hp5 Reciprocity Effect Question

No, that is not all you need. You accused me of taking another person's work without giving him credit, and you did it in a public forum. I did no such thing, and the article that I published proves it. Furthermore, by your own rule, you should not have made those comments without seeing the work that you claim violated your rules of referring. You have made no valid criticism of the accuracy or usefullness of my work. You have warned others not to use it or to do so at their own peril. These are not the actions of an honest, unbiased man. The graph that you saw had a reference attached to it to which you should have referred before you jumped to conclusions about where I got it. That graph did nothing but put to use standard methods of deriving empirical relationships, and the original text described that process in detail, including the method I used to determine the probable error of fit, and the reason for choosing one value of one of the arbitrary coefficients to represent 400TX, 400TMX, 100TMX, HP5+, and 100Delta. Everything I did was aboveboard, and the mathematical evidence was thorough and convincing to anyone who had no extracurricular bones to pick.

Patrick;

 

I responded to what you posted. You posted no references, but the posting itself constituted a publication and therefore required refereces itself and not recursive references to your own work.

 

Here -> http://silvergrain.org/Photo-Tech/reciprocity.html <- is a properly written and documented 'posting' on the subject with graphs that predates your article by several years and includes almost all of the references you need.

 

Above, in this thread, you have denigrated Ryuji's work. I offer his posting on reciprocity as evidence to the contrary and evidence of having a plot similar to yours. He has simply taken published data and plotted it on a log - log scale just as you did and come up with the same reciprocity failure behavior, long known in the literature.

 

It is amusing to contemplate the fact that hundreds of people over a century of work have done probably thousands of reciprocity tests and used the same formula you published, but in the years to come people may refer to reciprocity plots and formulae as the Bond-Gainer reciprocity law. And it will happed due to the manner in which they were posted on the internet. I submit to you that the audience here is probably far larger than that of the magazine article, and therefore demanded as good or better presentation of something so technical.

 

Ron Mowrey

Ryuji did not plot the same data, nor did he plot it in the same way I did. I plot only the correction that is to be added, not the corrected times. There is a huge difference. He did not have the experimental data that Howard Bond produced. The 5 films that I plotted were within 1/3 f-stop of the straight lines on log-log paper. There is one line for each film. The slopes of the lines are not statitically different. I did my work independently of anything anyone else did, which was the analysis of available data. You are not excused from following your own pontification on proper reporting. You have no basis for saying that my data analysis is unreliable.

Patrick;

 

I did not say your data was unreliable. Not in any way. Please don't misquote me.

 

I said your work did not give proper recognition of the fact that the formula you derived was done nearly 100 years ago.

 

I said that the data was dangerous to use due to the difficulty in getting reliable data from one point when reciprocity changes the entire curve unpredictably. Ryuji supported this view.

 

I agreed that flare was important and mentioned that my work was done in the absence of flare and asked if yours had accounted for flare.

 

I merely referred others to Ryuji's page to see a complete compilation of reciprocity with plots on log - log paper. I don't mean to imply that he did the same work as you did. But he did a very credible work with references, and that was my intended point. He did a full scientific report on the subject. It was intended for all to see what such a report would look like.

 

When this discussion started, you didn't know of the work in Mees, nor the references that I cited. That is what counts.

 

Ron Mowrey

And you still do not know of my work. None of what I did was done 100 years ago. Howard bond is not that old. Neither am I. I analyzed data that Howard Bond produced by experiments with current films, which were 400TX, 400TMX, 100TMX, HP5+, and 100Delta. I used methods of analysis that were no different than what I used to analyze flight, wind tunnel and human factors research data for NACA-NASA. These are long in the public domain, and so far as I know are used universally by research scientists and engineers. Along with these methods of analysis, one must use some powers of observation. I did not steal those from anyone else, either. I probably should credit my parents, if anyone. I know they were proud of me while they lived.

 

As I said, Ryuji did not separate the indicated exposure from the corrected exposure in his plots. I said I could not use the data because they gave no possible way of extending the results in graphical form and the numerical values were not present. I gave both in my paper in Photo Techniques and credited Howard Bond with his contribution. Anyone who wants to try a different analysis has the data I worked with at hand by readinf the paper. The plots that you could not read were unfortunately limited in size and content by the internet. They are quite readable in the paper. I am quite sure that anyone who uses the data and curves will get good results that are as good as the film manufacturers make the film. I have not touted my method on the basis that I found that all the films I tested showed the same basic curve within the probable error of the tests. I could find, in other words, no statistically significant difference between exponents derived from least squares analysis of the individual filme and the exponent derived by lumping all together. To whom do I owe that kind of analysis? Certainly not to Ryuji nor to you. Must I cite every math or statistics or information theory text and author that went into my way ot analysis? BS!

 

You are HE double hockey sticks bent on proving that I don't know what I'm talking about by showing that you don't know what you are talking about. It doesn't work that way. You've got to read what I wrote before you can say it is wrong or not to be trusted or anything else about it. So far as anyone on this forum can tell, you are basing everything you have said about my treatment of reciprocity on an out-of-context post that I sent to try to help someone. Grow up!

Patrick;

 

I never said you were wrong. I said that the data derived by your method was done that way for ages, you didn't credit that, and that deriving that data could be dangerous from only one point on a curve and due to large possible variations.

 

I have far more experience to judge this than you, but - you have a big following who will believe you no matter what, and you will never admit that what I have said might have an element of truth.

 

In previous posts, I actually ran experiments and showed that volumetric vs gravimetric measure of solids could vary by as much as 20% depending on crystal habit, and you refused to believe me even though I was supported by other chemists. You still continue to support the volumetric measurement method. I could go on and on, but you know very little about chemistry and photographic chemistry and engineering in particular. So, as a result it is very difficult to explain things to you which come second nature to me due to my background.

 

This is not intended to be derogatory. I know little about aeronautical engineering and mechanical engineering. The difference is that I don't try to advise people on how to build aircraft. Trying to extrapolate chemical knowledge into aeronautical engineering would be irresponsible.

 

I have read many of Mr. Bond's articles and find them informative and useful, particularly the latest ones on unsharp masks. I have cited them positively here on PN as a matter of fact. I find your articles and posts, on the matter of chemistry and photography, to be of much less value as explained in my comments above and elsewhere on PN.

 

Ron Mowrey

Whatever you may think you know about photography and chemistry and other things, you still do not know how I analyzed those data that Howard Bond produced. I did NOT use one data point to define a curve. I did a statistical analysis of the fit of the curve that I observed would fit the data. If it had been done 100 years ago, that doesn't keep me from using it. The method of analysis is the same no matter what field of knowledge we are talking about. There is a scientific method that you should read up on before you make such judgements as you have. It does not begin with a theory. It begins with observations. My work stands alone as the only analysis of its kind of Howard Bond's experiments. You would find that out if you would bend your stiff neck long enough to read it.

 

I worked and taught in topographic computing in the Corps of Engineers. I studied chemical engineering 3 years before changing majors to aeronautical engineering. I designed star charts for the Mercury astronauts. I became an internationally known expert in non-linear mathematical models of the human operator. I designed apparatus for measuring eye movements of pilots and methods of analyzing those movements. I analyzed wind tunnel data for comparison with linearized potential flow theory. I was principal oboist of the Norfolk Symphony Orchestra for 7 years. What was it you said I don't know anything about? At least I know what method I used to analyze the reciprocity corrections that Howard Bond deternined experimentally, and I know that method has been in the public domain for centuries and is used by scientists and engineers in every field. I also know that if anyone had analyzed it the way I did before I did it, it would have been widely known and used by every serious photographer.

 

When you tell me that you have read my article, and you have comments specific to it, I will think a little better of you. As to that other garbage about volumetric measurements, how many loaves of bread have you made by weighing the ingredients to the milligram? Do not flour, sugar. salt, yeast have the same kind of variability? Bah! Humbug.

Patrick;

 

To star with, baking is not an exact science. It is an art.

 

Further, I meant one spot on the curve such as in the toe. Please read my posts more carefully. Curve shape varies and therefore you can take a million points from the toe and measure them, but the statistical average of those points from one point (or spot or position) on the curve are often meaningless because reciprocity changes the curve shape. Since you appear not to understand this, then it just shows your own lack of understanding of the subject.

 

We measure toe, midscale and shoulder points with multiple tests and plot all 3 curves to fully understand reciprocity.

 

Without an understanding of the change in characteristic curve as a function of reciprocity, the data are of less use. It depends on where you measure and where the user places his exposure relative to what you measure.

 

Please think about what I am trying to explain. Don't reject it out of hand. I know what I'm talking about.

 

Ron Mowrey

As I said, no matter how much you know about photography, you still do not know what I did or wrote about. What should I listen to you about? You haven't said anything I don't know. If someone else divided the reciprocal relationship into indicated time and correction increment, and plotted only the one on log-log paper, and did the necessary curve fitting, the that person would have gotten the same result as I and we all would have known about it because the manufacturers would have put it on their data sheets. Instead, Ryuji, for example, decided he needed 3 arbitrary constants to put the curve on log-log paper. Why? Because he did not separate indicated exposure from total exposure. He of all people should know that the log of a sum is not the sum of the logs. If you make the separation, you need only a slope and an intercept and a straight edge to draw the correction increment as a function of indicated time on log-log paper. Prove to me otherwise by data, not by your favorite hero's theory. Exery set of axioms at least as rich as arithmetic is either incomplete or inconsistent. What we see is real. The way we describe it is imaginary. You need a good imagination to get good descriptions.

 

If reciprocity changes the shape of the characteristic curve, so what? The only thing we can do about that is change development to controll contrast. If you do not have shadow detail, you have nothing.

 

I did not do the experiments. You really should read Bond's paper as well as mine. He used step density wedge images to decide the correct exposure. He found no significant change in contrast index due to reciprocity failure in the current films. The only thing the practicing photographer CAN do at the time of exposure is to adjust the exposure so as not to lose the shadows. Any other magic must be done in the darkroom. Considering that the majority of those who run into reciprocity effects are the users of large format cameras, whose priority is the setting of aperture for artistic purposes, Bond chose not to consider the changing of aperture to be an option.

 

You are a worry wart. That is what my father used to call me, come to think of it. You say I used someone else's equation, but it won't work anyway because the film changes in ways we cannot control. You say we cannot use volumetric measurements of developer chemicals because something may have change since we opened the bottle to change the specific gravity. Of course, the most likely thing is taking up miosture, which affects the composition of the chemical, which is not compensated by weighing it out. In the first place, I never suggested using volumetric measurements for scientific measurements. When I use them in experiments to find a formula, be advised that that has been done also for a hundred years or more. When I find what I want, I can weight spoonfuls, publish the formulas to the nearest milligram, and pretend that those measurements are to be precisely observed. Phooey!

 

When did I say anything about only one point on the characteristic curve? If I said one data point, it was referring to one indicated exposure time with the associated reciprocity correction. That is a point on any reciprocity chart I have ever seen published. As to the criteria for what constitutes a proper correction, I have no say. That was up to Bond, who is a careful worker in photography. Who am I to argue with him?

 

Oh, what's the use. You got your nose bent out of shape a long time ago, and instead of trying to straighten it, you're working on bending mine. I think everyone can see that except you.

Patrick;

 

Please read again my reference to Mees above. I am not the originator of the precaution given about using this method.

 

Please also note that you have made no mention previously in your posts about possible contrast changes, or where to take measurements on the curve, or adjusting development. It is only when I brought up those points that you begin to make cautionary statements along those lines which now begins to bring you more into alignment with my comments to the readers that they use this method with caution. It also alerts the readers to the fact that you have not posted a lot of information and technique that they should know inorder to make this a workable method.

 

These readers need to know everything necessary to reproduce your method.

 

I'll warrant that most of them have not read your article. One reader of this thread lives in Europe, and he has told me that he is unable to get the magazine there, and must rely on these posts for his entire source for any descriptions of methodology. So, I have to balance my repsonse based on what the readers of PN are capable of doing based on your posts and not magazine articles with limited availability. You must understand that as well!

 

Now, once again. An ideal film exposed to a series with decreasing light intensity and increasing exposure will either undergo no change (No LIRF) or a speed loss only (LIRF in speed). This type of behavior generates a family of identical or parallel characteristic curves and will obey your method (Kron's method) perfectly, provided the tester derives your constants with enough accuracy by using statistical methods (another thing you didn't explain to the readers).

 

That was an ideal film. Many real films will change contrast however and therefore produce families of curves with decreasing contrast (slope) with LIRF or increasing contrast with HIRF (in the experience I had with certain products - it varies from product to product and may even introduce kinked curves if blended emulsions are used). This latter type of film will give different results according to your formula depending on what point on the curve is measured and will fail utterly if the user does not place the center of his exposure on that measured point. Changing development will not work unless the reciprocity test is done with each longer exposure processed at the development time that reproduces the exact curve shape of the original exposure. This often cannot be done, as the change in time introduces errors itself. It therefore becomes a case of a 'dog chasing its tail'.

 

Therefore, you often see mfgrs curves not meeting your measurements as they may have made them at different points on the curve, or with different development times, or in another developer than you used.

 

I hope that I have presented enough information so that the users of your method can understand that it will work, but only as a rough guideline, and not an exact measure. And, it can fail with certain films or if they do not determine the constant(s) with enough precision through statistical methods as you describe above. And since it is a rough guideline, taking a lot of work on the part of the user, it is just as useful to use the mfgrs data and bracket, because it takes less time, costs less money, and will work just as well or perhaps even better in the long run.

 

I have tried to clarify this as much as I can, but obviously you don't believe or understand my reasoning. Or both. I'm truly sorry that this has happened. We have much to learn from each other, but it seems that you will not open your mind to hear anything that I try to say on any subject. I have tried to learn from your methodology and I have learned this - again repeating - it is old, and can be inaccurate. It has an error factor which depends on the film's response to LIRF or HIRF and the users ability to derive accurate constants.

 

Ron Mowrey

I can't believe I missed out on this interesting thread for so long. I have read both Patrick's article and Bond's article from Patrick derived his experimental data.

 

Anyway, Patrick, you have taken some data, done an analysis of it, and then proposed an equation that does a pretty good job of replicating the data. That's great!

 

I think you are missing the first point that Ron is repeatedly trying to make - that similar work was done a long time ago by others.

 

So the simplified equation of Kron as supplied by Ron above is:

Log It = K - a Log I

 

At the start of your article, you say, "Some time ago, I noticed that available reciprocity behaviour data showed a pattern. Although each film required different corrections for long exposures, the corrections of any one film were quite well expressed by the following equation:

 

t(a) = t(m) + a*(t(m))^b" (End quote)

 

And then you derive your second equation from it:

Log(tc) = log(tc,1) + 1.62 log

 

Well, the problem is that Kron's and your second equation are very similar.

 

But now, when this has been pointed out to you, you say what you did is different.

 

"And you still do not know of my work. None of what I did was done 100 years ago. Howard bond is not that old. Neither am I. I analyzed data that Howard Bond produced by experiments with current films, which were 400TX, 400TMX, 100TMX, HP5+, and 100Delta."

 

The method you used do derive hose formuals is not necessarily significant. You seem to be puting a lot of weight in your data reduction techniques. And how do you know that Kron did not use the same data reduction techniques?

 

Anyway, what is significant are the final equations and their similarity.

 

So you say that your equation is different - "My equation only deals with the correction to be added to the indicated exposure."

 

Well, you find that significant, Ron does not seem to. Perhaps Patrick, you have the time to play with the math and the curves to determine if it is or not. Perhaps it would help defend your claims.

 

Secondly, the question of referencing similar work. Ron is used to working in the industrial, patent-driven world when it comes to photographic processes. Patrick is not. Patrick looked at some data, came up with an idea, tried it out, and in the space of one magazine publishing cycle, had an article published. In a popular magazine. Not in something that could be called a scientific journal. Although PhotoTechniques is one of my personnal favorites, it cannot in any way be considered a scientific journal.

 

And as such, I think that Patrick can be forgiven for not doing the amount of backend research that one should have done for a scientific journal. Perhaps even if he had been handed references, in the past, that contained all the info he needed for this research.

 

But it would have been really great if you had, Patrick, as I suspect we would not be having this discussion right now if you had.

 

OK - so Patrick may have never seen the equation that Kron derived. Well, things like that can happen. You live and learn. Afterall, I got the impression that you enjoy learning - otherwise, why would you have been playing with all that reciprocity data instead of doing something else?

 

Ron used Ryuji's web site as an example of a well written web site that deals with scientific subjects - he has given credit for where his information and ideas came from and those that did previous research. We should all strive to do that.

 

Patrick, you say that your article is based on the work of Bond, and any credits that he has given. I have his article in front of me, and I just read it. Bond thanks Ilford and Kodak for suppling him with film, and refers to their reciprocity charts. That's it. No references to previous theory on reciprocity failure. That's fine, Bond's article merely sought to present the findings of his reciprocity testing. And again, that is really to be expected, as PhotoTechniques is not a scientific journal. And we should understand that space for articles in that magazine, as with many others, is often at a minimum and I'm sure the first thing an editor would want to chop is the references anyway. (Although Robert Chapman in his Photochemistry column always seemed to do a good job of giving references.)

 

Patrick, You keep bringing up the point to Ron - "As I said, no matter how much you know about photography, you still do not know what I did or wrote about." (In reference to Ron not having read your article.)

 

Haven't you written enough about your equation, graphs, and methodology in this post and the one on APUG about what you wrote? I just read you article, and other than some more talk about about your method of analysis, and how your "b" number happens to coincidentally be close to the golden mean (does that matter or are we just looking for more patterns?), and how to make a "calculator" of the graphs, there is not much more there to learn. If you think there is more that needs to be added, you are free to add it here after all.

 

I hope we can all learn some lessons from this - do some research on things you print, whether on paper or the Internet. And have at least a second set of eyes, and ones that are qualified to do so, take a look at your work before it is published. As questions about what you did and how you came to your conclusion are bound to be raised at some point.

 

Kirk

 

Patrick did touch on, in a recent post, some of what Bond's methodology was for deriving his reciprocity numbers. Here's what I found significant, in regards to several questions that were raised in this thread earlier.

 

Bond made his exposures in camera. He set up a white card, placed a 5 inch 30-step wedge against the film in a holder, and made his exposures. He found that he was getting a significant amount of flare in the film which initially inflated his Zone III density readings due to the oversized white card, and so he made a little baffle that he placed inside of his 5x7 camera to try and reduce the amount of flare - he was using a 4x5 reducing back so he had some extra room in the bellows. (Bond did suppose that if he had used a smaller white card that was surrounded by a black card he could have better minimized the amound of flare he was getting.)

 

Lighting levels and exposures were determined so that he recieved a Zone III density on the neg. For Bond, he used 0.28 for this value. (I assume this is above base+fog as I missed it if he said so.)

 

Bond appears to have made no corrections to his development times, and he says that upon checking higher Zones, he decided that he had no significant change in contrast. Quoting him, "In the past, films typically yielded increased density ranges with long exposures. The extra exposure that rendered Zone III as planned was less needed in the high zones, so there were elevated, increasing the density ranges. This situation is much improved. At 240 seconds indicated, T-Max 400 and Delta 100 showed no elevation of Zone VIII. Tri-X was up slightly, but only slightly more than the typical variation from one trial to another. The Zone VIII densities from HP-5+ and T-Max 100 were elevated about 2/3 zone. Statins an elevation in terms of zones is very approximate, since the width of the zone (expressed as a range of negative densities) varies greatly with development." Howard Bond, "Black-and-White Reciprocity Departure Revisited", Photo Techniques, July/Aug 2003.

 

So while Bond did check two points on his film curve, he did not actually check for changes in local contrast (shadow or highlight) that may have changed.

 

Kirk

Kirk;

 

In this case, with respect to Bond's work, increased density range can be either lower contrast of the product due to LIRF or higher flare in the optics or both. We cannot tell from his experiment.

 

The tests I ran were in a low flare model 1B Sensitometer with 1:1 imaging on a step tablet and constant light energy and temperature. The intensity was varied using carbon neutral density and a cast carbon 21 step wedge (silver neutral density is slightly colored and also not very uniform from batch to batch).

 

Exposure times went from as much as 1/1000 th sec to 100 sec. The variations in exposure time were sometimes done in power of ten sequence, but varying by stop increments was sometimes done due to the calibration of the ND filters being in 1 stop units.

 

The speed approximately described the hyperbola of the prior art. Contrast was higher at short exposure, high intensity, and lower at long exposure, low intensity.

 

To combat this problem in the color arena, where emulsions diverged widely under these conditions, EK made two families of negative film (Ektacolor S and Ektacolor L) and two families of color paper (Ektacolor and Ektacolor Professional). One was designed for the HIRF region and the other was designed for the LIRF region.

 

Ektacolor paper for photofinishers was tested at 1/10 second and shorter, and Professional was tested at 10 seconds or longer. The aim was to get the same speed and contrast (curve shape) in the two products under the 'average' optimum exposure conditions for the product. This difference was specifically due to LIRF vs HIRF.

 

Aims were constructed assuming similar optical flare factors, but it was known that different printers, enlargers and cameras contributed to this part of the equation differently. It was impossible to predict this figure. OTOH, it was known, for example, that certain demographical areas used certain photofinishing equipment, and this led, for example, to a special aim curve for color papers used in Europe.

 

To insure adherence to these aims, 4 points on the curve were measured and plotted for every reciprocity test. Of course, multiple exposures were made and a statistical analysis was done to insure that the results were valid.

 

Bottom line - I ran a lot of reciprocity tests. So many that I kept a file folder full of the special blank graph paper here at home, and just recently found them at the back of a file cabinet. I used to bring the raw data home and plot them up here. I have a sheet sitting next to me right now as I type this.

 

Regards.

 

Ron Mowrey

On the other hand, there is something to be said for doing these tests under conditions of flare such as might be expected in real life. In any case, you ought not blame me for those shortcomings. The straight line that I saw on log-log graph paper when only the correction was plotted extended also to all the commercially available data I could find, including old Kodak data Ilford data, and the one set of data that Ryuji did divulge, that for Polaroid positive-negative film.

As I said, I did not produce the data nor do the experiments. What experiments I have done were with an enlarger set up with a sort of tent from the lens to the easel, using a 35 mm camera body with the step wedge in contact.

 

How does flare affect the curve shape when the step wedge is contact printed onto the film? It might change the measurement og the incident light, but the same light is hitting the whole wedge, so the curve shape should only be affected if the lighting is uneven along the wedge. The wedge is a small part of the light rectangle to boot.

 

Now if you want to say the data I analyzed was corrupted to start, fine. Maybe it was. But IIRC, Howard allowed as how he had tested the times he came up with in pictorial situations and found them satisfactory. If my analysis can predict the times he used and found satisfactory, what are you complaining about?

 

I recognize that all experiments are subject to errors, both systematic and random. I acknowledged this as fact in my article. That is the reason for the use of various statistical curve-fitting algorithms. I started doing that in 1952 when I went to work at NACA and we had flight test data to analyze. We did find both systematic and random errors. Some of the systematic errors looked random, but turned out to be the haste with which the person who read the oscillograph film on the machine at the computer center. This person was moving the cursor to the next trace before hitting the foot button to record the previous trace. All the random looking spikes were on the same side of the trace. Whatever you say about Howard's methods, the data he collected lie on the same kind of curve that also fits the available commercial data.

You don't like it that I did not tell everyone who introduced the method of least squares. I could have referred to one of my NACA reports where I described a method of adding a single observation to an existing least squares solution.

Patrick;

 

Obivously a contact printed wedge has low flare. That is what I used. If I had not, then any flare introduced by focusing an image of a wedge onto the film would introduce flare.

 

However, a silver wedge can introduce a small amount of flare if contact printed base to emulsion. The amount observed depends on the size of the steps, in a manner similar to an edge effect, and on the size of the aperture on the densitometer when you measure the steps on your experimental film. A carbon wedge should not show this problem as there is no base, but we used a large step rather than the more common small step size wedge.

 

But, you still miss the point that regardless of what statistics are used, the curve derived for one film can vary as a function of the point on the characteristic curve from which it is derived if the contrast is changed by reciprocity.

 

Therefore, one measurement using 'your' method taken at a density of say 0.3, and another taken at 0.5 and another at 1.0 etc will give different Bond-Gainer curves. The problem is, which one do you use? And, if you change development time to match the contrast, then the curve no longer fits.

 

My explanation above will only be false if the curve shapes are exactly identical at all exposure times, and vary only by the change in speed due to LIRF. This is so rarely the case, that I believe that I am giving an appropriate caution to people.

 

The least you can do is acknowledge this fact of science that I bring to your attention, and take a look at it rather than continually dance around this issue.

 

Ron Mowrey

Patrick wrote, "How does flare affect the curve shape when the step wedge is contact printed onto the film? It might change the measurement og the incident light, but the same light is hitting the whole wedge, so the curve shape should only be affected if the lighting is uneven along the wedge. The wedge is a small part of the light rectangle to boot."

 

I can see where if the flare light is illuminating a step wedge in contact with the film, it would raise the level of the lower zones due to the increase in overall illumination. It depends on your lighting situation ultimately - Bond did say he found this to be true for his tests. This would increase your measured film speed and indicate less reciprocity failure. It may not change the curve shape due to the contact printing as the lighting/flare amount should be uniform over the wedge, as you point out.

 

If you have a real world subject and you are not contacting the subject (i.e. not a step wedge) against the film - just shooting a real world subject, the flare would affect the lower zones more, the upper zones less, and this would change the shape of the curve. Just as flare always does in all situations.

 

I find it interesting that Bond did not find the curve changed much, because the amount of exposure from low zones to high zones changes by a factor of 1:100 or so. Look at the changes in expsoure that the reciprocity tables or your graph give, Patrick, over a range like that. So it seems like it should be significant if you are measuring reciprocity at more than just one point on the curve. I wonder if Bond actually graphed the long exposure curves compared with the normal exposures he did as a reference - perhaps he may have seen differences then?

 

I have some tests I did this last summer so I guess I could do that. But I haven't found the time to do the data reductions yet...

 

Kirk

Ron wrote, "Therefore, one measurement using 'your' method taken at a density of say 0.3, and another taken at 0.5 and another at 1.0 etc will give different Bond-Gainer curves. The problem is, which one do you use? And, if you change development time to match the contrast, then the curve no longer fits."

 

That's a good point, Ron. We can't tell that, or even really derive that from the data that we have from Bond to use - other than his indication that at Zone VIII things did not change much. But that doesn't really give us a full picture of what may be happening at the other exposures.

 

Kirk

Kirk;

 

And you have there the exact crux of my position on this.

 

I have had this same problem and had to 'solve' it at work. My supervisor took me through this as a newbie at EK, holding my hand and explaining exactly that statement of mine, giving me the graph paper and saying, in effect, 'go forth young man and do reciprocity tests that are as useful as possible, but make sure you measure the entire curve'.

 

Messers Bond and Gainer should do the same and see what effect it has on their conclusions and recommendations to others. I must say, it opened my eyes. And across products, it was a jaw dropper (to photographic engineers at least). Of course, todays products are much better in this regard, but still not perfect.

 

Ron Mowrey

What kind of skirmish around the facts are you empolying now? You know that Howard used a step density wedge in contact with the film to make his exposures. How is that just using 1 or 2 or 4 readings? I never, ever said to use just so many readings. I can do as well with the reciprocity information at one or two values of indicated time, but I used many more to observe the data and to determine the best fit of it to the curve I chose before I discovered that.

 

Let us suppose that by detailed reading of the step wedge negatives, Bond had discovered kinks in the characteristic curve at long exposures. What could anyone have done to correct them? The ONLY thing one can do at the time of exposure is to change the exposure either by time or alerture. He chose to preserve Zone III by chooeing the exposure time to do so because for him and most large format users aperture is set by other conditions. He examined the higher zones for increased contrast and found it was at worst small and for most of the films not significant.

 

You didn't answer my question about flare. Exactly what effect could it have on contact prints of a step density wedge? The flare has already happened by the time it hits the contact frame. It is no more than a change in the distribution of light, which could, of course change the shape of the H&D curve if it is uneneven over that small area, but can you disregard his work for on speculation that it is in eror? We could do that with any of Kodak's work as well. None of the reciprocity curves that I have seen from them were worth two hoots up a hollow.

Patrick;

 

I never said flare was a problem in a contact print. I did point out some potential sources of flare even when contact printing, in my last post. All of my experiments were designed to completely eliminate even this flare. Thats all. Those sources would lead to miniscule flare, but we avioded them all.

 

I see that you ignored Kirk's post however and seem to completely misunderstand the potential for problem in your methodology. Kirk seems to grasp it well enough, why can't you? You profess to be such a good engineer, you should be able to understand what Kirk can. I understand it. I was taught it by my mentors at EK.

 

Note that I did not say 'error', I have said 'problem' of the sort which can mislead people and result in improper estimation of exposure.

 

I really don't know why it is so hard for you to admit that someone who has worked in this field might know something you don't and might be able to point out a potential problem. I'm trying to HELP you make your method better, I'm not trying to demolish it! I have several thoughts on that to create a better plot. Ah well, you don't seem to care. You demonstrate a principle I have known for years 'NIH' or 'Not Invented Here'. If you don't 'invent' it, it isn't right. Am I on the right track?

 

Ron Mowrey

I can't believe this thread is still ongoing. I thought the topic was well covered a few days ago.

 

Indeed, as I see it, there is no new major issue and I think this thread has departed from reciprocity failure issue. Instead, it is getting into problems with Gainer's general approaches, incorporating several examples drawn from other threads and another forum.

 

I think enough problems have been pointed out about the "Gainer's" equation, teaspoon method, etc. in the past weeks. I personally did some of that in other threads as well. Some of them are misconceptions, some are inaccuracies, and some limitations of arbitrary choices made during experimentation and data analysis. To me they sound like a valuable feedback that Gainer might greatly benefit from for his future authoring activity. As far as I see it, everyone is giving some partial marks to Gainer's articles, but he seems he must defend himself form every possible criticism using all sorts of far fetching arguments instead of accepting limitations, inaccuracy, advantages of alternative methods, etc. Same goes to the common sense of searching for prior art and ethical standard of crediting people who opened up the background for the present knowledge. If the guy who published the equation for the first time were still alive and were participating in this forum or APUG, he would be pissed off long ago. Of course there can be irrelevant/irrational criticisms and in fact there were some "off" remarks but most of what I saw here made sense to me. If the author of an article does not listen to rational suggestions, criticisms, feedback, etc. to improve the work, what is the point of the author asking for the manuscript vetted by other experts?

 

I dislike one favorite phrase of Gainer, which goes like data is real and theory is imaginary. This is a very common irrational criticism against things people don't understand in general. Mechanism of photography is a science and all sorts of theories in the field are a highly condensed form of data obtained before the theory, and each surviving theory has sustained through various challanges and experimental verifications. In this very sense, if the data is real, the theory is also real. Of course those who understand the theory must know the limitations and would apply the theory carefully. I am not an industrial emulsion expert or anything but I at least have some grasp of how commercial emulsions evolved as the theoretical understanding progressed. Experimental tecniques, scientific knowledge, theoretical understanding of things that cannot be directly mesured, and application of the knowledge to products are all linked in the history of photography. Same thing applies to processing, particularly developers. Experimentalists making a blanket criticism about theory is like throwing a piece of stone to the sky.

 

Of course there is nothing wrong not to understand these things. They;re mostly the tricks of the trade for people in the field and industry. End users pay for the products and use them. This is the major distinction from the 19th century photography and it is perfectly acceptable in the 20th and 21st centuries. But if someone publishes something about photography at a deeper level than the end users' how-to, then I would expect the author to understand the principle.

 

At this point, I am not expecting anything new to happen to this thread.

Let's just end this thread instead. For fairness Gainer should be given a chance for one more post. This is not a political debate. I'm here strictly for the utility value as a knowledge source, and I'm getting tired. I don't care who is/are right.

I still don't know where all this is coming from. I did a simple analysis of some data that another produced. My analysis was not for the purpose of testing the validity of the data, and I never said it was. My analysis was for the sake of making the data more easily useful to a working (or playing) photographer. An estimation of the arbitrary constants in the curve I found representative of the existing data, including whatever amount of Ron's work Kodak was willing to divulge and some Ilford data, requires only two experimental estimates of reciprocity correction for each film. Ideally, these should be some average of repeated experiments, and should be photographically usable data. However the data are obtained is not the concern of the curve fitter. I presume they are data that someone wants to make use of and simply needs a compact way of carrying them in his backpack. I think I did that. When I get some kind of reference to Kron's work, I will know if his equation is like mine. All I can tell you about that is that if he is a genius for having arrived at it, then so am I because I had not seen that equation applied to reciprocity before I did it, and that was before 1998, though not published until 2003.

 

I can also tell you that any log-log plot of the total exposure corrected for reciprocity will not plot as mine did as straight, parallel lines. Also, I can say that if and only if it turns out that 1.62 is a universal constant for all films, an adequate descriptive equation only lacks a multiplicative factor which can be estimated from only one valid reciprocity correction.

 

I agree with Ryuji on most things. I agree with Ron on most things. I'm not sure they know it, and I don't know what else I can say.

 

The scientific method starts with some set of observations and a desire to extend them or reduce them to first principles. No other creature has a language as capable of abstraction as ours. That is imagination, the conversion of what we see into verbal (including mathematical) descriptions. What does E = M*(c ^ 2) mean? It is a concept that was derived from observation and so far as we know the words describe a real fact, but the same theory that produced that relationship also change the relationships described by Newton. It was the object of two men to describe the same reality, but they used different words, and one was more successful than the other, even though both were geniuses. Which is the reality, observation or theory? Godel proved that any set of axioms is either incomplete or inconsistent.

 

Yea, we should go on to other things.