Understanding characteristic curve

[Craig]

I'm not sure if this is the right section to put this, but after reading PE's "Photo System Engineering", I think so.

I now understand that the slope of the straightline portion of the curve is contrast, but what about the densities? Are they important in practical photography in terms of matching a film to a subject?

I was looking at the curve for Pan F and HP5+, by my rough calculations I make the slope of Pan F to be 0.8 and HP5 to be 0.66. However, the density of Pan F at an exposure of 3 is roughly 0.75 vs almost 1.5 for HP5. Does that mean a Pan F negative will be thinner than an HP5 negative? I've always associated a thin negative with low contrast, but the slope tells me that isn't so.

Pan F seems to build density quickly with exposure, does that mean I should use it with a subject of a small brightness range, say a cloudy day and a range of 3 or 4 stops from shadow to highlight?. Would HP5 be more forgiving of a larger brightness range? Or am I misunderstanding the information within the curve?

 

[Kino]

Sorry, I'm really not going to answer your question directly...

The slope of the straightline portion is gamma; which indicates relative contrast. The actual contrast is determined by a combination of the film stock, the scene photographed and the film stock response to a particular developer, (at a specific time and temperature).

Assuming your calculations are correct, then the Pan F was processed to a higher gamma than the HP5, but we know that the Pan F is ISO 50 and the HP5 is rated at 400 with normal development. Typically, low speed films have more capacity for graduations of density than higher speed films, but that depends on how they are developed.

The charts you are looking at are one developer (and dilution), at a certain temperature, with a certain agitation, developed by unknown methods (mechanical or hand?) Chances are, if you did duplicate the results yourself, there would be a difference in some aspect of what they used to arrive at their results.

A more meaningful comparison would be to generate a personalized "Family of Curves" for each stock by exposing multiple copies of the density target at a set exposure and then developing those targets at varying times and dilutions of developer. Of course, being this fine-grained gets tricky unless you are machine processing the film, as being consistent is problematic.

Instead, I suggest this...

Shoot an entire 36 exposure roll of a scene (preferably the same shot for each exposure) with each film stock. Include a good range of highlights and shadows. Then cut it into 3 sections in the dark and load them into 3 tanks. Start at the recommended time/temperature/agitation and double or halve those to see how the negative varies.

Process one at recommended time/temp/agitation. Process one at half that time, but maintain temp and agitation. Process the last at double the time and the same temp and agitation.

Now make a contact sheet for both rolls and have a look.

This will tell you far more than reading manufacturer's curves or specifications.

Plus, you will see how the stock curve responds to increasing/decreasing time in the developer AS YOU TEND TO DEVELOP THEM.

 

[markbarendt]

I now understand that the slope of the straightline portion of the curve is contrast, but what about the densities? Are they important in practical photography in terms of matching a film to a subject?

The horizontal scale across the bottom shows how much light was given to the film. This part of the scale is directly related to the scene and the camera settings.

With regard to subject matter in the scene their relationship (spacing on the horizontal scale) remains constant. The difference between say a subjects face and their shirt might be 2-stops, on the horizontal scale that translates to a spacing of 0.66666667 between those two tones. (A spacing of 1 on the log scale equals 3-stops in the scene. 2-stops is 2/3 of 1 or 0.66666667ish.)

With regard to the camera settings, adding or reducing exposure shifts all the tones/exposure levels equally right or left. So imagine that a face gets an exposure of 1.8ish on the horizontal scale.

Using the 2 examples you provided notice that for HP5 that places the face nicely on the straight line at 0.8ish density. With Pan F though that exposure level just barely touches the toe and the density is hardly different than Film base + fog; for Pan F that camera setting gives us an underexposure. Pan F is just a slower film so if we add three stops of exposure, so add 1.0 on the horizontal scale to get 2.8, we can see that the face comes up nicely onto Pan F’s straight line and that face falls at 0.8ish density.

With HP5 and the face exposed at 2.8 on the horizontal scale you should see that the face is still nicely on the straight line and the density on the vertical is now up to 1.3ish, it is not over exposed and still very printable. What needs to happen to print it ‘normally’ is simply more enlarger exposure to place the face properly on the paper.

So, does density matter? No, as long as you get enough exposure to get all your subject matter up off the toe a bit.

 

[Craig]

Thanks Mark, that helps. I was getting fooled by the "relative" portion of the log exposure scale, I thought that it had been normalized and film speed accounted for. If it hasn't, than makes much more sense in the way the curves are shifted.

 

[David Allen]

You do not need to understand a characteristic curve and there is no need to use a densimeter , here is the system that I use with my students. They all find it quite boring but, in one day, they are able to achieve results better than many people who have spent years 'experimenting'. By the way, if you choose to use a two-bath developer (such as Barry Thornton's) you can ignore the second development time tests.

The real key to testing a film/developer combination is to use a consistent and repeatable system. For your information, here is the testing system that I have taught for many years:

Now the key to achieving consistently good negatives is the correct placement of your shadows when exposing the film and ascertaining the correct development time for achieving good separation without losing the highlights. A simple and relatively quick way to way to pin all this down for the future is to do the following (WARNING: reading these instructions is more time consuming and a lot more laborious than actually doing it!!):

1. Find a scene with with a good range of tones
2. Using the box speed, meter the darkest area in which you wish to retain shadow detail
3. Move the camera so that you are only photographing this shadow area
4. From the meter's reading close down the aperture by 2 stops or increase the shutter speed by two stops and then expose 6 frames at: the given exposure then +1 stop, +2 stops, -1 stop, -2 stops and -3 stops less than the meter has indicated

5. Process the film

6. Using the frame that was exposed at -3 stops less than the meter indicated (which should be practically clear but will have received lens flair and fogging - i.e a real world maximum black rather than an exposed piece of film that has processing fog) and do a test strip to find out what is the minimum exposure to achieve maximum black - Print must be fully dry before assessing this
7. Do another test strip with the first exposure being what you have selected for achieving maximum black minus your dry-down compensation then plus 1 second, 2 seconds, etc
8. The time that achieves full black inclusive of compensation for dry-down is you minimum exposure to achieve maximum black for all future printing sessions - print must be fully dry before assessing
9 You now know the minimum time to achieve full black inclusive of exposure reduction to accommodate dry-down
10. Using this minimum exposure to achieve maximum black exposure time, expose all of the other test frames.
11. The test print that has good shadow detail indicates which exposure will render good shadow detail and achieve maximum black and provides you with your personal EI for the tested film/developer combination

12 If the negative exposed at the meter reading gives good shadows, your EI is (when metering shadows where you wish to retain good detail) the box speed (i.e. for 400 film you need to set your meter at 400)
13. If the negative exposed at +1 stop more than the meter reading gives good shadows, your EI is (when metering shadows where you wish to retain good detail) 1/2 the box speed (i.e. for 400 film you need to set your meter at 200)
14. If the negative exposed at +2 stops more than the meter reading gives good shadows, you EI is (when metering shadows where you wish to retain good detail) 1/4 box speed (i.e. for 400 film you need to set your meter at 100)
15. If the negative exposed at -1 stop less than the meter reading gives good shadows, you EI is (when metering shadows where you wish to retain good detail) double the box speed (i.e. for 400 film you need to set your meter at 800)
16. If the negative exposed at -2 stop less than the meter reading gives good shadows, you EI is (when metering shadows where you wish to retain good detail) 4x the box speed (i.e. for 400 film you need to set your meter at 1600)

You have now fixed your personal EI but there is one more testing stage to go.

1. Find a scene with with a good range of tones
2. Using your EI, meter the brightest area in which you wish to retain highlight detail
3. Move the camera so that you are only photographing this highlight area
4. From the meter's reading open up the aperture by 3 stops or decrease the shutter speed by three stops
5. Expose the whole roll at this setting
6. In the darkroom, process one third of the film for recommended development time

7. When dry put negative in the enlarger and make a three section test strip exposing for half the minimum black time established earlier, for the established minimum black time and for double the minimum black time.
8. Process print and dry it.
9. If the section of the test strip exposed for 1/2 the minimum black time gives bright highlights with a trace of detail then the film requires 20% more development
10. If the section of the test strip exposed for the minimum black time gives bright highlights with a trace of detail then the film is correctly developed
11. If the section of the test strip exposed for double the minimum black time gives bright highlights with a trace of detail then the film requires 20% less development
12. You can use the rest of the exposed highlight test film to fine tune the development time.

YES - it is VERY boring but . . .for the investment of minimal materials and a few of hours you will have pinned down so many variables that it is really worth doing.

Back in the real world, all you need to do in future is meter the shadows that you wish to retain good detail with meter set at your EI and then stop down the aperture 2 stops or increase the shutter speed by 2 stops. In the darkroom start your first test print with the minimum exposure to achieve maximum black (inclusive of dry-down compensation) and go from there.

Best,

David
www.dsallen.de
D.S. Allen, fotograf.

www.dsallen.de

Neu / New: http://www.dsallen.de/video.php

 

[Mr Bill]

(A spacing of 1 on the log scale equals 3-stops in the scene. 2-stops is 2/3 of 1 or 0.66666667ish.)

Mark, I think you may have missed your coffee today - I know that you know this, but please allow me to make the correction. On the log scale, a spacing of 1 correlates to a factor of 10, which is more than 3 f-stops (which would only be a factor of 8). For the photographer, it's probably more convenient to think in terms of each change of 0.30 on the log scale as an f-stop change (a change of 0.60 is 2 f-stops, and a change of 0.90 is 3 f-stops, etc.).

 

[markbarendt]

Mark, I think you may have missed your coffee today - I know that you know this, but please allow me to make the correction. On the log scale, a spacing of 1 correlates to a factor of 10, which is more than 3 f-stops (which would only be a factor of 8). For the photographer, it's probably more convenient to think in terms of each change of 0.30 on the log scale as an f-stop change (a change of 0.60 is 2 f-stops, and a change of 0.90 is 3 f-stops, etc.).

I'm ok with being corrected. The precepts still work even if the math needs refining.

 

[Mr Bill]

I now understand that the slope of the straightline portion of the curve is contrast, but what about the densities? Are they important in practical photography in terms of matching a film to a subject?

Hi, actually the more complete use of the characteristic curve is to also combine it with a characteristic curve of the printing paper. (But hardly anyone does this, as far as I know.) Now, the way the film density enters into this is that it reduces the amount of light reaching a specific spot on the paper. Density also uses a log scale, so it also correlates to log exposure. Meaning, for example, that if you make a contact print (negative is pressed against the paper for the exposure) with some arbitrary amount of light, film with density = 0.30 cuts that light in half (same as 1 f-stop). Film with density = 0.60 cuts the light down to one-quarter (1/2 times 1/2 = 1/4), same as 2 f-stops.

But again, mostly this is too complicated for most people, unless they are studying "tone reproduction."

What you would probably be doing is to simply make some prints, then say, oh, this has too much contrast, or whatever, so I want to change my film contrast. If you have a set of characteristic curves for different development times, and perhaps different developers, so you could just pick out a different curve and try that setup for your next roll. But you have to develop a sense for how much is enough. If you were trying to work to Ansel Adams' Zone System, I believe that he gives you some aim values for film density; he can do this by knowing ahead of time what an approximate paper characteristic curve looks like.

I think the most significant thing, other than overall contrast, that you might see is whether the film has a curve where the high end either sweeps up or rolls off - this tells you if the highlights will tend to blow out white, or to roll off more gently (assuming you have exposed them on that part of the curve). Or you might notice that your midtone areas always print a little darker than you like; to counteract this you might look for a film/developer combination where the middle part of the curve has a slight upward bulge.

Alternatively, I think the way most photographers would work is to just try several film/developer combinations, and say, oh, I like this combo best for a certain type of subject.

As a note, if you want to try working from sensitometric data, different enlarger types (diffusion vs condenser) need different contrast values with b&w films. Also, "flare" in the scene affects your shadow areas quite a lot, and this is not included in a characteristic curve. If you want to read up on this, look for older articles here by Stephen Benskin.

Sorry if it got overly complicated.

 

[Gerald C Koch]

Hi, actually the more complete use of the characteristic curve is to also combine it with a characteristic curve of the printing paper. (But hardly anyone does this, as far as I know.) Now, the way the film density enters into this is that it reduces the amount of light reaching a specific spot on the paper. Density also uses a log scale, so it also correlates to log exposure. Meaning, for example, that if you make a contact print (negative is pressed against the paper for the exposure) with some arbitrary amount of light, film with density = 0.30 cuts that light in half (same as 1 f-stop). Film with density = 0.60 cuts the light down to one-quarter (1/2 times 1/2 = 1/4), same as 2 f-stops.

But again, mostly this is too complicated for most people, unless they are studying "tone reproduction."

What you would probably be doing is to simply make some prints, then say, oh, this has too much contrast, or whatever, so I want to change my film contrast. If you have a set of characteristic curves for different development times, and perhaps different developers, so you could just pick out a different curve and try that setup for your next roll. But you have to develop a sense for how much is enough. If you were trying to work to Ansel Adams' Zone System, I believe that he gives you some aim values for film density; he can do this by knowing ahead of time what an approximate paper characteristic curve looks like.

I think the most significant thing, other than overall contrast, that you might see is whether the film has a curve where the high end either sweeps up or rolls off - this tells you if the highlights will tend to blow out white, or to roll off more gently (assuming you have exposed them on that part of the curve). Or you might notice that your midtone areas always print a little darker than you like; to counteract this you might look for a film/developer combination where the middle part of the curve has a slight upward bulge.

Alternatively, I think the way most photographers would work is to just try several film/developer combinations, and say, oh, I like this combo best for a certain type of subject.

As a note, if you want to try working from sensitometric data, different enlarger types (diffusion vs condenser) need different contrast values with b&w films. Also, "flare" in the scene affects your shadow areas quite a lot, and this is not included in a characteristic curve. If you want to read up on this, look for older articles here by Stephen Benskin.

Sorry if it got overly complicated.

Combining the use of the two characteristic curves is what is done in the Zone System. So more than just a few are interested in these curves. Anset Adams has a long discussion on the Zone System and how to do this in The Print, 2nd Ed.

 

[removed account4]

this is why i love this place .. !
i just wing it, but its nice to sort of get it

 

[markbarendt]

Thanks Mark, that helps. I was getting fooled by the "relative" portion of the log exposure scale, I thought that it had been normalized and film speed accounted for. If it hasn't, than makes much more sense in the way the curves are shifted.

You can see the same shift when you compare Kodak's charts too, Kodak does number things differently but that is irrelevant, the numbers are just labels on ideas.

 

[markbarendt]

this is why i love this place .. !
i just wing it, but its nice to sort of get it

Like a dog catching a treat thrown it's way, you don't have to be able to put it on paper to understand trajectories.

 

[Bill Burk]

I always feel that those stylized graphs are difficult to interpret because you can’t read the numbers with enough precision.

I offer one of my more informative graphs which you can read to 0.01 precision. I subtract base and fog. Instead of giving relative exposure I purport to give actual exposure (probably accurate +/- 1/6 stop). The log meter candle seconds scale at top tells how much light hit the film. The bottom scale (like the 0 1 2 3 4 at the bottom of Ilford’s scale) goes from 0 to 3 but right to left on my graph, because I am showing how much light is “held back” from a full blast exposure (full blast would have 0 density blocking the light and my “step wedge” goes up to 3).
The large dashed triangle shows the ASA aim... the line closest to that hypotenuse is the one that best tells the film speed. In my case 13:30 meets the requirement and this film could be called ISO 100 because of where it hits 0.1 density near -2.1 log mcs.

 

[Craig]

In my case 13:30 meets the requirement and this film could be called ISO 100 because of where it hits 0.1 density near -2.1 log mcs.

That's a fair bit different from the datasheet time for TMX in D76 1:1 of 9:30

 

[RalphLambrecht]

I'm not sure if this is the right section to put this, but after reading PE's "Photo System Engineering", I think so.

I now understand that the slope of the straightline portion of the curve is contrast, but what about the densities? Are they important in practical photography in terms of matching a film to a subject?

I was looking at the curve for Pan F and HP5+, by my rough calculations I make the slope of Pan F to be 0.8 and HP5 to be 0.66. However, the density of Pan F at an exposure of 3 is roughly 0.75 vs almost 1.5 for HP5. Does that mean a Pan F negative will be thinner than an HP5 negative? I've always associated a thin negative with low contrast, but the slope tells me that isn't so.

Pan F seems to build density quickly with exposure, does that mean I should use it with a subject of a small brightness range, say a cloudy day and a range of 3 or 4 stops from shadow to highlight?. Would HP5 be more forgiving of a larger brightness range? Or am I misunderstanding the information within the curve?
View attachment 190676 View attachment 190677

No, I think, you are starting to understand. a thin negative can (and often is) high in contrast.low exposure may have lead to thin midtowns and shadows but highlights had enough exposure to develop their full density potential. The result is a thin but contrasty negative, best printed on a medium to harder paper.

 

[Europan]

Should like to say something about the thread title. I think you want to understand a photographic material’s character, not a graph on paper. The so-called characteristic curve is only a means for explaining something.

As others have said you have only the negative side of the story here, the positive one needs to be added unless you want to look at a negative, I mean, why not, an 8 × 10 image framed on the wall can be attractive.

In theory, an image on photo paper should have the same contrast that was with the object. Or gamma 1.0 which is a logarithmic value. Negative films generally end up with a gamma of around 0.6 or 0.7 when processed according to the manufacturer’s instruction. Photo paper develops to a gamma of about 2.0. The product of these values comes to something around 1.3 or a positive image a tad more contrasty, harder. When making positives on transparent materials for projection, slides or ciné film, you want a gamma of 1.55 or so to compensate the Callier effect. Briefly, the graphs serve her/him who wants to keep everything under control rather in commercial work. A photographer judges by her/himself, uses her/his eyes and common sense.

 

[markbarendt]

Craig, Europan is starting the conversation about contrast control. I had purposefully ignored it above for clarity regarding speed difference.

When considering print contrast there is some context that needs to be understood. For example that paper has a fixed range, it goes from paper white to ‘max black’ and that’s it. The only other significant variables for a finished print are how glossy or matte the finish is and the lighting the print sits under.

A high gloss RC paper will look more contrasty than a glossy fiber because the former can produce a darker black. Brighter lighting can help you discern a bit more range but the print has to be printed darker to get there. These limits are hard physical realities that limit how much of the negative can be straight printed.

You can represent these limits with horizontal lines drawn across the film curves. Those lines have a given spacing, for example grade 2 will be spaced on the vertical axis about 1.0 apart. The height of the lines is controlled by enlarger exposure and is adjusted to place subject matter like the face I used as an example above or whatever other tone you please. For a straight print you only get to place one tone the rest ‘fall’ as they will. Only the tones between the lines show detail on paper. Anything above the upper line prints as paper white, anything below the lower line prints as max black. But you can choose a different paper grade and extend or shorten the spacing between the lines.

As you can also see in Bill’s example the steepness of the film curve is a variable too. That’s true of Pan F and all others.

The point here is that Pan F like the rest can be used in most any lighting contrast situation, the bigger issue is typically the practicality of the film speed.

 

[Craig]

Should like to say something about the thread title. I think you want to understand a photographic material’s character, not a graph on paper. The so-called characteristic curve is only a means for explaining something.

Only to a certain extent. I understand about gammas and matching the film and paper, but perhaps I misphrased my question in the first post, so let me have another try:
These curves are included on the data sheet, what are they trying to tell me? What information can I obtain from them that will affect my use of the film, either in exposure or development?

I wasn't really asking about film testing ( although I appreciate those who took the time to type out their methods) but wondering why those curves are presented and what information is contained within them. It seemed like it was a tool that I had no users manual for and no idea how to use.

 

[Europan]

Alrighty. Let’s assume all film manufacturers published graphs of their products in equal manner, identical scales on abscissa and ordinate, same size. Then we would be able to simply put one over the other for comparison of minimal densities, contrast, max densities, and more. I think that’s about it.

 

[markbarendt]

The curve gives us a way to talk about the way a film responds to light.

For example we can see that HP5’s toe is longer than Pan F’s. With experience we might decide that that longer toe helps make us make nice portraits but we like landscapes with better shadow separation, so all the important tones on the straight line. Understanding the shape of the toe and how to place our subject matter to use the toe, or not, gives us a way to make better photos.

On the attached pdf you can see more examples of curve shapes etc...

 

[Photo Engineer]

A contrast of 0.6 to 0.7 is considered the norm for negative films. The slightly higher contrast is often used to compensate for higher flare. This contrast range is used to match a contrast of 2.5 in a paper with a grade of 2. Soft toe in the film can give muddy shadows, but soft shoulder in paper, which might correspond to this loss of detail, actually gives more and the reverse is true of soft shoulder in film vs soft toe in paper. Soft toe in paper greatly enhances highlight detail. In these cases, the mid scales should all be held constant.

This intertwining of film and paper is a complex action of moving splines and differential equations which would be useless to go into here. It is sufficient to say that you should know the curves of both your film and paper in your favorite developer(s). Once you do, you can guesstimate the overall effect in a print by multiplying the mid scale gamma of the paper x midscale of the film (assuming proper exposure) and you should aim for the print gamma that gives the best result. This is often found to be about 1.5.

The zone system is a compressed version of the real picture revealed by 21 step sensitometry. The zone system is virtually useless for precise tuning of your workflow. I've found this to be true over many many years.

By copying the graphs of many manufacturers films and papers, you can use Photo Shop (or the equivalent) to match the X and Y axes of them and then you can superimpose and compare the different products. I had the luxury of actually testing many of these products myself and printing the H&D curves on our onionskin graph paper and then using a light table to compare the results. I spent many hours doing this for films, papers and then final prints and this included many manufacturers.

PE

 

[Bill Burk]

That's a fair bit different from the datasheet time for TMX in D76 1:1 of 9:30

Yes it takes longer to develop to the ASA specifications than most data sheet starting point times.

I think it’s because thinner negatives help you hold qualities like sharpness and resolution, so the published times are a little shorter than the ASA times.

ASA times fit the average scene to Grade 2 paper with a diffusion enlarger.

 

[Craig]

Sounds like I need to start doing some testing then!

 

[Mr Bill]

These curves are included on the data sheet, what are they trying to tell me? What information can I obtain from them that will affect my use of the film, either in exposure or development?

Hi, I think that going into this cold, with only one data sheet and very little practical experience, there is not much useful information for you in a characteristic curve.

Further, any "useful information" that you might get could most likely also be found through a series of systematic tests. Now, when I say "useful," I'm talking in terms of the effect on pictorial photographs.

The benefits of published characteristic curves are mostly in your ability to judge how a specific problem you encounter would show on a characteristic curve, then how to judge different film and developer curves to find a better combination. In other words, you would do much more limited testing because you know how to target it better.

 

[Ted Baker]

This is often found to be about 1.5.

Why 1.5 instead of 1? (colour seems to be about 2) Is this to compensate for reduced intensities of the reflected light from a print versus that those found in the real scene?