BTZS Testing on negatives developed in Pyrocat-HD

[Kirk Keyes]

Glass - The 1980 Kodak book says, yes - only glass. I missed that. Well that takes the fun out of doing a fun little test...

And as to our previous discussion, it is probably not float or crown glass, but something a bit more fancy.

Did you see this web page - Dead Link Removed
They say that Tiffen makes a similar filter and B&W as well - B&W 403. They show a Schott UG1 that looks perfect for this test as well.

I haven't done it, but check around www.edmundoptics.com web site too.

Kirk

 

[gainer]

Patrick - I would suggest using a high cal point that is a bit higher than 0.6. You're using an ND filter right? Glass or gel? The smaller Wratten filters are allowed to vary density within +/- 5% across the surface of the filter. Also, as you are calibrating at the bottom end of the practical range of densities you will encounter in film testing (0.6 is only two stops of density), you may have some larger errors for readings outside of your calibration range. Remember it only takes a little error in the setting of your high point to extrapolate to a larger error farther out - it tweaks the slope of the response line. "
I can always cut the 0.6 and double it or triple it. That is probably the best way to test linearity of the log function.

"It's best to have calibration points that "bracket" the range of values that you wish to measure. And since your diode has a pretty linear response, you can be reasonably certain that the values in between the cal points fall on your calibration curve. (But it is always best to use a check standard to verify!)"

I hate to say this, but you are wasting your time telling me what I have known for years, and have learned not to worry too much about. In order to use these precise measurements, you must know more about your materials than you can know. You and I are never certain that we are measuring the effect of a developer or the effect of the variations of the film.

"If you like, I can send you a copy of my Stouffer Step tablet. I sure I've got some that will have densities above 2.0, probably more. Of course I can also "calibrate" it by reading the RGB-Vis densities with my densitometer. Let me know if you are interested."

I have and have had a number of Stouffer tablets, calibrated and uncalibrated.


"So you make adjustments so the your meter reads in "steps" instead of actual optical density?"

No, the meter reads in zones according to custom, but because of the non linearity of paper, it cannot read all zones exactly right. It can be set, however, to read any 2 zones of a negative and assure me that those zones will print as such on my paper.


"Well, that indirectly relates to this thread. And I'm sure Sandy would remind you of the complications involved in reading stained negs, especially when the spectrum of the measurments is not controlled. "

The spectrum of the measurements is not what controls the response. It is the part of the spectrum transmitted by the negative that the paper is able to "see" and has no direct bearing on what we measure.

"Sure, you could adjust it to read "effective stops" at two points, but I think the non-linear printing response of stained negs (at least on VC paper) will give you wacky results for the zones in between 2 and 8. Especially if you have no filtration on your densitometer - since it is going to be more red-sensitive than a densitometer than one that has a blue filter on it. "

You are telling me I might be able to do things I have already done. Are you also telling me that knowing the exact densities of a negative in any and all parts of the spectrum will somehow cause the paper to respond obediently? The best you or anyone can do on a straight print is to set two printed densities where you want them, unless both the paper and the film are linear, or just happen to complement each other's non-linerities. Measurements of anything do not mean a whole lot unless they are usable. We can use them to test theories or to make prints. Prints are seldom used to test theories, and theories do not always make good prints. Now I'm telling you something you already know.

 

[Kirk Keyes]

Also check out the Hoya U-360, U-340, or U-320 filters.

 

[Kirk Keyes]

Sandy - Here you go:

http://www.bhphotovideo.com/bnh/controller/home?O=productlist&A=details&Q=&sku=8207&is=REG

That's cheap enough I might get one to play with!

Kirk

 

[Kirk Keyes]

Look for a Tiffen "Hot Mirror" filter to go with it for blocking the IR. Under $50 for sizes up to 55 mm.

 

[sanking]

Sandy - Here you go:

http://www.bhphotovideo.com/bnh/controller/home?O=productlist&A=details&Q=&sku=8207&is=REG

That's cheap enough I might get one to play with!

Kirk

According to information at the B&H site the 403 filter completely blocks wavelengths longer than 360nm.

I am curious why you believe this is good filter for the kind of testing I discussed given the fact that the peak reading in UV mode of the Gretag D-200-II is 373 nm, with a bandwith of 45 nm?

Sandy

 

[gainer]

I think I have cooled down a little.

Let me reiterate something I think I said a long time ago, but maybe not in this forum.

If a densitometer has a truly logarithmic response to input current, i.e. if its output follows the equation:

Eout = k log(Iin)

and we make the constant k = 1 by adjusting the gain, then:

Eout = log(Iin)

Now you see there are no requirements placed on the source of the input current. It could come from a photosensor or it could come from some other source. The meter's output will still be the logarithm of the input current. In our case, we would like to have the input current be a linear function of the illumination on the sensor. Phototransistors are very linear over a long range when used in the current mode, without a load resistor.

Suppose then that my phototransistor has a transfer funstion like:

I = K*L

where L is the illumination of the cell.

When this current is used as input for the meter, the output reading is:

Eout = log(K*L)

or Eout = log(K) + log(L)

But log(K) is another constant that may be set to zero by the level adjustment on my meter.

Consequently, whether I use a step density wedge or a precise arrangement of resistors and a regulated power source, the slope of the output vs log(input current) equation will be the same. The only change will be in the intercept of that line. Hard as this may be to swallow, it is true.

Nowhere in this derivation is there any requirement placed on the color of the illumination. If I measure in a part of the spectrum to which the photosensor is either more or less sensitive, I will find a shift in the calibration curve, but not a change in slope.

The questions are, how and what do we measure? The questions are answered one way if for some reason we want to know the spectrum passed by a pyro developed negative or step wedge, and another way if we want to know how to estimate paper grade and exposure. In the first case, the H&D curve (I still prefer to honor the men who got us started) would have to be three dimensional, with the spectrum plotted at each point on the log relative illumination axis. Then we would know the spectral content of a pyro-developed piece of film. In the second case, we would want to know the spectral response of our printing material as well as the content of each part of the negative we want to print. I doubt that any of us are willing to undertake all the experiments and concatenation of transfer functions that would be necessary. It will not be done simply by taking densitometer readings. If you are making contact prints by a printing-out process, I'll bet you will still lift up a corner now and again to see if it's done.

 

[Kirk Keyes]

Sandy writes, "According to information at the B&H site the 403 filter completely blocks wavelengths longer than 360nm."

"I am curious why you believe this is good filter for the kind of testing I discussed given the fact that the peak reading in UV mode of the Gretag D-200-II is 373 nm, with a bandwith of 45 nm?"

Sandy - did you look at this link I gave you on the previous page?

Dead Link Removed

Again, I say "They say that Tiffen makes a similar filter and B&W as well - B&W 403. They show a Schott UG1 that looks perfect for this test as well."

Appearantly you did not. From the graphs they show on that web page, it looks to me that the B+W 403 may actually be a better match to your Gretag UV filter than the Wratten 18A.

I like B&H a lot, but I'm not sure they are a better source of scientific information than a web site devoted to Medical and Scientific Photography. Perhaps B&H has a typo on their page. I hope that they at least got the price right...

Look at the other link I gave you and then please let me know what you think about that filter.

Kirk

 

[sanking]

Sandy - did you look at this link I gave you on the previous page?

Dead Link Removed


Kirk

No, I missed the link on first reading of your message. The charts do indeed suggest that the 403 would be a good inexpensive substitute for the 18A.

Sandy

 

[Kirk Keyes]

Patrick - I'm sorry if my comments upset you. That was not my intent.

Patrick wrote, "Are you also telling me that knowing the exact densities of a negative in any and all parts of the spectrum will somehow cause the paper to respond obediently?"

No. But it may lead us to making some measurements that may allow us to predict how the paper responds.

"We can use them to test theories or to make prints. Prints are seldom used to test theories, and theories do not always make good prints. Now I'm telling you something you already know."

There is an entire science called sensitometry that does this. And we are attemping to scratch the surface of it ourselves here.

"Let me reiterate something I think I said a long time ago, but maybe not in this forum. If a densitometer has a truly logarithmic response to input current, i.e. if its output follows the equation: Eout = k log(Iin) [...]"

I believe you did say this in this thread. And I even understand the math and electronics behind it.

"Nowhere in this derivation is there any requirement placed on the color of the illumination. If I measure in a part of the spectrum to which the photosensor is either more or less sensitive, I will find a shift in the calibration curve, but not a change in slope."

I'm not saying that your densitometer's slope will change with wavelength - otherwise it would be impossible to build things like spectrophotometers.

I'm saying the density of the materials you are trying to measure may change with wavelength. And sometimes significantly. That's why it could be important to control both the souce of illumination as well as the spectral response of your densitometer.

If your densitometer is responding to light from blue to near IR, without some sort of control of the light source and the sensor response, how will it be able to give us an accurate measure (say +/- 0.02D) of the density on a piece of stained film? How will it tell us about the UV absorption? I can't, as it is being influenced by all the other wavelengths that are striking the sensor - it is integrating the densities of all the wavelengths into one number. That is what my response to Donald's question was about.

And what about making measurements on a something that is obviously not neutral - say a piece of color transparency film. Without controlling the light source and the sensor, any measurements made would not be useful for much.

Patrick - I understand you are concerned about what wavelengths all these measurements are made at - it's a valid concern. That's why at this point, we should be using standardized materials for our density readings, things like Status filters or filters with specified properties, like the ones in Sandy's Gretag densitometer. At least that we we have some way to try and make comparable numbers. And there are probably better suited filters that could be used for stained negs, but until that has been determined, we must use what we have.

"In the first case, the H&D curve (I still prefer to honor the men who got us started) would have to be three dimensional, with the spectrum plotted at each point on the log relative illumination axis."

Yes, I've come to that conclusion as well. I really would like to see a graph of that...

And I've got no problem with honoring Messrs. H and D. - please do, we owe them a lot.

"Then we would know the spectral content of a pyro-developed piece of film. In the second case, we would want to know the spectral response of our printing material as well as the content of each part of the negative we want to print."

I believe that point is what got this entire thread started.

"If you are making contact prints by a printing-out process, I'll bet you will still lift up a corner now and again to see if it's done."

Yeah, that's why I don't like printing out papers...

Kirk

 

[gainer]

Okay. Now my point is that I have a meter that reads out the log of its input current. This kind of meter is well suited to use with silicon sensors in the current mode, both for linearity of photocurrent with illumination and for response speed. I used photoresistors at first, but the risetime of a change in reading was too long. Furthermore. they also had problems with the spectral response.

The available phototransistors have changed since I built my contraption, I'm sure. I have heard of silicon blue cells that have increased response in blue or decreased in red, I don't know which. as I said, I would like to have a flat response, because then I know what my filters are actually doing. We had a light meter at NASA that none of us could afford that had a flat response. The sensing element seemed to be about the size of a nickel and would have been of little use as a densitometer for our purposes anyway.

You may have missed the point that I have both slope and intercept adjustable. There are scales on the pots so I can return to any previous setting. I have one setting of slope for measurement of density, and several others for different contrasts of printing paper. The intercept setting varies with paper speed when I am set to read Zones, and with the maximum light level when I am reading density. In my next version, I am going to use 10 turn precision pots which can be set and reset to 1/100 of a turn for the adjustments. Another point I want to stress is that the setting of the slope of the meter response for reading density can just as well be done with a pot, some resistors and a regulated power supply. In fact, that part of the calibration will be built into my next version.

The secret to using this meter to set exposure for projection prints from stained negatives is to make a test print based on zone 3 where the stain is least. Now I look at the print and see where one of the higher zones lies. I then read that part of the negative and set my slope and intercept if needed to make those two parts of the negative read the two zones that I observed. Now I am pretty well set for deciding which grade of paper to use. Heaven help me when I go to VC!

 

[Kirk Keyes]

Patrick wrote, "The available phototransistors have changed since I built my contraption, I'm sure. I have heard of silicon blue cells that have increased response in blue or decreased in red, I don't know which."

I don't think it's that a Silicon Blue diode has an increased response to blue so much as they have added a filter to reduce the red response. Here's a blurb from a Perkin Elmer data sheet: Blue "series devices have a built-in infrared rejection filter for those applications where a detector is needed that approximates the human eye. Typical transmission of wavelengths greater than 750 nm is less than 3% when measured with an incandescent source operating at 2850 K."

Check out the Absolute Spectral Resonse graph with the filtered Absolute Spectral Response graph on this data sheet.
Dead Link Removed
The filtered diode does look pretty close to the response curve that one finds for human vision. A sensor like this would be good in a densitometer that was trying to approximate the resonse of the eye.

"...as I said, I would like to have a flat response, because then I know what my filters are actually doing."

That's true. However, if we are using filters to examine a (relatively) small range of wavelengths, then the response curve of the sensor will become negligible as it is the filter that will be in control of the wavelengths for our density measurements.

"The secret to using this meter to set exposure for projection prints from stained negatives is to make a test print based on zone 3 where the stain is least. Now I look at the print and see where one of the higher zones lies. I then read that part of the negative and set my slope and intercept if needed to make those two parts of the negative read the two zones that I observed. Now I am pretty well set for deciding which grade of paper to use. Heaven help me when I go to VC!"

I'm amused by that last sentence, because VC papers are where things get hairy, and more interesting.

For graded papers, I find that stained negs print just like non-stained negs. I probably would not bother using a staining developer unless I was using VC papers. (And that's only because I dont need the extra density or a dual purpose neg like Sandy has designed the pyrocat to do.) Although I do think stained negs are lower in grain - so I guess I would stick with them afterall...

Kirk

 

[Donald Miller]

Kirk Keyes said:
For graded papers, I find that stained negs print just like non-stained negs. I probably would not bother using a staining developer unless I was using VC papers. (And that's only because I dont need the extra density or a dual purpose neg like Sandy has designed the pyrocat to do.) Although I do think stained negs are lower in grain - so I guess I would stick with them afterall...

My results are exactly opposite of your suppositions in this regard. The proportional stain exhibited by certain staining developers is counter productive to highlight tonal separation on VC materials. Most notable in this regard is the stain color of the PMK pyro formulation.

In fact the proportional stain is most effective with graded materials, I can not understand your statement about stained negatives printing the same as non stained negatives on graded materials. Have you done any testing of the reflection density in the zone VI and higher tonal regions on prints made from negatives developed in both types of developers? If you haven't I think that you may be surprised at the results of these density readings. At least my testing has indicated this.

Regarding your statement about less grain with staining developers, I have found that ABC pyro exhibits noticeably more grain then some non staining developers.

 

[gainer]

Kirk,

If I do not know what the response of my sensor is without a filter, then I do not know what it is with a filter, regardless of how much I know about the filter. True, I may know which wavelengths are not passed, but of the ones that are passed what gets through is a combination of the passband of the filter and the sensitivity band of the sensor.

 

[Kirk Keyes]

Patrick wrote "If I do not know what the response of my sensor is without a filter, then I do not know what it is with a filter, regardless of how much I know about the filter."

If you have calibration standards, you do know the response - or at least you can measure/calculate it.

 

[gainer]

OK then, tell me the procedure for determining the spectral response of an arbitrary photocell. I will want to know the amplitude ratio and phase angle at enough frequencies to enable me to do a convolution with another such response curve that describes a filter. What kind of calibration standards will do this or anything like it?

 

[Kirk Keyes]

Patrick, I think we are thinking in different directions. I've been trying to discuss design considerations of a densitometer - for testing stained negs. A a color densitometer essentially, so that we can measure the density of negatives at various wavelengths. You seem to be telling me that it can't be done. I don't believe that you actually believe that.

There are certainly a lot of people that are selling color densitometers out there.

And I don't think you are trying to convince me that these instruments can't function. They certainly do. They are constructed with care being taken to the spectral properties of the light source, the optical pathway of the design, the spectral properties of the filters used, and the spectral response of the sensor. These manufacturers also supply calibration standards that have been measured under specified conditions. Calibration standards are needed because no matter how ideal our design, electronic parts, or the materials used to built the instrument, it will not have a perfect response. And since the standards have been measured under specified conditions, they should allow us to be able to exchange measurements and then hopefully have a reasonable expectation of comparable results.

You appear to be headed somewhere else - so I guess I'm not quite sure where you are going with the last post.

Kirk

 

[Kirk Keyes]

Donald - I knew I should have gone into more details when I made that post...

First let me say that my comments were based on my experience with PMK, and almost exclusively with PMK. I have used ABC-Pyro, but that was nearly 25 years ago... So my comments in that post and this one here are intended towards using PMK (The concepts I've covered the last 10 pages of this thread could be applied to any staining developer.) And I would only apply my comment about the appearantly lower grain to my experience with PMK. I do understand that ABC-Pyro has a reputation for more grain.

Donald wrote, "My results are exactly opposite of your suppositions in this regard."

Which of my suppositions?

Donald wrote, "The proportional stain exhibited by certain staining developers is counter productive to highlight tonal separation on VC materials. Most notable in this regard is the stain color of the PMK pyro formulation. "

Donald, I agree that there is not better "highlight separation" when using PMK with VC papers. Not only is not better highlight separations, there is highlight compression. I did not say anything to the contrary about that above.

What I said was "For graded papers, I find that stained negs print just like non-stained negs. I probably would not bother using a staining developer unless I was using VC papers."

I would use PMK to NOT get better highlight separation - I would use it to compress highlights - by using VC papers. I was trying to respond to Patrick saying, "Now I am pretty well set for deciding which grade of paper to use. [By using his densitometer.] Heaven help me when I go to VC!" Again, I probably would not bother using PMK with graded papers, although it would give you the option of using a VC paper and then being able to compress the highlights. Kind of a dual-purpose option, again like the PyroCat with alt processes.

Donald wrote, "In fact the proportional stain is most effective with graded materials, I can not understand your statement about stained negatives printing the same as non stained negatives on graded materials. "

I guess I need a clarification of what you mean by "most effective". Perhaps we have different goals in mind with our choice of developers. And see the response to the next paragraph -

Donald wrote, "Have you done any testing of the reflection density in the zone VI and higher tonal regions on prints made from negatives developed in both types of developers?"

Yes, I certainly have. I found that once you get past the issue of trying to match development times, the shape of a PMK neg printed on a graded paper looks very much like the shape of a non-stained neg printed on a graded paper. The curve has a nice, straight-lined portion in the middle. As far as the graded paper is concerned, the stain looks just like silver. The graded paper can't tell the difference.

When printing a PMK neg on VC paper, the highlights are compressed and then print darker than a similar exposure would have if a non-staining developer was used (or if the same PMK neg was printed on graded paper). Instead of a straight-lined portion to the curve, you have a curve that is essentially all toe until you get into the darker zone numbers. The toe of the curve on the print is very long.

I'm not the only one that has found this - it is exactly what Howard Bond found in his second, much better researched article on PMK that was just published in the Sept./Octo. 2004 PhotoTechniques magazine. See his Figure 5. I get the exact same results.

Are you finding results different from this?

Kirk

 

[Donald Miller]

Kirk Keyes wrote:
First let me say that my comments were based on my experience with PMK, and almost exclusively with PMK. I have used ABC-Pyro, but that was nearly 25 years ago... So my comments in that post and this one here are intended towards using PMK (The concepts I've covered the last 10 pages of this thread could be applied to any staining developer.) And I would only apply my comment about the appearantly lower grain to my experience with PMK. I do understand that ABC-Pyro has a reputation for more grain.

I understand your suppositions based on your use of PMK. I have found that other staining developers have different characteristics when used with VC materials. I have found that Pyrocat is neutral insofar as any compression or expansion effects of the highlight tonal scale when used with VC materials. This is based in my testing and experience.

Kirk Keyes wrote:
What I said was "For graded papers, I find that stained negs print just like non-stained negs. I probably would not bother using a staining developer unless I was using VC papers."

I would use PMK to NOT get better highlight separation - I would use it to compress highlights - by using VC papers. I was trying to respond to Patrick saying, "Now I am pretty well set for deciding which grade of paper to use. [By using his densitometer.] Heaven help me when I go to VC!" Again, I probably would not bother using PMK with graded papers, although it would give you the option of using a VC paper and then being able to compress the highlights. Kind of a dual-purpose option, again like the PyroCat with alt processes.

My results show that proportional stained negatives do print differently then conventionally developed negatives with graded papers. My tests have indicated this to me. This especially true when I print on Azo. In which case not all films will build the density range required of this paper without using staining developers. The proportional stain acts as additional density in the upper density regions.

I can't understand anyone wanting to compress the highlight tonal scale in a print. That is the reason that I use proportional staining developers...to effectively gain increased separation. It would appear that if compression is the desired result that reduced development would suffice.

 

[sanking]

My results show that proportional stained negatives do print differently then conventionally developed negatives with graded papers. My tests have indicated this to me. This especially true when I print on Ago. In which case not all films will build the density range required of this paper without using staining developers. The proportional stain acts as additional density in the upper density regions.

Donald,

I don't think you and Kirk have any fundamental disagreement. He is saying that with graded papers stained negs print just like non-stained negs. You are saying that the proportional stain acts as additional density in the upper density regions. Both statements are essentially true.

What one can say is that when developed to the same effective printing contrast, as measured by the spectral response of the process, stained and conventional negatives print virtually the same. I am not going to say exactly the same because there is one area where I find that even with graded papers tanning/staining developers give better results, and that is when printing strongly backlit scenes. Since development with tanning/staining developers is mostly at or near the surface there is less light scattering than with conventional developers, and this results in keeping detail even in extreme lighting conditions, say when photographing a tree with leaves against a very bright object like the sun.


Sandy

 

[Kirk Keyes]

Sandy wrote, "I don't think you and Kirk have any fundamental disagreement. He is saying that with graded papers stained negs print just like non-stained negs. You are saying that the proportional stain acts as additional density in the upper density regions. Both statements are essentially true."

Sandy, I do agree with everything said in that paragraph.

Sandy wrote, "What one can say is that when developed to the same effective printing contrast, as measured by the spectral response of the process, stained and conventional negatives print virtually the same. I am not going to say exactly the same because there is one area where I find that even with graded papers tanning/staining developers give better results, and that is when printing strongly backlit scenes."

I also agree here - please note that I said that "the shape of a PMK neg printed on a graded paper looks very much like the shape of a non-stained neg printed on a graded paper." I can not say that they are identical, but they are very similar.

Sandy, thanks for giving an example that one would not observe from printing step wedges but only from shooting real world subjects! It is easy to get wrapped up in testing stuff and forgetting the ultimate application. I'm not saying we should not do testing, but it is good to remember that the tests we usually do, do not always tell us everything about our materials.

 

[rwyoung]

I've used logarithms a LOT, having been in the "pre-PC" era.

... snip ...

Wow... what a memory exercise this has been. This is equal to trying to remember how to calculate square .. or cube roots - by hand!!

Searching for other information came across this old thread and couldn't resist. Ed, and I'm sure you do remember this,

(any base log would work but I'll just use 10 for this example and ^ for "raise to the power")

log10(X^Y) = Y * log10(X)

So to find the square root of 99,

99 ^ 0.5, now get out the tables or slide rule:

Do the log : log10(99) = 1.996
Multiply for the power : 0.5 * 1.996 = 0.998
Do the antilog to get the result : 10 ^ 0.998 = 9.950

Cube root of 99? 4.626 by same method...