Film Characteristic Curve

[jsimoespedro]

Hi all,

I understand the characteristic curve to be S-Shaped. My question is: Why do manufacturers do sometimes post the entire curve and others only the upwards section?

See exemples:
HP5+
http://www.ilfordphoto.com/Webfiles/20106281054152313.pdf

FP4+
http://www.ilfordphoto.com/Webfiles/2010712125850702.pdf

Thanks.

 

[AgX]

In text books you typically find those curves up to the change into the horizontal (which still would be no "complete" curve, as it would go on as horizontal line in case there is solarisation).

In data-sheets often the curve ends at the rising leg.
Reason is that manufacturers often limit their graphics to an exposure of about 1000x beyond inertia. If the curve is steep it will then not yet have turned into the horizontal.

 

[Bill Burk]

Most of the time, highlights (the upper part of the curve) are around 1.0 density. Some alternative processes, or if you seriously overexpose and need to know what will happen, can use densities that are higher, around 2.0 might be interesting. But much beyond that can't be used for pictorial photography (except to make pure white on the prints).

Much of the currently available film has a very long straight line, so you don't see the traditional s-curve until you perform extensive testing... It's interesting, but not very useful. The datasheets are often designed to give you the most useful information.

p.s. I work for Kodak but this opinion is my own and not necessarily that of EKC.

 

[AgX]

Reason is that manufacturers often limit their graphics to an exposure of about 1000x beyond inertia. If the curve is steep it will then not yet have turned into the horizontal.

In also know of curves ending open at only 100X but at the verge of the standard exposure scale the manufacturer employs.
Keeping things in this scale makes it more easy to relate curves. If the typical exposure range falls within this limit nothing is to criticize.


This all depernds on the setting of the manufacturer and the intended use of the material.

 

[jsimoespedro]

My question is motivated by the fact that I was trying to find the dynamic range of several films, and then choose one to push. Higher dynamic range films will tolerate underexpose better, right?

 

[georgegrosu]

Test strips has 21 beaches (step - commonly used).
At a unit log H (exposure) correspond three diaphragms and something.
The confusion comes from the fact the characteristic curves are not figured density readings.
The difference between beaches test strips is half of the diaphragm.
http://www.flickr.com/photos/21121448@N06/4507764748/in/photostream/lightbox/
Graph illustrate only the coordinates used.
Caracteristic curve is for something else.

George

 

[Gerald C Koch]

My question is motivated by the fact that I was trying to find the dynamic range of several films, and then choose one to push. Higher dynamic range films will tolerate underexpose better, right?

Why do you need to push? Among other things you will sacrifice shadow detail. Just use a faster film like ilford 3200 or a speed increasing developer like Acufine.

 

[Photo Engineer]

If you look at those curves, the second one flattens out in the diagram but the first one continues upward. The first film therefore has more latitude and the data goes off the chart for this reason. The second film thus has less latitude and is able to be plotted.

However, I must point out that in my experience at Kodak, films went to a density of 3.0 before flattening and thus had long latitude like film #1 in your reference. Note that film #2, with less latidue, flattens out at density 2.0 but film #1 appears to be able to reach a level beyond 2.0

PE

 

[Rudeofus]

My question is motivated by the fact that I was trying to find the dynamic range of several films, and then choose one to push. Higher dynamic range films will tolerate underexpose better, right?

I think you got something mixed up here. The film you look at is negative material, which means the right part of the characteristic curve is the part where illumination was stronger and resulting film density is higher. This part of the characteristic curve is completely irrelevant for underexposure!

As far as tolerance to underexposure is concerned, it's the same for all films: about four stops. ISO speed is/was defined that way. A gray card shot four stops above rated ISO speed (i.e. underexposed by four stops) would record with 0.1D above background and fog, which is considered about the end of discernible detail.

 

[markbarendt]

Bill Burk has the right thought going.

Regardless of what the whole curve looks like, the density range that naturally prints (a straight print) from any negative onto grade 2 paper is about 1.0 unit of density wide.

So if you were to follow Ansel Adams ideal for a normal scene what actually prints on paper may come from the negative density range between 0.1 to 1.1 on the negative.

Anything at a density of 0.1 or lower on the negative simply prints pure black, anything at a density of 1.1 or higher on the negative prints pure white. So, none the info on the negative above the 1.1 density point (in this specific case) matters because it simply won't straight print.

Ilford seems to have chosen to cut off their charts at about 2.0 negative density. So for FP4 and a normal scene their curve illustrates roughly 3-stops of extra room on the curve that won't naturally print. (That extra room runs on the relative log exposure from about 2.6 to 3.6 on their chart. 1 relative log = 3 stops of exposure.)

This is where the latitude PE is talking about is comes from. Using FP4 for a normal scene you could "over" expose this film by 2-3 stops without affecting the print much. With HP5 the curve shows that you have even more room for over exposure. The printable range is still "1.0 wide" though; so that might be from a negative density of 0.7 to 1.7. Under 0.7 everything would print black, over 1.7 everything would print pure paper white, no detail or tone.

You can burn and dodge to print things outside the "1.0 range" but the question becomes "how much curve do we need to illustrate?" For most of us mortals 2.0, even 1.8 is plenty.

One of the problems I see in charts that run way on out is that we might actually expect to be able to straight print the whole negative, physics is a tough taskmaster and a density range of 2 or 3 simply won't straight print on the "normal" papers we use.

We see a fair number of threads that talk about compensating developers or techniques like stand developing and for staining developers/Pyro. What a lot of these tools are trying to do is make printing the whole negative easier, they are trying to bend the film curve (get it to shoulder at a lower density) or limit the maximum density to fit a longer scale scene inside the "1.0 print range".

 

[ic-racer]

Hi all,

I understand the characteristic curve to be S-Shaped. My question is: Why do manufacturers do sometimes post the entire curve and others only the upwards section?

See exemples:
HP5+
http://www.ilfordphoto.com/Webfiles/20106281054152313.pdf

FP4+
http://www.ilfordphoto.com/Webfiles/2010712125850702.pdf

Thanks.

The traditional 21 step wedge (used to make those graphs) does not provide enough information to show the shoulder when testing many films.
More specifically, the the last datapoint on the graph will be demonstrated on the last step which will be 'clear.' If the shoulder is not demonstrated at that intensity, the system does not provide additional calibrated exposure to show the shoulder.

 

[AgX]

So far we got about half a dozen somewhat different answers...

 

[Xmas]

So far we got about half a dozen somewhat different answers...

Yes but the curve does not tell you too much instead how the fog, contrast and ISO change with development time is more useful? All you can do is expose for shadows and burn the highlights if the scene is too contrasty, or use a softer working developer, the key thing is where you put your zone 1 shadows.

Few film suppliers do characteristics with POTA or D23 type developers, one could assume that no one uses?

If you are scanning it is more difficult if or when the highlights are denser than the scanners DMAX.

Seven and counting...

 

[jsimoespedro]

I think you got something mixed up here. The film you look at is negative material, which means the right part of the characteristic curve is the part where illumination was stronger and resulting film density is higher. This part of the characteristic curve is completely irrelevant for underexposure!

As far as tolerance to underexposure is concerned, it's the same for all films: about four stops. ISO speed is/was defined that way. A gray card shot four stops above rated ISO speed (i.e. underexposed by four stops) would record with 0.1D above background and fog, which is considered about the end of discernible detail.

I actually read ISO determination is more complicated than that. Also involves average contrast over a portion of the curve.

http://en.wikipedia.org/wiki/Film_speed Read topic: "Determining film speed".

 

[jsimoespedro]

Bill Burk has the right thought going.

Regardless of what the whole curve looks like, the density range that naturally prints (a straight print) from any negative onto grade 2 paper is about 1.0 unit of density wide.

So if you were to follow Ansel Adams ideal for a normal scene what actually prints on paper may come from the negative density range between 0.1 to 1.1 on the negative.

Anything at a density of 0.1 or lower on the negative simply prints pure black, anything at a density of 1.1 or higher on the negative prints pure white. So, none the info on the negative above the 1.1 density point (in this specific case) matters because it simply won't straight print.

Ilford seems to have chosen to cut off their charts at about 2.0 negative density. So for FP4 and a normal scene their curve illustrates roughly 3-stops of extra room on the curve that won't naturally print. (That extra room runs on the relative log exposure from about 2.6 to 3.6 on their chart. 1 relative log = 3 stops of exposure.)

This is where the latitude PE is talking about is comes from. Using FP4 for a normal scene you could "over" expose this film by 2-3 stops without affecting the print much. With HP5 the curve shows that you have even more room for over exposure. The printable range is still "1.0 wide" though; so that might be from a negative density of 0.7 to 1.7. Under 0.7 everything would print black, over 1.7 everything would print pure paper white, no detail or tone.

You can burn and dodge to print things outside the "1.0 range" but the question becomes "how much curve do we need to illustrate?" For most of us mortals 2.0, even 1.8 is plenty.

One of the problems I see in charts that run way on out is that we might actually expect to be able to straight print the whole negative, physics is a tough taskmaster and a density range of 2 or 3 simply won't straight print on the "normal" papers we use.

We see a fair number of threads that talk about compensating developers or techniques like stand developing and for staining developers/Pyro. What a lot of these tools are trying to do is make printing the whole negative easier, they are trying to bend the film curve (get it to shoulder at a lower density) or limit the maximum density to fit a longer scale scene inside the "1.0 print range".

Excellent reply. Thank you.

I have an extra question. When I shoot the film at box speed, where does the mid-tone lie in terms of density?

 

[jsimoespedro]

The traditional 21 step wedge (used to make those graphs) does not provide enough information to show the shoulder when testing many films.
More specifically, the the last datapoint on the graph will be demonstrated on the last step which will be 'clear.' If the shoulder is not demonstrated at that intensity, the system does not provide additional calibrated exposure to show the shoulder.

Thank you. That explains the incomplete curve.

 

[DREW WILEY]

Characteristic curves differ from one film to another, one method of development vs another. That why these things are published in the first
place, or why some of us might plot and study our own curves for particular applications. For example, some curves have a long straight line
with a very short toe, while others have a longer toe. They are not all "S" curves by any means. The nature of the toe determines how far into
the shadows you can realistically reproduce discrete recognizable values. Therefore what Rudeofus stated above is largely nonsense. Likewise,
highlight reproduction can vary from one film to another, based upon exposure and developer variables. Published box speed (ASA) is just a
suggested starting point. There a number of models how to turn such things into practical knowledge, like the Zone System, but none of them are perfect by any means. It takes a lot of shooting and printing to real understand what is going on in a practical sense, but sometimes this can be facilitated by getting ahold of a good sensitometry text to help you deciper the curves.

 

[jsimoespedro]

I actually have the three Ansel Adams books. I believe they are very well written and accessible to the beginner (myself). But... it seems arbitrary that (any) film will have a ten zone/stop dynamic range. This is certainly not true for E6 film (I used this a lot when I was a teen). Sensia and Velvia have definitely very different dynamic ranges. So zones and stops cannot be the same thing, but Adams treats them as equivalent.

 

[georgegrosu]

jsimoespedro „Film Characteristic Curve
Hi all,

I understand the characteristic curve to be S-Shaped. My question is: Why do manufacturers do sometimes post the entire curve and others only the upwards section?

See exemples:
HP5+
http://www.ilfordphoto.com/Webfiles/...1054152313.pdf

FP4+
http://www.ilfordphoto.com/Webfiles/...2125850702.pdf

Thanks. "

Your characteristic curve presented shows a variation of exposure 4 log units of exposure (H) which translated into diaphragms are about 12.
Usual, cline of exposure has 10 diaphragms. So, the characteristic curves are more than complete.
To get S-shaped characteristic curve should make another cline of exposure (sensitometric).
This should have a larger number of beaches or change the constate of cline.
I think it is very difficult and expensive to fulfill your desire.

George

 

[Xmas]

Excellent reply. Thank you.

I have an extra question. When I shoot the film at box speed, where does the mid-tone lie in terms of density?

Excellent question : the answer mainly depends on the film the developer and the time in the developer, your light meter measures an average (reflected or) incident or a spot reflected reading you need to translate this to an exposure knowing the film,... etc.,

Which is back to the OPs question,

note I cheat and use a Weston meter and its zone system calculator, cause I cannot recall enough when it is raining and cold and I need time to bracket several exposures. You can print out the calculator if you don't have a Weston. I normally meter for a zone one shadow, ignoring mid tones, easy with a Weston.

Noel

 

[DREW WILEY]

This is true. Films differ with respect to how many "Zones" they will realistically capture, which in turn is dependent upon how you print them.
And I'm just referring to black and white technique. If you are getting precise, you'd have to hypothetically engineer a slightly different Zone
System for every combination of film and developer you intend to work with. But in the real world, things don't need to get that complicated.
The Zone System doesn't need to become a religion. It's just a convenience tool, and after awhile, you learn that you can fudge one kind of
film a little more one way or another, while a different film might require you to restrict its application. But I've always considered the Zone
System to be clumsy or worthless for color film strategy. In Zone System work, you pigeonhole your individual shots according to whether
you plan to develop them standard, or plus, or minus, etc - something which obviously works best for sheet film, but which can be extrapolated to roll film too under many circumstances. And it's no substitute for understanding the finer nuances of any given film curve,
which takes experience. In other words, the Zone System with its reproducable detail parameters of Zones II to VII is a convenient set of
garden-variety rules to learn, but afterwards, you soon learn how to break those very rules.

 

[markbarendt]

Excellent reply. Thank you.

I have an extra question. When I shoot the film at box speed, where does the mid-tone lie in terms of density?

In the 0.5-0.6 area in a perfect world.

Your metering methods can have a marked effect on that.

It's not a perfect world and within the workable range of the film's latitude it's to a great extent irrelevant.

When in doubt use a little extra exposure.

 

[Rudeofus]

I actually read ISO determination is more complicated than that. Also involves average contrast over a portion of the curve.

Of course they have found a way to make things more complicated, but from what I know the four stop below standard gray rule still holds for normal development.

 

[DREW WILEY]

When I take a spotmeter reading I am actually almost visualizing the placement of a shadow value or highlight value on a specific place on the film curve geometry. At this point, I do this almost subconsciously, instinctively. I never work in Zone System mode anymore, unless it is to simply label a shot or roll for plus or minus development. The ASA printed on a box of black and white film means little in the real world. When I'm out in the redwoods the luminance range can easily exceed twelve stops when the sun is out, so I know I need a film with a very long straight line, and will be placing those deepest shadows way down there, equivalent to what is called Zone I or even 0. Diffuse lighting under fog is a completely different ballgame, but might occur on the same day, a few hours earlier or later. Simply using a compensating developer or contracted development (conventional Zone System wisdom) might not solve the problem at all, at least to my satisfaction, because it cruches the midtone microtonality. I want to keep my cake and eat it too. But then, if I'm just out snapshooting with a Nikon on a rainy day, I might choose an entirely different kind of film with respect to curve. If one looks at very different film curves, say, old-school Super-XX versus something like Pan F, it is apparent that the usable lighting range of these films is very different indeed. Stereotypes just don't work, nor do simple ASA standards, which are just a starting point to experimentation. With color films, however, the published box speeds make a lot of sense, because one rarely deviates from standardized process variables. It can be done, but with limited results, and always with a side-effect. But in black and white, you control the processing all kinds of ways. And there
are all kind of variable with the paper too, including VC paper.

 

[BetterSense]

So for FP4 and a normal scene their curve illustrates roughly 3-stops of extra room on the curve that won't naturally print. (That extra room runs on the relative log exposure from about 2.6 to 3.6 on their chart. 1 relative log = 3 stops of exposure.)

This is a great answer but it illustrates the very problem I have with these incomplete curves. That's the attitude that certain parts of the curve are "extra", so it's ok to leave that info out. No matter how you spin it, they are leaving out vital information about the film.

The curve is not complete until it goes flat at the shoulder. There IS a shoulder, for every film; it's a physical fact. The curve WILL shoulder off and go flat at Dmax. The contrast behavior in the shoulder, and most importantly where the shoulder falls, and what Dmax is, is an entire half of the H&D curve, not "extra" information.

The manufacturers are simply opting not to show the shoulder half of the curve. They are printing only a portion of the graph-admittedly the most important portion for most purposes-either because they are embarassed, because they are incompetent, or because they don't want to confuse the audience. None of them are acceptable reasons. It's like if car companies published horsepower curves but left out everything below 1000rpm and over 4000rpm and said "well, for best performance you will stay in that RPM range so that's all we decided to show".

I would love to be able to view complete H&D curves for various films to compare Dmax, latitude and shoulder behavior, but the way the manufactures only print the bottom 6-10 stops, it's impossible. There are plenty of reasons to be interested in the shoulder behavior including to make predictions about highlight rendition, exposure latitude, suitability of the film for masking or lithographic use, suitability for UV contact printing, or basically most of the things you need the curve for in the first place.