film speed

[LoveMinusZero]

in terms of an h&d curve, is film speed simply where the first density is, or is it relative to the whole toe of the curve? i believe it is the first density off the base, but suppose you have a really big toe, the meter's reading might be still zone 3 or something low, and not zone 5. could this be possible? thank you

 

[bobbysandstrom]

Some people judge film speed as the point where density is .1 above filmbase + fog. Others use the iso standard. The longer the toe (the longer it takes for density to build) the slower the film speed is considered to be. The shorter the toe, the faster the film speed. The film speed (the point at which there is that amount of density mentioned above) changes with development. That is why when you reduce development to contract the contrast you must expose longer(meaning reduce the film speed). Otherwise, your low values will not reach their intended density point. When you expand the contrast by overdeveloping, you must increase the film speed so as the lows don't get overexposed. The lows will come into that proper density range when they are underexposed and overdeveloped.

Hope this makes sense.

 

[gainer]

Not really. The variable that gives you an image to print is contrast, the difference in density between neighboring exposures. If contrast at a certain point in the shadows scales up or down the same as that in the highlights with development, you can get an image on some grade of paper or you can intensify the image by one of several means. Absolute density is not a good criterion for proper exposure. If the contrast at 0.05 density is sufficient to show the contrast you want in visual Zone III, here's no reason not to use it. It often happens that on reducing development to accommodate a wide scene brightness range, the best straight print you can make is dull. That is because the eye viewing the scene scans with automatic sensitivity adjustment at each saccadic movement. The eye is blind between rest points. The camera takes away this option. As a result, whether you adjust development to fit the ideal #2 paper or develop as if the scene brightness can actually be rendered by the print, you will have to dodge or burn on some grade of paper to make the picture fit the paper and make the local contrast more vivid.

 

[LoveMinusZero]

i might be confused about the toe. i think it is the point that the film densities first rises from base, to the point where it starts to straighten. my thought is that if it does not straighten quickly, the of course you will need more exposure to get to a certain density. so i wonder if it takes you an extra stop over the meter to get to middle gray, why not up the film speed by one, even if the point where it first rises off the base density tells you its 100 speed?

i hope i am a bit clearer, i just think that the point where the toe starts to raise above base density ( this is what i believe to be film speed, correct me if i am wrong) is almost not important, because toe could be really long and shallow, thus needing more exposure to reach a certain level compared to the light meter.

 

[Donald Miller]

This is a valid and vital question. The misunderstanding of the ISO standards is why so many Zone System practitioners find that they derate films almost as a matter of course.

There certainly are films that are overrated. But not nearly to the degree that most would believe.

Additionally, the speed point of the film is not a stationary fixed point. It is movable with the development to which the film is subjected.

 

[bobbysandstrom]

loveminus:

I think you might be thinking of the way this works backwards. You set your light meter according to the film speed and your intended development plan ie either minus, normal, or plus. You find your film speed by doing tests. Most people find their film speed by metering a grey card and placing it on zone 1 then proceed to make several exposures starting at 1/2 the box speed up to twice the box speed. The film is processed and a densitometer reading of each will reveal the exposure that is closest to .1 above fb+f. The setting you used to attain that density is considered your personal EI or film speed. From there (using your personal EI) you do further tests using Zone VIII exposures and develop for different times in an effort to find out how long it takes (development time) to produce the density that your paper is able to print just short of pure white.


Keep in mind your toe will change with development. Think about this... Imagine developing your film (that you shot at your personal EI) for only 30 seconds. If you plotted the densities they'd form all toe. Now imagine you developed the film normally. The toe would be shorter of course and would work its way up to form the straight line part of the curve.

So, if you need to compress the range of the scene so it fits your film and of course your paper you must underdevelop. What happens to the toe when you underdevelop? It gets longer or in other words, the film speed changes. It gets slower. The opposite is also true if you want to expand the tones of the scene to fit your film/paper.

It might not be a bad idea to find yourself a book on this subject. You might try AA's "The Negative" It will answer a lot of your questions.

Keep at it. It will eventually all make sense.

 

[LoveMinusZero]

thanks bobbysandstrom, i will pick up a copy of the negative. what i was really proposingis that even films with the same personal EI, after tests and all, still could need different exposures to get to a density above what what used to find the speed. i.e. maybe 2 films rated 100, one might use the meters reading for zone v, another film might need an extra 1/2 stop to get to the same density, even if the 2 films are both shown to be 100 EI.

so i was thinking that having your film speed exactly what it actually is, is not important. but, knowing the relationship between whatever speed, exposure and development is, even if the speed your using is not exactly what the films is. anyways, you always can change the speed and compensate with exposure, right?

 

[bobbysandstrom]

What you are talking about is film curve. Different films have different curves. Some more "S" shaped, some more straight line etc. Yes, if you have 2 films that you tested and rated at 100 but these films have different curves you are absolutely right... zone V may be at a different density and hence produce a different shade of gray. That is why in general some people prefer the look of one film over another. I think you got it.

One other thing... you adjust your development time based on how you expose. You base your exposure on what the scene offers in terms of total zones. If the scene is flat and contains only a few zones the common thing to do is develop longer so you can stretch those few zones offered by the scene so your print will contain a wider range of tones. If the scene is very contrasty meaning has many zones you would then shorten your development time to compact the tones so they all fit on your paper. If the scene has just the right amount of tones you develop for the normal amount of time. You know what happens to your tones when you change development time. So if you are going to shorten the time you want to expose more so you don't lose shadow detail. If you lengthen development time you want to expose less so the shadows remain where you want them. If you develop normal you expose normal ie at your personal EI.

Bruce Barnbaum has a great explanation of all this in his book "The Art of Photography." I recommend this book to anyone that hasn't read it! There's a lot to be learned from it.

Good Luck

Bob

 

[gainer]

And Barnbaum invariably does local bleaching and other manipulations to make his prints look the way he saw the scene. You can interpret "expose for the shadows, develop for the highlights" two ways. If you try to keep the highlight on the paper, you will have to use a higher contrast grade of paper and burn it in. If you keep the contrast over the normal scene range so that that part of the scene fits paper, the highlights that wash out on the paper can be burnt in. If they are washed out of the negative, they are lost. You pays your money and you takes your choice.

 

[dancqu]

... does local bleaching and other manipulations
to make his prints look the way he saw the scene.

I don't think so. He will make a print that looks more like
the scene he would have photographed, did it exist. In a
sence some of his prints are of scenes non-existent.

I'm not interested in cosmetics.
My concern is fidelity. Dan

 

[Stephen Benskin]

in terms of an h&d curve, is film speed simply where the first density is, or is it relative to the whole toe of the curve? i believe it is the first density off the base, but suppose you have a really big toe, the meter's reading might be still zone 3 or something low, and not zone 5. could this be possible? thank you

The function of film speed is to produce a negative that can achieve a print which is considered to be of high quality. Since the idea of quality is psychological in nature, the idea of film speed must be linked to the perception of quality. This is done through psychophysical testing. Under controlled testing, the technical aspects of exposure and the physical aspects of each negative are evaluated for the ones that produce corresponding prints that are perceived as excellent. Various procedures are evaluated in order to achieve type of results in the quality test negatives under the greatest range of shooting conditions and emulsion types. The technique/procedure with the smallest variation is considered the optimum as a definition of film speed.

The method that produces the best negatives under the largest range of conditions is the Fractional Gradient Method determined by Lyod Jones. Jones found that speed is related less to density as it is to gradient. The highest quality was produced when the speed point was placed at a point on the toe of the curve where the toe’s gradient as 1/3 of the over film gradient.

While the FG method produces the highest level of quality, the determination of the FG speed point was difficult and subject to error. The method to determine B&W film speed was changed in 1960 using a fixed density to determine film speed; however, the fixed density did not constitute the aim point for the shadow exposure. The ANSI/ISO method is a technique to determine film speed, but the actual speed point isn’t at the 0.10 fixed density point. As a matter of fact, from the perspective of the 0.3 method, any fixed density method tends to underrate films that are processed to a lower than average contrast (overexpose) and overrate films that are processed to a higher than average contrast (underexpose).

These facts have gotten lost over time. The underlying principle behind today’s ISO method is surprisingly still the Fractional Gradient Method. The ISO method relates to it through an equation that is built into the testing procedures. This is known as the Delta X Speed Criterion. The Delta X speed criterion has managed to unite the more accurate factional gradient method with the easier to apply fixed density method. Contrary to what most think, the 0.3 method wasn’t superseded by a fixed density method. The fixed density method was really subverted by the 0.3 method.

The main reason why the fixed density method tends to underrate films that are processed to a lower than average contrast (overexpose) and overrate films that are processed to a higher than average contrast (underexpose) is because any fixed density point will vary considerably with any variation in processing. With the FG method linked to gradient the resulting film speed changes little for changes in processing. For example, a test I conducted found speeds using the fixed density method ranged over two-thirds of a stop from EI 100 to EI 160, where the processing was from CI 0.54 to CI 0.69 while the Delta X speed method rounded the speeds to 125 even though the range of processing is from an approximate contrast index of 0.54 (ΔD 0.70) to 0.69 (ΔD 0.90). The only difference in the test was the method used to calculate film speed. The same sensitometric strips were used in the application of both speed methods.

So, a fixed density method can be used for determining film speed, but it is not what film speed is about.

The second part of your question assumes each Zone has a corresponding print density. They don’t. The resulting print density is determinant upon both the film and paper curve. One of the reasons why you might have come to the conclusion you have is because the gray scale that is usually presented has equally spaced values from black to white. This is simply misleading. I don’t know if ZS examples use this for the sake of simplicity or because of its relation to the Munsel scale.

 

[Photo Engineer]

Stephen;

Very good explanation. You must work for Ilford, Fuji, Agfa or Kodak.

The only comments I have are regarding the last paragraph:

"The second part of your question assumes each Zone has a corresponding print density. They don’t."

In a negative, each zone is usually either 0.3 or 0.15 in Log E and changes in density as a function of the film curve. It is really an H&D curve. I don't follow this comment of yours at all.

"The resulting print density is determinant upon both the film and paper curve."

The final print scale is equal to the slope of the film x the slope of the paper at a given value of Log Exposure, or is the integral resulting from the multiplications of the equations that represent the two materials.

"One of the reasons why you might have come to the conclusion you have is because the gray scale that is usually presented has equally spaced values from black to white. This is simply misleading. I don’t know if ZS examples use this for the sake of simplicity or because of its relation to the Munsel scale."

All photographic engineers use equally spaced 21 step charts for tests like this and construct a so called zone representation. Ours has 21 'zones' though and uses either 0.3 or 0.15 density unit changes per step. I have two sheets of 8x10 silver step scale in my darkroom right now for tests, including one with each density increment. They are very useful. One of my posted pictures here shows one of the step charts on the right of the photo.

The Munsel scale I'm familiar with is for color rendition, not B&W. So, if there is a Munsel scale for B&W we never used it in desigining photo products. I used it all of the time though for color products.

PE

 

[Kirk Keyes]

Stephen;
Very good explanation. You must work for Ilford, Fuji, Agfa or Kodak.

No, he is just Loyd Jones reincarnated!

Kirk

 

[Kirk Keyes]

In a negative, each zone is usually either 0.3 or 0.15 in Log E and changes in density as a function of the film curve. It is really an H&D curve. I don't follow this comment of yours at all.

I think Stephen is talking about negative densities, not exposure amount. The exposure for each Zone is one stop, but the resulting neg density is also dependant on the CI of the processed film.

 

[Photo Engineer]

I think Stephen is talking about negative densities, not exposure amount. The exposure for each Zone is one stop, but the resulting neg density is also dependant on the CI of the processed film.

Kirk, for all practical purposes, with a properly exposed negative at the proper contrast, the same density in a negative should yield the same print density. Of course, this is as good as any other generalization, but that is where I was coming from.

If you deviate from that, you are outside of the zone for all practical purposes, however let me state some obvious exceptions.

The longer the latitude of the film, the more leeway you have in this situation and therefore with a film with a 3 stop latitude you have 3x the number of positions to place a proper print depending on over or under exposures. With a short latitude film, there is practically only one place (as I stated above) to place a proper print.

In any event, what I stated is a generalization. I could graph it out for you all, but I'm getting totally lazy.

How are your researches into sharpness going? Staining developers? I have not heard anything on these from you lately.

Regards.

PE

 

[Kino]

No, he is just Lyod Jones reincarnated!

Kirk

Not to be a pain but it was LOYD JONES; it was he and his associates who made the investigations that led to the ASA/BS film speed an exposure index specification in 1949 and they were part the American Eastman Kodak Research Park Lab in Rochester, New York.

OK, I copied that from a favorite book of mine, "Exposure Manual" 3rd edition by Jack F. Dunn and George L. Wakefield (1974), but I claim "fair use" under US Copyright laws! ;-)

Stephen Benskin:

Are the methods described (assuming you have a copy of this book) on pages 24 through 28 still pretty much valid for determining the speed point of film? Looks to be pretty much as you describe it...

I read this book often, but often don't understand everything...

 

[Kirk Keyes]

Not to be a pain but it was LOYD JONES

Ooops - typo fixed.

 

[Stephen Benskin]

Stephen Benskin:

Are the methods described (assuming you have a copy of this book) on pages 24 through 28 still pretty much valid for determining the speed point of film? Looks to be pretty much as you describe it...

I read this book often, but often don't understand everything...

Kino,

I have the 1981 edition, but yes it is still valid for determining film speed. The speed point should not be mistaken for the point of shadow exposure. Given the ISO contrast conditions, the fractional gradient speed point is at a constant 0.29 log-H units to the left of the ISO Hm speed point. That's a stop difference in exposure.

Exposure is based on a scene that has a subject luminance range of 2.2 (7 1/3 stops). Under these average conditions, the shadows fall 1.28 log-H units (~ 4⅓ stops) below the exposure meter’s calibration point and the highlights fall 0.92 log-H units (~3 stops) above. The ISO speed point falls only 1.0 log-H units below. The approximately 4 1/3 stops average condition should not be considered a validation of the ZS approach. I've attach a graph showing this. It has additional information Dunn knows but didn't get into.

Anyway, it gets rather involved. I wrote an article for Photo Techniques that dealt with some of this. It was in the Jan/Feb 2005 issue. I forget what they eventually titled it.

 

[Photo Engineer]

Stephen;

Seems you have ignored my queries.

PE

 

[Stephen Benskin]

Stephen;

Very good explanation. You must work for Ilford, Fuji, Agfa or Kodak.

The only comments I have are regarding the last paragraph:

"The second part of your question assumes each Zone has a corresponding print density. They don’t."

In a negative, each zone is usually either 0.3 or 0.15 in Log E and changes in density as a function of the film curve. It is really an H&D curve. I don't follow this comment of yours at all.

"The resulting print density is determinant upon both the film and paper curve."

The final print scale is equal to the slope of the film x the slope of the paper at a given value of Log Exposure, or is the integral resulting from the multiplications of the equations that represent the two materials.

"One of the reasons why you might have come to the conclusion you have is because the gray scale that is usually presented has equally spaced values from black to white. This is simply misleading. I don’t know if ZS examples use this for the sake of simplicity or because of its relation to the Munsel scale."

All photographic engineers use equally spaced 21 step charts for tests like this and construct a so called zone representation. Ours has 21 'zones' though and uses either 0.3 or 0.15 density unit changes per step. I have two sheets of 8x10 silver step scale in my darkroom right now for tests, including one with each density increment. They are very useful. One of my posted pictures here shows one of the step charts on the right of the photo.

The Munsel scale I'm familiar with is for color rendition, not B&W. So, if there is a Munsel scale for B&W we never used it in desigining photo products. I used it all of the time though for color products.

PE

I just wrote a rather long answer to your post and lost it when attempting to spell check. Forgive me, I just don't have the inclination to rewrite the whole thing. Still, I'll give the Readers Digest version.

Munsell’s color tree is comprised of the three component attributes of color. They are Value, later called Brightness, Hue, and Chroma, later called Saturation. Value is the apparent lightness of the color. Munsell used Value as the vertical trunk of the tree. Hue is the recognition of the base color, such as red, green, blue and it forms the branches of the tree. Finally, Chroma is the apparent concentration or purity of Hue and it is represented by the length of the branches.

For our purposes, we are only interested in the trunk. The height of the trunk consists of a spacing of 10 neutral Values in a perceived equidistance ranging vertically from black to white (Munsell notation N/0 – N/10). The standard Munsell value unit is defined as one-tenth the difference in value between a Lambertian perfect diffusion surface of 0 percent and 100 percent reflectance when both are viewed on a background whose reflectance is around middle gray. Neutral gray is used because it is, in fact, a color with an absence of Hue and Chroma so it has is a commonality to all colors. Called achromatic, its appearance will not alter the color balance of the illuminating light source.

There are a number of papers on this issue. Munsell Labs published one in the late 1920s. A Subcommittee of the Colorimetric Committee of the Optical Society of America performed the first systematic investigation of the psychophysical properties of the Munsell system and formulated what has come to be known as the Renotation Munsell. Eventually, it was adopted into CIE Color in the 1940s. A book from the committee's report was published as The Science of Color.

The second part of my earlier post was written quickly. I had spent so much time on the first part to avoid misunderstanding that I had little left for the second part.

My point is that film and paper curves aren't linear, therefore, the resulting image will not accurately refect the scene's subject luminance range. If that is the case, then you can't have specific print values corresponding to exposure. It's more of a range of potential values.

I've attached a file comparing the different Tri-X emulsions processed to identical contrast indexes (adjusted for speed differences). As you can see, a specific exposure can produce differing negative densities. In another test comparing TXP to TX (also adjusting for speed differences), a stop difference in exposure was required to produce the same negative density in the midtone region. The shadow and highlight region tended to produce similar negative densities to identical levels of exposure.

Correct me if I'm wrong, but your example with the 21 step scale as a reference makes me think you are dealing with copy or reproduction which has a different set of criteria. This might explain our difference of opinion.

That's the short version. Parts are copied out of a manuscript I was writing on Middle Gray. This material came from the section on the psychological and psychophysical middle gray. The other two sections were on the physical middle gray and the meter's middle gray. Photo Techniques showed no interest, so I abandoned the article(s).

 

[Stephen Benskin]

I found another example which might help illustrate my position. It is the Tone Reproduction curve from a four quadrant Tone Reproduction Graph. HP5P was used and the same paper type (I believe Ilford Multigrade). The idea was to match the CI of the film with the different paper grades, so that the Paper Log Exposure Range would be identical in all cases.

These curves are generate by comparing the original scene to the resulting paper densities. I've indicated with a dot the placement of the subject midtone. Notice how the tones have shifted. The straight line is a 1:1 reference line. Anything below that line is less dense (lighter) than the original scene, and anything over is more dense (darker). Compression of the highlights and shadows are perceptually acceptable as long as the mid section has around a 1.10 gradient.

 

[sanking]

The main reason why the fixed density method tends to underrate films that are processed to a lower than average contrast (overexpose) and overrate films that are processed to a higher than average contrast (underexpose) is because any fixed density point will vary considerably with any variation in processing. With the FG method linked to gradient the resulting film speed changes little for changes in processing. For example, a test I conducted found speeds using the fixed density method ranged over two-thirds of a stop from EI 100 to EI 160, where the processing was from CI 0.54 to CI 0.69 while the Delta X speed method rounded the speeds to 125 even though the range of processing is from an approximate contrast index of 0.54 (ΔD 0.70) to 0.69 (ΔD 0.90). The only difference in the test was the method used to calculate film speed. The same sensitometric strips were used in the application of both speed methods.

I am trying to reconcile the concept that with the "FG method linked to gradient the resulting film speed changes very little, with my own experience in exposing and developing negatives for processes with different exposure scales. For example, if I were to use the same exposure for a negtive that will be developed to the average contrast required by a #2 silver gelatin paper as for a negative that will be printed in palladium, the density in the shadows of the palladium print will be far more than necessary, perhaps as much as a full stop in excess.

Thanks for your comments on this.

Sandy

 

[Stephen Benskin]

I am trying to reconcile the concept that with the "FG method linked to gradient the resulting film speed changes very little, with my own experience in exposing and developing negatives for processes with different exposure scales. For example, if I were to use the same exposure for a negtive that will be developed to the average contrast required by a #2 silver gelatin paper as for a negative that will be printed in palladium, the density in the shadows of the palladium print will be far more than necessary, perhaps as much as a full stop in excess.

Thanks for your comments on this.

Sandy

That is to be expected. It is also thinking from a density perspective. The Fractional Gradient Method would have a negative with a greater density with extended development than one developed more normally. Jones was basing the method on reproduction quality. Of course, with small formats the additional density in some cases can have a noticeable reduction in sharpness and a increase in grain.

Obviously, the level of density you wish or like in your negative is a personal choice. I personally will overexpose a negative, in relation to the FG method, when I reduce development. Jones was determining a standard method. The Delta X Criterion is built into the ISO standard. My personal opinion is that what ever your choice, knowing the background can only help.

Delta X and it's relation to the ISO standard is covered in:

Nelson, C.N. and Simonds, J.L, Simple Methods for Approximating the Fractional Gradient Speeds of Photographic Materials, JOSA, vol. 46, n. 5, May 1956.

 

[Photo Engineer]

Stephen;

I'm afraid that your arguments avoid the point that the human eye integrates the data in the image. In that case, due to having a toe and shoulder in the photographic image, the toe and shoulder are 'soft' and the mid scale is higher in contrast than the 1:1 slope we would otherwise expect. Therefore, it is unavoidable that the 'contrast' of a print would be about 1.5 in the mid-scale to compensate for the toe and shoulder. This is what we designed for.

This 1.5 mid-scale contrast compensates for the loss in contrast in toe and shoulder and gives the human eye the overall impression of a curve with a slope of 1. It is, in effect, an optical 'delusion'.

As for the Munsel chart, well, it can be used as a B&W tool, but we never did. To use an analogy, you can remove an appendix via the mouth! Why bother when other methods are so much more convenient and less difficult. So, we just used the H&D curve for both B&W and color. We used the Munsel chart only for color space.

I'm afraid that our methods differ, but I am familiar with what you refer to. I'll defer to the experts that I know, and rely on the work of DeMarsh, Zwick and others that I worked with personally and who taught me the methodology one on one in the lab. I have no doubt that what you say will work, but it seems rather round about to get to simpler solutions to the problem.

PE

 

[Stephen Benskin]

Stephen;

I'm afraid that your arguments avoid the point that the human eye integrates the data in the image. In that case, due to having a toe and shoulder in the photographic image, the toe and shoulder are 'soft' and the mid scale is higher in contrast than the 1:1 slope we would otherwise expect. Therefore, it is unavoidable that the 'contrast' of a print would be about 1.5 in the mid-scale to compensate for the toe and shoulder. This is what we designed for.

This 1.5 mid-scale contrast compensates for the loss in contrast in toe and shoulder and gives the human eye the overall impression of a curve with a slope of 1. It is, in effect, an optical 'delusion'.

As for the Munsel chart, well, it can be used as a B&W tool, but we never did. To use an analogy, you can remove an appendix via the mouth! Why bother when other methods are so much more convenient and less difficult. So, we just used the H&D curve for both B&W and color. We used the Munsel chart only for color space.

I'm afraid that our methods differ, but I am familiar with what you refer to. I'll defer to the experts that I know, and rely on the work of DeMarsh, Zwick and others that I worked with personally and who taught me the methodology one on one in the lab. I have no doubt that what you say will work, but it seems rather round about to get to simpler solutions to the problem.

PE

The core of the Munsell tree is the N scale. He needed to define it before defining the branches of chroma and hue. Maybe you are thinking about the use of the Munsell neutral scale while I am referring the theorectical and historical significance of it.

I don't think I avoid the concept of perception. You should look back over the posts. It's a big subject and I may not have given it enough attention. So, I'm not sure what counterpoint you are making.

You might be confusing the paper gradient with the reproduction gradient which is what I was referring to, which means we are talking about two different types of curves. If not, the value of a 1.50 gradient for the middle section is simply not correct. I could refer you to some papers on this subject if you are interested.

I also don't know what you mean by "simpler solutions." I was discussing theory, not practice.