every camera ad, and some new system would be even more confusing. If there's culpability in all this, it rests with the companies (and more than a few "experts") that promote full-frame DSLRs with the idea that expensive cameras make either good photographs or good photographers.

    I'm guessing that anyone who has a 70-200 f/2.8 to attach to their camera body understands crop factors; and most (but certainly not all) who only have the kit zoom won't notice.
     
  8. Poki

    Poki TalkEmount Hall of Famer

    Aug 30, 2011
    Austria
    Crop factors are only there for one reason - to make systems with different sensor sizes easily comparable. That's about the same reason why the ISO rating still exists - it is simple. Why making everything utterly complex when you can explain everything with an accuracy of about 95% so simple? Those guys who need to know the details - i.e. the ones who design the sensors and lenses - will study these things for years anyway.
     
  9. roundball

    roundball TalkEmount Hall of Famer

    Oct 8, 2013
    USA
    Won't even pretend that I have a good grasp of these sorts of details, but just thinking out loud, this questions comes to mind:

    If there is a known offset from actual ISO per manufacturer...ie: 30% (1/3rd)...couldn't 1/3rd exposure compensation be set and left that way to counteract that offset to hold ISO to it's real baseline for a particular brand camera?
     
  10. Poki

    Poki TalkEmount Hall of Famer

    Aug 30, 2011
    Austria
    Yep, you can do that. But it doesn't solve any problems. The camera will always meter correctly (well, it should), so you're effectively over- or underexposing by a third of a stop. Except if you use manual mode, of course. If you are using manual mode, then the problem is that you don't know any "absolute" ISO values. Heck, you probably don't even know any "relative" values between cameras as long as you don't do excessive comparisons. And then the question comes up: Why should you care? Right, you shouldn't. It makes no actual difference for the user. And that's exactly the reason why current measurement systems are more than good enough.
     
  11. roundball

    roundball TalkEmount Hall of Famer

    Oct 8, 2013
    USA
    I'm was referencing the comment that a brand mis-represents their ISO by 30%.
    If all that is just advertising gobbly-gook, and an ISO100 setting on the camera really is still ISO100, then it's a non-issue.

    But if the camera setting of ISO100 is really ISO130, then a 1/3rd exposure compensation would basically offset the 30% back down to a real ISO100 and camera metering should be pretty much on target for ISO100...no PP to compensate as mentioned above.
     
  12. Poki

    Poki TalkEmount Hall of Famer

    Aug 30, 2011
    Austria
    Yep, this could work for camera tests - but the problem of determining just how much the ISO of a camera is off from the reference camera persists. But honestly, I don't see the problem. For consumers, it doesn't make any difference. For camera tests, well, the testers just have to look a little closer, than everything's fine.
     
  13. Bill

    Bill TalkEmount Veteran

    339
    Oct 22, 2012
    Brisbane, Australia
    Bill
    If you want to see the typical ISO measure of your camera, have a look at the DXO Mark sensor data. But it's important to remember that there are other factors such as shutter speed accuracy -- often off by 25 to 30% -- that will affect your final exposure.

    And when your camera analyses the scene, perfect exposure can be quite elusive. Using the histogram can be very helpful.

    Shooting in RAW helps because even if all the elements conspire to shift in on one direction (rather than cancelling one another out), you'll have enough latitude to fix it in post.

    Having an exposure meter (I suggest the Sekonic L-308s) can also be helpful. There are times when an incident reading is invaluable, but it also provides a good check against your equipment's overall exposure performance.