Have I Misunderstood ISO?

[138S]

I'm not going to disagree, but I would point out that the graph doesn't say anything about how contrast index is determined - it just says that it changes with changes in development - which isn't the issue, because it does.
What else can change it?

Matt, that graph tells what gradient the sensitometric "curve" will have depending only on development, use the EI you want and will you have exactly the same CI.

Of course EI may have a dramatic impact in how scene is depicted, but not from a change in the CI.

How CI is calculated has no mistery, if you want we may dicuss it.

 

[Anon Ymous]

I'm not going to disagree, but I would point out that the graph doesn't say anything about how contrast index is determined - it just says that it changes with changes in development - which isn't the issue, because it does.
What else can change it?

What else? Nothing. A different contrast index means a different characteristic curve. This can only be obtained by different development. For a given developer, dilution, agitation and temperature, only time can give a different curve.

 

[138S]

What else? Nothing. A different contrast index means a different characteristic curve. This can only be obtained by different development. For a given developer, dilution, agitation and temperature, only time can give a different curve.

Anon, anyway perhaps Matt's point comes form a different effect, if changing CI then film speed is marginally modified, a higher CI moves Speed point a bit to the left.

So it's not a two way relationship, modifiying EI does not modify CI, but a change in CI modifies a bit Film Speed thus having an interaction with exposure.

 

[Sirius Glass]

Under exposing by shooting at ISO 800, merely looses detail and increases the work necessary to make a good print. At least if it was shot at ISO 200 the negative would have been denser and more detail would be available to print. If one is not going to use box speed, at least show more shadow detail.

 

[pentaxuser]

Just to repeat my question which if it has been answered I admit to having failed to draw the necessary conclusion as yet. What should the correct CI be when developed for the same time at speed 800 compared to 400. Can this be worked out?

If the CI remains the same because the development time is the same then what will be the changes in a 800 negative compared to the identical negative exposed at 400 and will these changes be noticeable in the prints from the 400 neg and 800 neg respectively?

Thanks

pentaxuser

 

[138S]

What should the correct CI be when developed for the same time at speed 800 compared to 400. Can this be worked out?

If you develop the same time you obtain the same CI, see this graph:

Blue bar shows the right exposure as recommended by meter at EI400, and aprox 8 min gives the ISO normal contrast of 0.62 CI gradient: density woud rise additional 0.2D for each stop of additional exposure, If yo don't change the development (8min) you always have the sensitometric response labeled "8 min".

At the bottom I placed a sample scene dynamic range in green of 3 stops latitude, in red shows the range that it would take the same scene if closing 1 stop the diafragm. If using EI 800 you will have the same gradient but densities would be lower.

If the CI remains the same because the development time is the same then what will be the changes in a 800 negative compared to the identical negative exposed at 400 and will these changes be noticeable in the prints from the 400 neg and 800 neg respectively?

If you don't change development (say you alwys use 8min), a lower EI will produce lower densities, for the gren range (EI400) the 8min curve shows the densities you will obtain, at EI 800 (red range in x axis) densities will be lower.

________

A sensitometric curve tells what density you will reach for each level of exposure, exposure is shown as the decimal Logrithm of the exposure in Lux*second (in the horizontal axis).

"-3" is 0.001 lux · second, so 0.001 Lux during 1 second (or 1 lux reaching the film during 1/1000 second)

-2 is 0.01 lux · second

-1 is 0.1 Lux · second

0 is 1 Lux · second

1 is 10 Lux · second
-_______


So for each exposure level, you take the right curve in the family of curves for the development time, and that curve says what density you obtain for each level of exposure.

_________________


If you are interested in that I'd recommend you Beyond The Zone System Book, by Phil Davis.

 

[MattKing]

Under exposing by shooting at ISO 800, merely looses detail and increases the work necessary to make a good print.

With T-Max 400, under-exposing by shooting at EI of 800 and not changing the development time will give you a negative that will be easier to print and give you better highlights than under-exposing by shooting at EI of 800 and increasing the development time (a "push" development).
Kodak isn't telling you to expose it at 800 (because some shadow detail will be lost). Kodak is discouraging you from increasing development if you do expose it at 800.

 

[Bill Burk]

This might illustrate what’s going on. This is TMAX-100 but it’s similar to the study we need here. It’s a properly exposed full scale picture shot at ISO speed and developed to 0.5 CI.

Notice how “if” I had set the meter to 200 instead of 100, the darkest point in my picture (my dog’s eye)... would be shifted “to the left” by 0.3 units.

This would put the eye at 2.82 on the horizontal scale. Just about 0.00 density.

Notice that the curve has pretty good contrast all the way down.

You almost wouldn’t be able to tell the difference.

(I might have to say flare probably steals a lot of the contrast in the toe. So I probably wouldn’t get a clean zero.)

So TMAX-400 developed to 0.5 and exposed at 800 would probably just use up that little extra bit and put my dog’s eye at zero. It works because we have good meters and coated lenses. And that short, straight toe that is characteristic of TMAX films.

 

[Mr Bill]

Nope, there's nothing about exposure that would alter contrast index. The characteristic curve covers a wide range of exposure anyway. When calculating contrast index, you place one point of a straight line on the film base plus fog density level. The second point is 0,2 units away from the first, but must fall on the characteristic curve. The third point is on the aforementioned line, but also on the characteristic curve and 2 units away from the second point. You don't have a choice where to put these points, the rules dictate where they would fall. The slope of this straight line is the contrast

This is all true - except when you give the film less exposure (EI of 800), the resulting curve moves and changes its shape (a bit). So the three points you take the measurements from are at different actual positions (not relative positions) than the three points used for the measurement at an EI of 400.

Just for general info, here's what's going on: way back Kodak came up with this thing the called the "contrast index. It is exactly as Anon Ymous says, all the way.

I don't know exactly when it came about, but I was using it in the 1970s (I was a young pup, using the Kodak overlay template, doing sensitometry work in a large finishing lab). When people start in on this sort of thing they tend to think that there is always a "straight-line portion" of b&w film curves. If there is then one can just fit a line to that part of the curve and calculate the "slope" aka "gamma." But often there is no good straight line to work from. So Ilford came up with a thing called "G-bar" and Kodak devised the thing called "contrast index."

Today hardly anyone seems to know what these things properly are, and maybe it doesn't make much difference. I suspect that Matt doesn't know, so is just using some sort of invented contrast measure pertinent to the actual use of the negative under different exposure conditions.

Anyway, this is why there seem to be contrary views to this situation.

 

[Bill Burk]

We had this discussion 5 years ago...

https://www.photrio.com/forum/threads/0-6-standard-contrast.120976/page-2#post-1603500

My opinion has not changed.

Or has it?

Anyway you don't get 0.62 CI at the same development time as 0.56 CI... If you shoot TMAX-400 at 400 your aim CI should be 0.56. Don't bother developing longer to push to 800 because you will get excellent prints with 0.56 CI - just give it the same development. The datasheet as much as says so.

But if you really are into "precision" and want to give the "right" CI for a one-stop push, the right CI is 0.62. For me that's 13:30 in small tank D-76 1:1 68-degrees with inversion agitation every 30 seconds.

 

[Bill Burk]

And I believe Mr Bill is right about Ilford and Kodak coming up with their own contrast measures.

I'm a fan of CI. I like the way when you develop a really short time or a really long time, the first point moves to pick up where you would probably have something interesting (like my dog's eye) in the negative. Then the other end is far enough out to cover most of what you would use when making a print.

Here's the transparency, scaled to fit my graphs if you desire...
http://beefalobill.com/imgs/cntrastindexmeter.pdf

 

[Bill Burk]

MattKing,

The definition of CI does allow this experiment...

You can take two rolls of film, fog one roll of film slightly and then expose, develop and read sensitometry exposures on both films... developed in the same tank for the same time and get different CI.

Another thought... In reality you are right to think the overall contrast of two pictures taken with two different exposures can be different, and the lesser exposed picture will have less contrast typically..... it’s just not called CI

You can carry on with your explanation of the differences how “that” contrast measure is less significant with TMAX-400 than it is for other films such as those with long toes, because TMAX-400 has a sharp toe that makes the straight line seem to go all the way to zero.

 

[ic-racer]

In the majority of photographic applications, the level of exposure is such that much of the toe of the D-log e curve is involved. The use of CONTRAST INDEX [rather than GAMMA] is then likely to be advantageous. "Contrast Index"; C. J. Niederpruem, C. N. Nelson et al 1966.

 

[MattKing]

MattKing,

The definition of CI does allow this experiment...

You can take two rolls of film, fog one roll of film slightly and then expose, develop and read sensitometry exposures on both films... developed in the same tank for the same time and get different CI.

Another thought... In reality you are right to think the overall contrast of two pictures taken with two different exposures can be different, and the lesser exposed picture will have less contrast typically..... it’s just not called CI

You can carry on with your explanation of the differences how “that” contrast measure is less significant with TMAX-400 than it is for other films such as those with long toes, because TMAX-400 has a sharp toe that makes the straight line seem to go all the way to zero.

Thanks Bill (and others).

 

[pentaxuser]

A great discussion. I may not have grasped all that has been said at this point but I will persevere. Thanks to all contributors for taking the time to teach what they know.

pentaxuser

 

[Rudeofus]

I may not have grasped all that has been said at this point but I will persevere.

This whole topic is a lot simpler than it may look. In the end we want to compare contrast, and contrast is defined as the slope of the curve "density over exposure". Accurate measurement of slope in a situation with detector noise is very difficult, therefore numerically stable estimates for slope were defined. Examples for such slope estimates are shown in the document posted by ic-racer. When these books were written, even pocket calculators were out of reach for most people, so some of these explanations sound a lot more complicated at first glance than they should. It doesn't really matter, which estimate you use, as long as you use the same type of estimate in comparisons.

 

[ic-racer]

In this computer era, I measure the contrast on my films with something easy to do on a computer (rather than using a clear plastic template on a hand-drawn graph). I have the spreadsheet do a least-square computation (LINEST()) on the first ten steps beyond 0.1. Takes only a millisecond to get the answer. I call it PCI (Personal Contrast Index).

 

[Rudeofus]

Computers made numerical calculations a lot easier, but did not change the basic problems of these contrast comparisons. There are films with short and long toes, which means b+f+0.1 can be in the toe region or in the straight line region. There are film developers which amplify or shorten the toe. There are developers, which create shoulders within those 10 stops past D=b+f+0.1. You will be in for some great surprises, if you blindly accept what your spreadsheet program spits out.

The main problem arises, if some people search for one precise number to compare D-76 stock to Rodinal 1:200 stand. Your "PCI" is perfect if you want to measure the difference between slight developer modifications, such as D-76 with 100g/l or just 80g/l Sodium Sulfite.

 

[focus_on_infinity]

So, I recently shot a roll of Tri-X metered at 100, and developed regularly at 400. And ... it was pretty reasonably overexposed. Which, to me, was a relief, because I had yet to overexpose anything on Tri-X, ever. So I'm thinking that I'll probably just shoot it at 200, and that will probably be my solution.

Ironically, shortly after I did this, I found that exact advice in the Zone VI Workshop book. So, I'm confident that my experiences are in line with reality, my development processes are now reasonably consistent, and that I'm not imagining things, there are no excess cosmic rays being beamed through my house due to secret gov't experiments, and the world isn't an illusion created by an evil demon, etc ...

Thanks, everyone.