My densitometer has adjustments for both slope and intercept. I designed it that way. I calibrate it to read density of film by setting the slope to read two points on the step density wedge. Now that is a problem if the spectral transmission of the wedge is different from that of the film. In fact, there are so many such problems that I do not put a great deal of faith in densitometer readings for estimating printing exposure. Still, they are good for comparative analyses of film and developer combinations, and give an estimate of contrast index that is good enough for government work, so I can tell what grade of paper or filter I need.
Photographers (but not old-timers) seem to put a lot of faith in open loop systems. Engineers would rather use a closed loop system with negative feedback. A high gain amplifier with small phase shift can be made to have an accurate gain that is independent of moderate variations in supply voltage. Our feedback is in the form of test strips and working prints.
As the saying goes, the proof of the pudding is in the tasting. Measuring the ingredients of the pudding with a laboratory balance will not always guarantee a good pudding. There are variations in ingredients that predetermined measurements cannot account for. Film, paper and developer ingredients are manufactured to be within certain tolerances which are not nearly as fine as the measurements we can make.
If you always use graded paper or one of the alternative processes for printing, you have a predictable spectral response. You can measure film densities by the light you are going to print by. If you are doing enlargements, your paper will be sensitive fairly well up into the visible blue. A large part of ultraviolet is filtered out by all the glass in the system.
And we haven't even mentioned diffuse vs spectral densities yet. Oops. Did I say something?
I also use the densitometer primarily for film testing. In most cases printing exposures fall within very narrrow ranges, and since a test print or two is needed for best exposure the densitometer is often less useful than my eyes when dealing with familiar materials.
Originally Posted by gainer
For film testing with staining developers a densitometer can give a fairly accurate representation of effective printing contrast for graded silver papers, if the reading is made in Blue mode, and a very accuate representation of printing contrast with alterntive printing, if the reading is made in UV mode. The results are much less accurate with VC papers because these papers are sensitive to both blue and green light, and in varying proportions according to the contrast filter used. Which means that there is no single filter that could be used for measuring printing densities of stained negatives with these papers for all contrasts.
Stained negatives are a bit of a special case for densitometry. The stain part of the image should have the same diffuse and specular density, the silver part will have differing diffuse and specular densities – so the stain may behave the same while the silver may behave differently when projection printed by a condenser and when projection printed by diffuse light or contacted.
Expanding on the words of Patrick and Sandy, it would probably take quite a bit of work to be able to predict how a stained negative would print using VC paper. The diffuse/specular issue aside, here are my rambling thoughts. I suspect that they will remain in my Mental Exercise/Hey Life’s Too Short Dept, because I can’t see myself actually doing it this way. Anyway, here goes.
First: change the filters in your densitometer. Look through the Wratten curves, or Melles Griot or whoever, and, in the best engineering tradition, take a WAG at the filters that would most closely resemble the two ends of the VC paper’s response (filter method A'). The printing filters themselves won’t do, because they could pass wavelengths that the paper is not sensitive to – but a combination of one or two other filters (that would correspond to the overall spectral sensitivity of the paper) with the printing filters might be worth investigating ('filter method B'). Swap out your two least favourite filters for these and produce calibration curves. Those Status M filters aren’t good for much anyway*.
Measure the step wedge image on film using those two filters. Interpolate for the intermediate curves (method A) or produce a series of curves over the range of printing filters (method B).
Print the step wedge using a series of printing contrasts. You now have two sets of corresponding curves.
Well, it's just a thought.
*Of course they are good for something - standardisation - but they don’t represent the spectral response of colour print materials, and never really have done.
Originally Posted by Helen B
Good suggestions. I am with you in the life's too short mode to even think of doing this kind of testing myself, and since I don't use VC papers it would be pointless anyway. However, there are few people I know who are pretty keen to carry out this kind of testing and I am going to forward your message to them.
Sponsored Ad. (Subscribers to APUG have the option to remove this ad.)
This thread really points out that many of us do not really understand how our tools work, and I think this is a large part of why many of us cannot get comparable results even when we try.
The first thing I noticed that that no one, until Helen pointed it out about 3 pages into it, mentioned anything about calibration of the densitometer. The first step to solving this issue is to calibrate the densitometer. As Patrick pointed out there are open loop and closed loop systems. You can use you densitometer without much effort made towards calibration, and those numbers may work for you if your system is relatively stable, as you have "standardized" on whatever state you system happens to be at. But as soon as we all start comparing numbers with each other, then we all need to make more effort into calibration. You cannot just turn on a densitometer and expect it to be calibrated, especially over long periods of time.
Secondly, Helen, Jorge, and Patrick are all right in stating that when making measurements of the step wedge, the measurements need to be made using the type of light that will be passing through the wedge and onto the photographic material. If you are exposing daylight film and your light source is daylight, then you need to make measurements in the Visual channel. If you are exposing graded paper through the wedge, then you need the blue channel. If you are exposing Pt prints, then you need to use UV. This is, as Sandy says, because step wedges may not had the same transmittance at the different wavelengths. And this also needs to be done as we are trying to measure the amount of light that is passing through the wedge and exposing our materials. I think we all recognize this fact when measuring stained negs - this principle holds for our original step wedges too.
Sandy said, "However, as to which results are correct I again offer the suggestion that the log density units of measurements on the X-axis and y-axis need to be the same, i.e, taken with the same measuring instrument or mode of measurement. Otherwise the distance between equal units of log measurement would have to be different on the x and y axis. So you have apple to orange log units."
Sandy, there are not apple and orange log units. There are just log units. Plain and simple. You are very probably right that your step wedges have different measurements in different color channels. It is very difficult to make materials that have even absorbances/transmission curves over wide ranges of wavelengths. And that is why we do need to pay attention to the type of light that the log units are being measured in, and that is determined by the materials you are using.
So when plotting step wedges, you do need to measure your original step wedge - unless you have one that has been calibrated and it is not out of calibration date. And if you do have a calibrated step wedge, you should be measuring it on a regular basis (even if it is just a spot check of a few steps) to verify that your densitometer IS properly calibrated. I strongly recommend that your densitometer be calibrated every day (if not more, determined with spot checks) if you are doing critical work like making film curves.
The measured values for the step wedge should be used for the x-axis (exposure) densities in your H&D curves.
The y-axis measurements (negative density) do not have to be made in the same channel even. So if you are exposing onto FP4+ and then processing it in Pyrocat-HD, you should be measuring your original step wedge with the Visual channel, and then measuring the resulting FP4+ neg with the blue, green, or a combination of those two channels for VC paper. If using it on graded paper, then you should use the blue channel. And if printing it on Pt/Pd paper, then the FP4+ neg should be measured in UV. But you need to have used the visual channel for that initial measurement of the step wedge.
This issue also brings up something I ran across recently. I have been working on a project with someone that is 1000 miles away from me. I've been processing the film, and he was making measurements of the films and then doing plotting calculations. After a while, I noticed his density measurements did not agree with my spot checks of the films I was sending. We finally determined that while our two densitometers both were within 0.03D for the visual channel, they disagreed with blue channel readings. His instrument was low at lower densities, and much higher at high densities. Well, how do we tell which instrument was right? I had recently purchased both reflection and transmission calibration standards from X-rite. It cost a bit more than $100 for the pair. Since I had these standards, it was quite easy to demonstrate that my densitometer was the one that was more accurate. And not only more accurate, it was within the acceptable limits set forth by X-rite for the values on the transmission calibration standard. We tried to work on his densitometer's calibration of the blue channel, and were not able to get it accurate enough, at least not without sending it out for service. So now are now using my densitometer for all further measurements for this project. A major source of possible future problems have been elimintated for us.
(If any one is interested, the X-rite transmission standard part nr. 810-68 costs around $40, Check out this page, it lists both reflection and transmission standards suitable for what we are talking about here: http://www.xrite.com/product_accessories.aspx?Line=17 )
So getting back to the original question, JMoore DOES need to measure his step wedge with his densitometer using the visible channel since he is exposing TMX-400. He should not be measing the step wedge in the UV channel. He also should not use the default values that are in his BTZS software. He should also be calibrating his densitometer with a calibration standard of known quality.
As a side note, what to people here do as far as calibration their densitometer? Do you simply turn it on and maybe let it warm up, do you have an old piece of film with one number written on it, or do you have an actual calibration standard, and one that has not gotten so old that it has expired? Is your calibration standard one that is calibrated in visual only, or VRGB, or even UV? I'm curious.
PS Helen, are you a scientist? You have a very good grasp on this problem!
A few comments to your message.
“This thread really points out that many of us do not really understand how our tools work, and I think this is a large part of why many of us cannot get comparable results even when we try.
The first thing I noticed that that no one, until Helen pointed it out about 3 pages into it, mentioned anything about calibration of the densitometer. The first step to solving this issue is to calibrate the densitometer.”
One of the reasons that I did not comment on the calibration issue is because I assumed that I was communicating with people who understand the issue of calibration and would not have discussed comparison data unless they were confident of working with calibrated equipment. I have three densitometers, two Gretag D-200 units and an X-Rite 810. All were calibrated recently and all read within log 0.01 of each other in Visual and Blue Mode. As I noted, they read in Visual and Blue from 0.05 to 3.05. However in UV both of the Gretag D-200 units read exactly the same from 0.10 to 2.87.
As for your conclusion,
“So getting back to the original question, JMoore DOES need to measure his step wedge with his densitometer using the visible channel since he is exposing TMX-400. He should not be measuring the step wedge in the UV channel.”
I don’t believe this is correct. The fact is this. If the readings in Visual and UV mode of the step wedge give different values then you will definitely get different results in plotting stained test strips made with Davis’s Plotter program. Just try it and see for yourself. And once you have done the tests explain for me why the results made using readings made of the step wedge in Visual mode are correct and those made using readings made of the step wedge in UV mode are incorrect?
This is essentially Helen's question and a very valid one. Which results are correct, and why? I have offered an explanation for the reason that can be tested. So test it and explain why it is wrong. How would you explain the results?
The color of the light used to expose the film is of course important in that films have different response in both sensitivity and contrast to different color light, but in my opinion the color of the light used to expose the film is absolutely irrelevant to the plotting. Regardless of what light you use to expose the film the plotting results will be different in the above situation, depending on whether you use the UV readings of the step wedge or the Visual readings as the step tablet value for Plotter. The graph must be made in relation to x and y units of equal value, assuming equal distance in length on the graph of these units. If not, the slope will be distorted.
Last edited by sanking; 08-10-2004 at 11:50 PM. Click to view previous post history.
Just one more note.
Are you working with the WinPlotter program? If so I would be glad to send you a few actual files of film tests and you can plug in different values for the program step tablet, i.e. readings made in Visual and UV mode, and you can see for yourself how the plotting changes.
To repeat. The plotting changes depending on which values you use. Why? Which values are correct? Why?
"Are you working with the WinPlotter program?"
No, I don't have the Winplotter program. I usually do all my calculations in a spread sheet that I've written. If you want to get me the Winplotter software I would really love to try it! :^)
"The plotting changes depending on which values you use. Why?"
I know that you are right about the slope of curves changing and that the resulting calculations based on the slope change as well as the range of values used for the x-axis changes. It's simple mathematics. I agree with you.
"Which values are correct? Why?"
That's really the question here. The issue here is: Does the color of light being used in our sensitometric experiments matter? I'm claiming that it does, and it is actually very important, if not fundamental. It is important in both the color spectrum of the light that is used in making the exposure as well as in the color channel of the densitometer that is used in making our readings.
You :"The fact is this. If the readings in Visual and UV mode of the step wedge give different values then you will definitely get different results in plotting stained test strips made with Davis’s Plotter program."
I think you have just demonstrated with that statement the exact reason why these we need to match these things.
Perhaps a Gedankenexperiment would help. (I don't get to use that word very often!) We want to expose Tri-X for a calibration test. What color of light should we use? If we are going to use it for normal, daylight conditions, when we should use a daylight balanced light source. Looking at the spectral sensitivity curve for Tri-X (http://www.kodak.com/global/en/profe...009_0506ac.gif) shows us that Tri-X is sensitive from 300nm to about 650nm. As our lenses will filter out much of the light below 400 nm, we don't have to worry about having a UV rich-light source. And since the Tri-X is not sensitive to wavelengths above 650nm, we don't need a source that is infra-red rich. So a source that is similar to daylight will be best. So now we get a light source that is close to daylight balanced, a sunlit wall, a studio flash unit, an enlarger that has been filtered to daylight.
Now we need to determine the density of the steps in our step wedge. It is quite difficult to make a step wedge for a wide range of the light spectrum. that is neutral in color. And we have proved this to ourselves by measuring our step wedge with a calibrated densitometer. (A scanning spectrophotometer would make this very clear.)
Since we have just demonstrated that the step values change density depending on the type of light we use with them, we need to make sure that we use the appropriate color of light when we make our densitometer readings.
We could measure our step wedge with IR but since our film is not sensitive to IR that would not be appropriate. We could use X-rays to measure our step wedge with, and it would certainly have readings, but they most likely would not be appropriate. We could use UV and we do know that the values differ from visible readings, but since our film will not have much exposure from UV due to our glass lenses, that is not appropriate either.
We could use the blue, green, red, or visible channel - and they may all work well. Depending on how neutral the color of our step wedge is, we may get equivalent results, or at least very close results. But our goal here is to get the most accurate results we can and really try and duplicate the color of light that we will be exposing our test film with. So since we are using daylight, we should pick the visible channel since that most closely matches our light source.
So now we make our exposure of light, through the step wedge, onto our piece of test film. We then take it and dunk it into our developer and process it. At this point, we need to measure the resulting steps that we just made on the test film. Again, we need to first ask ourselves what part of the spectrum will we be passing through this film and onto our printing paper? If we are using Pt/Pd, then we probably want to use UV. If we are making prints on graded paper, then we should use the blue channel. If prints are going to be on VC paper, then perhaps the blue, green, or even visible channel. We may find that our test neg is neutral enough in color that is may not matter which one we choose.
But if we used a staining developer, there may be a big difference between the blue, green, and visible channel densities. Many people like to use blue, some a combination of blue and green. But the important thing is to match (as best we can) the color of the light that will be making the exposure on to the print. I think you can see that readings in IR or X-rays or some other non-actinic light source don't make any sense. So pick the one that most closely matches our materials.
We are somewhat limited by the tools we are using here - our densitometers may not actually cover the ranges of light that we need to properly make these measurements. If we are using a film with extended red sensitivity, the filter in our densitometer for the visible channel may cut off too soon in the IR. It may give us readings that indicate less light is making it through our step wedge than what is really going through the wedge when exposing our film. A wider-range visible filter may be more appropriate for this type of film. But then we are limited by the design of our tools and so we need to pick a filter setting that is most appropriate.
Making measurements with stained materials is no different - we need to try and match our tools. The light source, the spectral sensitivity of our material, and the filtration used in the densitometer readings all need to match as best we can at each individual step - i.e. at the film exposure step and at the paper exposure step. Two completely different step and therefore they both need to be treated separately.
That's why the reading on JMoore's step wedge which will be printed onto his TMax film should be made with a visible filter setting and then his Tmax film should be read with the UV setting if he is going to be using a material like Pt/Pd for printing. And one he has made prints of his test neg on say his Pt/Pd paper, we would use a visible filter setting for the reflection readings, not a UV channel, despite having made the exposure with UV light. 3 different steps with differing sets of needs.
I hope this reasoning makes sense to you!
I have a few questions for you about your densitometer calibration - You mention that your densitometers were recently calibrated. You say recently - do you mean that they were serviced recently and adjustments were made by the service technician? Or that you used your calibration standard that you have from someone like X-rite or MacBeth/Gretag and that you went through your densitometer's calibration routine and the was some time ago and you haven't recalibrated since? Or perhaps something else?
I'm suggesting that we all take the effort to have valid (i.e. not too old) calibration standards for our instruments to make sure that the numbers we read off of them are at least of a known quality, i.e. they are within a known precision and accuracy. I'm an analytical chemist, and no number I produce is worth anything if I can't demonstrate that the quality of my data is known.
I have a 35 mm camera body with a TP 35 step wedge taped into the film plane so I can make contact prints of the wedge on film, a whole roll at a time. I use the enlarger with Quartz-halogen lamp as the light source. I can put a TP 35 in the enlarger and use it to calibrate my easel densitometer. Still, the spectral response of the film is different from the response of my phototransistor. I can, of course, use various filters in the calibration to see if the difference is a matter of slope or intercept or both. Ideally, the photocell and the film I am testing should have the same spectral resoponse within the spectral output of the light source. That is not going to be generally true.
Now when I develop my film image of the wedge and want to get an idea of how printing paper of any kind will respond to that image, what must I do? My densitometer should have the same spectral response as the printing paper. That also is not going to be generally true. The best situation is when the paper has a very narrow response spectrum and my photocell can be made to respond to that spectrum by use of a filter. Not only that, but the meter must be calibrated to read log of illumination in that spectrum.
As far as the meter is concerned, the photocurrent is only a current. The one I designed, and I hope the ones that are commercially available, reads the logarithm of whatever current comes into it. The sensitivity of the photocell has no bearing on that part of the reading. The photocell is only required to have a linear relationship between input light and output current. If that is true, any photocell you put in will give a meter reading =a(log illumination) + b. The constant a is a function of the meter and the constant b now depends on the sensitivity of the photocell. The first is set by one potentiometer and the second by another, and the settings are independent. You can change the slope without changing the intercept and you can change the intercept without changing the slope of the meter's readings. If you calibrate the slope to read log of a current provided by any current source, it will then read log of the current provided by any other current source.
In point of fact, my meter reads log(I1 / I2) where I1 is a photo current and I2 is a current generated internally.
Why am I rambling on about this? I'm not exactly sure myself, but it may stimulate some ideas. The main result is that once a densitometer has been calibrated to read log of an input current, the spectral response of the photosensor will have no effect on the measurement of log(I1) - log(I2) provided that the photosensor is used in the current mode. The meaning of that statement is that when we use the densitometer only to measure contrast, we need not worry about spectrum.
A phototransistor may be used in a voltage mode by putting a resistor in series with it and using the voltage at the junction between transistor and resistor as input to a meter designed to read log of input voltage. Phototransistors are vastly more linear when used in the current mode without a load resistor, where the meter is designed to read log of input current and the transistor is connected directly between power supply and meter input.