ISO speed determination constants - question for Stephen, Bill etc

Discussion in 'Exposure Discussion' started by michael_r, Feb 27, 2013.

  1. michael_r

    michael_r Subscriber

    Messages:
    6,539
    Joined:
    Feb 2, 2010
    Location:
    Montreal, Canada
    Shooter:
    Multi Format
    How was the delta log H of 1.3 decided upon in the ISO "triangle"? I guess I could ask a similar question for the 0.8 delta D. Is there a relationship to the fractional gradient method that led to these numbers?

    I'm also trying to understand why there was a need to move to this system versus the fractional gradient. Henry (granted, not the best discussion of speed determination methods) writes the fractional gradient method was difficult to use and also that it couldn't gain international acceptance, which led to the ISO method. Obviously this is an oversimplification, but I don't understand why the ISO method is easier to use, other than it having only one gradient. So what?

    I guess the direction I'm going in here, is in theory I think I prefer a fractional gradient determination. Not sure yet. Is it possible ISO was preferable because it allows films with different toe shapes to have the same speed, which might have satisfied more manufacturers?

    For me, film speed or an EI is merely a means to ensuring I have good toe contrast. So the thing that is bothering me here is that there are too many ways to satisfy the ISO method. A short and long toe film can both have the same ISO speed.
     
  2. Stephen Benskin

    Stephen Benskin Member

    Messages:
    1,646
    Joined:
    Jan 7, 2005
    Location:
    Los Angeles
    Shooter:
    4x5 Format
    Michael, as you know this is one of my favorite subjects. You may come to regret asking.
     
  3. michael_r

    michael_r Subscriber

    Messages:
    6,539
    Joined:
    Feb 2, 2010
    Location:
    Montreal, Canada
    Shooter:
    Multi Format
    I certainly won't regret asking :D, I can assure you. I find these discussions interesting and informative.

    I'm trying to follow the path from fractional gradients to ISO (.1 above B+F satisfying the .8 delta D:1.3 delta log H relationship), to ZS testing etc. and figure out why ISO speeds don't "work" for me.
     
  4. Bill Burk

    Bill Burk Subscriber

    Messages:
    4,984
    Joined:
    Feb 9, 2010
    Shooter:
    4x5 Format
    My idea - and it may not be right, it just sounds easy to comprehend...

    I think it all comes down to how hard it is to "find" the tangent of the curve with a line that is 0.3 times the average gradient. If that were easy to find, we'd have picked that point long ago and been done with it.

    That 0.3 Gradient point is the "real" best place to determine speed. It corresponds with the studies of people saying which print they liked best.

    With everyone in the labs wanting a fixed point to look for, and where the curve meets 0.1 is easy to find, 0.1 became a popular point to look at.

    So the ASA triangle was drawn up because when you hit the triangle - then there is a direct relationship between 0.1 (that is easy to find), and 0.3 Gradient (that you want but which is hard to find). So they made finding the point easy. In case there are philosophical disagreements between those who liked 0.1 and those who liked 0.3 Gradient... the argument is rendered moot because now you can look at it either way and get the same outcome. Actually a really good way to end an argument -- make everyone happy.

    I think it also has something to do with being a good gradient to fit average subjects on paper.
     
  5. michael_r

    michael_r Subscriber

    Messages:
    6,539
    Joined:
    Feb 2, 2010
    Location:
    Montreal, Canada
    Shooter:
    Multi Format
    Do you really get the same outcome either way though?

    With the fractional gradient, you need a decent gradient right at the bottom. With ISO, you only need a gradient over a 4 1/3 stop range. So I'm thinking about two hypothetical 400 ISO speed films. One with a longer flatter toe might not satisfy the fractional gradient and/or "excellent print" criteria.

    So what is the relationship, mathematical or otherwise, between the two methods (eg how did they settle on 1.3 delta log H? in the ISO method)

    By the way I'm not necessarily agreeing with 0.3G either. It's too low for me. What I'm ultimately getting at (I think) is that if we're using a gradient in some way to determine speed in either case (which I agree with because I'm all about contrast), I would prefer an ISO method that had a shorter delta log H, or perhaps something higher than 0.1 net D for a speed point.
     
  6. Stephen Benskin

    Stephen Benskin Member

    Messages:
    1,646
    Joined:
    Jan 7, 2005
    Location:
    Los Angeles
    Shooter:
    4x5 Format
    Michael, I’m not able to address your questions tonight to the degree in which I’d like, so this should get you started. Pretty much everything you want to know is in these two papers. Nelson and Simonds' Simple Methods for Approximating the Fractional Gradient Speeds of Photographic Materials (which is too big to upload. Contact me and I'll email it to you.) and my Delta-X Criterion (which can be uploaded).
     

    Attached Files:

    Last edited by a moderator: Feb 27, 2013
  7. michael_r

    michael_r Subscriber

    Messages:
    6,539
    Joined:
    Feb 2, 2010
    Location:
    Montreal, Canada
    Shooter:
    Multi Format
    Thanks, Stephen - I'm going to read the Delta X paper and see what happens.
     
  8. AndreasT

    AndreasT Member

    Messages:
    355
    Joined:
    Sep 15, 2006
    Location:
    Berlin
    Shooter:
    Multi Format
    Ooh this is fun.
     
  9. michael_r

    michael_r Subscriber

    Messages:
    6,539
    Joined:
    Feb 2, 2010
    Location:
    Montreal, Canada
    Shooter:
    Multi Format
    Andreas, I'm definitely out of my league when it comes to discussing exposure theory, tone reproduction etc with Stephen and Bill, but I still try my best because I've learnt a lot in the Exposure forum. After all there are only so many threads I can read about magic developers and people trying to push Tri-X to EI 25,000. :smile:
     
  10. Stephen Benskin

    Stephen Benskin Member

    Messages:
    1,646
    Joined:
    Jan 7, 2005
    Location:
    Los Angeles
    Shooter:
    4x5 Format
    Only someone with a good deal of knowledge on the subject is able to ask this kind of question. At it's core, this is a question about the nature of film speed.
     
    Last edited by a moderator: Feb 28, 2013
  11. AndreasT

    AndreasT Member

    Messages:
    355
    Joined:
    Sep 15, 2006
    Location:
    Berlin
    Shooter:
    Multi Format
    Michael I enjoy these threats and post of you Stephen and Bill including a few others. I have learnt a lot more, get confused very often. This certainly helps to understand all this better.
    Exposure is the most important single factor in photography in my opionion and it is discussed too little.
     
  12. AndreasT

    AndreasT Member

    Messages:
    355
    Joined:
    Sep 15, 2006
    Location:
    Berlin
    Shooter:
    Multi Format
    Am I correct or incorrect in say that the Delta-X criterion is more similar to th CI from Kodak. Looking at curves near the foot it seems to make sense regarding film speed. Now am I over ambitious in saying when I get my EI there is no need to change the EI when pushing or pulling?
     
  13. Stephen Benskin

    Stephen Benskin Member

    Messages:
    1,646
    Joined:
    Jan 7, 2005
    Location:
    Los Angeles
    Shooter:
    4x5 Format
    The lower gradient part of CI uses the same basic theory as the fractional gradient method. Long toed films are where the influence is greatest.

    Film speeds don't change much with fraction gradient / Delta-X. The pre 1960 standard didn't have an aim contrast for film development (it did have a minimum). While the Delta-X Criterion uses a fixed density point, the Delta-X point moves in relation to the fixed point as the average contrast changes. This tends to compensate for the shifting of the fixed density point.
     
    Last edited by a moderator: Feb 28, 2013
  14. Sponsored Ad
  15. Rafal Lukawiecki

    Rafal Lukawiecki Subscriber

    Messages:
    809
    Joined:
    Feb 23, 2006
    Location:
    Co. Wicklow,
    Shooter:
    Multi Format
    I am excited to see the beginning of a (I hope) another illuminating chapter in the saga of the variables of the tone reproduction cycle and their relationships. Count me in to lurk over you. Stephen, Bill, Michael, and others—thank you.
     
  16. Stephen Benskin

    Stephen Benskin Member

    Messages:
    1,646
    Joined:
    Jan 7, 2005
    Location:
    Los Angeles
    Shooter:
    4x5 Format
    Here's a comparison of speeds derived from the fixed density method and the Delta-X Criterion.

    Relationship Between Fixed Density Speeds and Delta X Speeds.jpg

    Except this is not as much a comparison between two methods as it is between an accurate method, Delta-X, and the erroneous and inaccurate use of another.

    This is why the Delta-X speeds change little with changes in development.

    Comparisons of Delta X at different Delta Ds.jpg

    As the gradient changes with development and the fixed density point shifts left and right, the fractional gradient point's relationship to the fixed density point changes but it's log-H remains relatively constant.
     
    Last edited by a moderator: Mar 1, 2013
  17. henry finley

    henry finley Member

    Messages:
    302
    Joined:
    Dec 26, 2012
    Location:
    Marshville N
    Shooter:
    Medium Format
    You guys could start your own film company. More power to ya. I like reading after people who are smarter than I am, and I'm no mental midget. It's all I can do to be a competent repairman/calibrator of the things guys like you engineer and use for your test instruments. Carry on, men.
     
  18. ic-racer

    ic-racer Member

    Messages:
    7,473
    Joined:
    Feb 25, 2007
    Location:
    Midwest USA
    Shooter:
    Multi Format
    Whoa, you can't start a thread about fractional gradient approximations without discussing "W speed"...:smile:
     

    Attached Files:

    Last edited by a moderator: Mar 1, 2013
  19. Stephen Benskin

    Stephen Benskin Member

    Messages:
    1,646
    Joined:
    Jan 7, 2005
    Location:
    Los Angeles
    Shooter:
    4x5 Format
    Delta-X was adopted for the ISO standard. W speed is more of a historic footnote for being the runner up. As the paper says, W speed "is of particular interest in the mathematical theory of the shape of the sensitometric curve and its relation to speed." It's worth knowing for that, but it has never been used as a film speed method. Seriously, film speed can be a confusing topic and W speed will only add to the confusion unnecessarily.
     
    Last edited by a moderator: Mar 1, 2013
  20. Bill Burk

    Bill Burk Subscriber

    Messages:
    4,984
    Joined:
    Feb 9, 2010
    Shooter:
    4x5 Format
    My take-away from the W speed article that these different speed ratings are compared based on how well they correlate to the "Picture Tests".

    Then the writer makes an outlandish statement that I will challenge: "It is, of course, too time-consuming to use the print-judgment method."

    "Of course"? "Too time-consuming"?

    Look at how much time we spend talking about film speed testing.

    It might be fair to say that is more time-consuming than making some prints and showing them to your friends.

    I'm not suggesting to throw out the tests... Just providing some necessary perspective.

    I still like to test to ASA triangle and 0.1 density. Then I accept Delta-X Criterion as justification that I can use whatever speed I discover when I get as close as I can to ASA parameters... And I use that speed for all my varied development times (N-1, N and N+1 etc.), because it is easier to pick one speed than use a sliding EI... and because it is probably correct.
     
  21. Stephen Benskin

    Stephen Benskin Member

    Messages:
    1,646
    Joined:
    Jan 7, 2005
    Location:
    Los Angeles
    Shooter:
    4x5 Format
    The ring-around test can probably be thought of as a very stripped down personalized version of the first excellent print test.

    I have to agree with Nelson. It took 18 months to produce the negatives for the first excellent print test.
     
    Last edited by a moderator: Mar 2, 2013
  22. Bill Burk

    Bill Burk Subscriber

    Messages:
    4,984
    Joined:
    Feb 9, 2010
    Shooter:
    4x5 Format
    I would have enjoyed watching Nelson and Jones at work.

    I meant to emphasize that my take-away was...

    The ultimate speed criterion is the Picture Tests themselves.

    Anything else is an attempt to correlate to the Picture Tests.

    As you like to say, Stephen, hiding in plain sight.
     
  23. ic-racer

    ic-racer Member

    Messages:
    7,473
    Joined:
    Feb 25, 2007
    Location:
    Midwest USA
    Shooter:
    Multi Format
    Isn't that what this thread is about? :smile: My understanding is that one reason W-speed was not adopted because of lack of computing power. That is no longer the case.
     
  24. ic-racer

    ic-racer Member

    Messages:
    7,473
    Joined:
    Feb 25, 2007
    Location:
    Midwest USA
    Shooter:
    Multi Format
    Of course they were interested in developing a universal standard. For one's own work an EI based on what gives one the best prints is the way to go.
     
  25. Stephen Benskin

    Stephen Benskin Member

    Messages:
    1,646
    Joined:
    Jan 7, 2005
    Location:
    Los Angeles
    Shooter:
    4x5 Format
    Pt 1

    What makes the fractional gradient speed method significant and still relevant is its connection to the print judgment speed test. Photography is a subjective medium and the only way to determine what process works or not is to evaluated the result from the process. The first excellent print test was just such a psychophysical test. Photographs were made of a range of subjects over a period of time. Detailed information for each scene and exposure was recorded. Sensitometric tests were run at each step. All of this is thoroughly covered in the Loyd Jones paper The Evaluation of Negative Film Speeds in Terms of Print Quality.

    A series of prints were made from each negative and these prints were judged by a panel to determine the quality of the prints. The negatives were then evaluated to determine the correlation. The conclusion was print quality increased as negative exposure progressively increase, but only to a point. Then the level of quality leveled off.

    first excellent print.jpg

    The initial point where the prints were deemed excellent is call the first excellent print point. The results from the test indicates there is no ideal or correct exposure with black and white negatives. Only that a negative shouldn’t be exposed below a certain point. In practice, there are advantages to using the minimum necessary exposure. Exposure times are increased. Sharpness is maximized and grain is reduced. When the tests were conducted in the late 30s, the average negative size would be considered large format today and didn’t require a high degree of enlargement. Consequently, the range of “over exposure” without a change in print quality would be larger than with 35mm film. This still leaves a bit of room to chose from. What is key about determining the first excellent print point is that it defines the minimum, the base to work from.

    As Nelson and Simonds writes, “it is too time consuming to use the print judgment method except for basic studies. A sensitometric method is much more rapid.” Once the basic study was finished, it was necessary to find a sensitometric alternative that would yield consistent results with was many different types of film agreeing closely to the print judgment speed tests. The legitimacy of any method is contingent upon this point.
     
  26. Bill Burk

    Bill Burk Subscriber

    Messages:
    4,984
    Joined:
    Feb 9, 2010
    Shooter:
    4x5 Format
    I love that chart. My challenge to the method, while I accept as true that the "First Excellent Print" is the best place to begin.

    I wonder how hard the darkroom workers had to work to make those prints.

    That is why I choose, my own preference, to pick a point slightly to the right. So I overexpose a little past "First Excellent" deliberately. That makes my work in the darkroom easier.