Paper speeds

[Nicholas Lindan]

There should be no bumps in the curves at all. They should all have straight line mid portions with no irregularities. See the D&Z article for full sets of curves.

VC papers resemble roller coasters. Graded papers are smooth if they are single emulsion. I saw the curves in PT and and I can only speculate the curves were generated with two or three measurements and a generic curve was then put through the points. The characteristic hump in VC papers is well known and well documented, some papers have horrid ones - Agfa MC Classic RC comes to mind - and all papers have a significant hump at the lowest grades.

The curves on the website are accurate and are repeatable to 0.05 stop exposure within a box of paper.

By the box run-run CV's are at the control of Ilford. From what I have seen Ilford maintains pretty tight control on paper characteristics, but when the regular supervisor goes on vacation ... Henry has reported a ~1/3 stop variation in paper speed for Ilfobrom when samples from 7 batches were compared.

Spend an hour with a step tablet, a 00 filter and a densitometer and you can see for yourself.

 

[Photo Engineer]

I didn't say that the results were bad, I said that VC papers should NOT be a roller coaster as seen in that set of curves. See the curves on page 19 of the D&Z article in the March-April issue of Photo Techniques.

I have spent hundreds of hours with a step tablet while designing products in person at Kodak and I know that the article is correct. A well designed VC paper will present a monotonic series of mid scale straight lines with varying slope as a function of contrast grade.

Some papers are better than others. Some batches of a given paper are better than others.

PE

 

[Nicholas Lindan]

I didn't say that the results were bad, I said that VC papers should NOT be a roller coaster

Should or shouldn't doesn't really matter: the curves for Ilford _do_ have humps in them and it is not hard to verify - spend an hour. Curves for graded Oriental don't have humps in them so humps aren't some product of the methodology. http://www.nolindan.com/da/support/osg2.jpg

What's more, from the look of Ilford's curves it looks like they are using three emulsions to try and get the low contrast curve smooth. At the high contrast end only one of the emulsions is active and the curve is smooth.

If you claim Kodak's curves for their variable contrast papers were marvelously monotonic with a curve as smooth as a baby's backside, well that's great, but sort of beside the point what with there not being much interest in Kodak photographic paper these days.

 

[Photo Engineer]

Nicholas; If a paper has a bump in it, it distorts the tone scale of the final print. I don't claim anything about Kodak papers, merely what is shown for 3 products in the article referenced. D&Z do not name the papers. They do point out the fact that some design methodologies are less good than others.

I do know that papers can be made with largely straight characteristic curves and to do otherwise when it is possible to get a straight line is not good design. The evidence for and against is also in the article.

Ilford may be using 3 emulsions, I would tend to agree but cannot be sure. The wedge spectrogram indicates 2 but the curve shows 3. It is impossible to judge from the wedge spectrogram which I have posted elsewhere, how many components are used. With sensitometric readings of the wedge spectrograms themselves, I might be able to tell whether it is 2, 3 or even greater. It could be 4 or 6.

Why argue about it. The bumps are fact, they are not the best in a paper, but I use Ilford myself. So, nothing more can be said regarding the quality of the paper. Nothing is perfect.

I can also state for informational purposes that there are 2 design philosophies for MG papers. One in which the curve swings around the mid density point giving the user the appearance of a constant exposure and speed when printing with various contrasts, and another where the speed varies with the contrast grade. The curves shown follow both of these somewhat, from what I can see. They combine both methods using one for one range of contrasts and another for the other range of contrasts.

This is just a matter of design philosophy, and procedure, it is neither good nor bad.

PE

 

[Photo Engineer]

I decided to add more here.

It is not uncommon to blend 2 or 3 emulsions in a paper of grade 1.0 or lower, but a grade of 2 or higher is usually a single emulsion. This is speaking of a graded paper.

Projecting this into MG papers, the same is needed to get the contrast range if one were to expose just one emulsion or the other or a mix of the two. However, since filters are not perfect, and there is 'punch through' of the blue light into the green sensitive layer, you get what is effectively 'cross contamination' of exposure and therefore bumps in what might otherwise be straight lines unless you are very careful and use the right emulsions.

The idea is to adjust speeds properly in each emulsion, and isolate exposure in each so that the curve shape increases or decreases in contrast without undue bumps. So, sometimes what can be achieved in the lab or in small scale runs turns out to be hard to control in large batches. Speed fluctuations such as described above of about 1/3 stop can cause this type of problem, especially if it takes place between emulsions in a blend. Typically, it is desired to keep fluctuations less than that. (BTW, in that comment is the 1/3 stop a total variation or is it + or - 1/3 stop? Is it between components in a blend or is it in grade 0 or is it in grade 5 or is it between the two? This is the type of problem I am referring to in manufacturing an MG paper.)

This will always be the case with MG papers. I like them and do split printing for the obvious advantages, but then again, I recognize the small imperfections that are introduced by using MG papers such as I saw in the curves above.

I would much rather prefer a 'perfect' negative and a graded paper. As you know, I am tailor making my own emulsions and have been able to actually tweak the Azo curve to give something just a tad better than the original product. This is not my opinion, but rather is those of the audience who have used the emulsion. The bottom line is that a tailored paper for a given print can be the best of all possible worlds.

To get back to speed fluctuations, I have been concerned in my research coatings at EK if speeds fluctuated by that much (I tried for less than 0.05 log E). In my hand made emulsions at home, I have seen virtually no fluctuation most of the time, in both curve shape and speed, but have seen it happen. The max here is about + or - 1/3 stop or less.

There is a good discussion of VC papers in "Post Exposure" by Ctein IIRC.

PE

 

[Nicholas Lindan]

> If a paper has a bump in it, it distorts the tone scale of the final print.

I think we are coming to agreement, if we ever disagreed. Most arguments eventually can be traced to misunderstanding. Yeah, the bumps are bad news.

> Ilford may be using 3 emulsions, I would tend to agree but cannot be sure.

Ditto. I am basing the assumption on two major humps with 3 'straight' lines.

> Why argue about it. The bumps are fact, they are not the best in a paper,
> but I use Ilford myself.

My conclusion also.

> I can also state for informational purposes that there are 2 design
> philosophies for MG papers. One in which the curve swings around the
> mid density point giving the user the appearance of a constant exposure
> and speed when printing with various contrasts, and another where the
> speed varies with the contrast grade.
> The curves shown follow both of these somewhat, from what I can see.
> They combine both methods using one for one range of contrasts and
> another for the other range of contrasts.

Good observation... The curves are a real tangle and it is hard to
sort them out.

The raw data is available in a spread sheet - Quattro Pro 8 - at
http://www.nolindan.com/da/support/

It can be translated to Excel but only the numbers can go across,
Excel hasn't got a linear interpolation function [@linterp()]

 

[Jim Noel]

MOst manufacturers list the ANSI speeds for their papers in their technical information. These are the best indicator of speed.

I will look later, after I cook dinner, and see if I still have some of the old ANSI speeds listed. I know I do not have them for Kodabromide or Polycontrast because I never liked those papers.

 

[Nicholas Lindan]

... However, since filters are not perfect, and there is 'punch through' of the blue light into the green sensitive layer, you get what is effectively 'cross contamination' of exposure and therefore bumps in what might otherwise be straight lines unless you are very careful and use the right emulsions.

You can really see this if you use Kodak Polycontrast filters with Ilford paper, the results are awful, and I am sure the same applies to generic filters from Freestyle and old Agfa and Varigam filters.

Bringing up an interesting point:

This may be the cause of the ephemeral difference between split-grade and filtered printing. If narrow band light sources, as the ZVI blue/green LED VC head or a dichoric head with narrow passbands are used the 'punch through' will be less(?) than it is with dye-based gelatine filters.

I have looked for spectral curves for poly/multi/vari contrast filters to no avail. I don't know how well the filters seperate the spectrum: how closely does the spectral curve of combined #00-#5 exposure compare with the spectral curve of a, say, #2 filter. Time to dust of the old spectrophotometer, or maybe not -- an old hand-cranked monochromater, it takes hours to plot absorbance curves.

It may be interesting to compare curve -shape- with respect to humps and bumps with various VC light sources. The ANSI paper grade specification ignores what happens in the middle of the curve so what should be a perfect match of paper grades with dichoric heads most likely isn't.

 

[Photo Engineer]

With a B&W VC paper, it is virtually if not completely impossible to separate the blue speed of the green sensitive emulsion(s) from the blue speed of the blue sensitive emulsion(s). With color, it is trivial.

The spectral characteristics of the filters are important. I find that my Dichro 45S gives an MGIV grade of 2 with a +30M for example. However, if there is any blue sensitivity to the green layer, then regardless of the cut of the filter it will be virtually impossible to entirely eliminate the effects.

One of the ways to do this is to add a yellow filter exactly equal to the apparent speed of the green layer and call this the zero point and then add filters (M or Y) to go in either direction from this center. Unfortunately, determining that filter factor is just about impossible.

Again, with color it is simple and involves the use of the standard 50Y 50M starting filter pack for printing, as the orange mask of the negative offsets the remainder. In fact, that is the reason for the 50R offset in color printing, the 'punch through' of the blue light to the G and R layers.

PE

 

[Nicholas Lindan]

Most manufacturers list the ANSI speeds for their papers in their technical information. These are the best indicator of speed.

I don't know, my experience is they aren't worth much, which is why - up to now - enlarging meters have been not very useful if you are going after making fine prints without a lot of test strip mucking about.

I found that only by looking at the speed of the paper at all points along curve was it possible to get a darkroom exposure system to really work. The term 'paper speed' as used in the DA system is really a bit of a misnomer, but then so is 'F-Stop Timer' - makes as much sense as a 'gallon ruller'.

The same technique will work with any meter/timer combination. The PITA is calibrating the system. EM-10 meters are a problem since no two are alike and they drift badly with age. The Analite III meters are a lot better, though of course no match for a meter made by ..., where production meters match to each other within .02 stops so calibration data can be shared among photographers.

Darkroom printing should be like film shooting. The more I tested paper speeds and development strategies the closer I came to the numbers Kodak prints inside the yellow box. It should be the same for paper.

 

[Nicholas Lindan]

if there is any blue sensitivity to the green layer, then regardless of the cut of the filter it will be virtually impossible to entirely eliminate the effects.

I know I read it some time ago, but I obviously didn't _read_ it:

http://www.ilfordphoto.com/Webfiles/2006130201152306.pdf

Ilford uses 3 emulsions: Blue ; Blue with a bit of green [Bg]; And a third equally sensitive to blue and green - white light for all intents [W].

Whooo boy, as you said, trying to tease those apart is going to be nigh impossible.

I posted a graph of just the 00 grade of MGIVRC

http://www.nolindan.com/da/support/mgivrc00.jpg

And the shoulders - humps - of the three emulsions can be seen at OD's .45 (paper speed 7), 1.2 (9) and ~2.2? (12+) - the graph doesn't make it all the way to max black.

 

[Helen B]

...

It can be translated to Excel but only the numbers can go across,
Excel hasn't got a linear interpolation function [@linterp()]

It is TREND, isn't it?

Best,
Helen

 

[Nicholas Lindan]

It is TREND, isn't it?

That's what I thought, that's what a lot of people thought. Trend does a least squares fit putting a line through the whole data set and returning a point on the line.

Linterp interpolates between values in a table. Very useful for turning data 'inside out'. In this case the data is densities at a set of given exposures and what is wanted is exposures for given densities. Naturally none of the density readings falls exactly on a Zone system value so it must be interpolated between two measurements.

The best I can find is:

=(INDEX($B$2:$B$9,MATCH(D2,$A$2:$A$9))-INDEX($B$2:$B$9,MATCH(D2,$A$2:$A$9)+1))/(INDEX($A$2:$A$9,MATCH(D2,$A$2:$A$9))-INDEX($A$2:$A$9,MATCH(D2,$A$2:$A$9)+1))*(D2-INDEX($A$2:$A$9,MATCH(D2,$A$2:$A$9)))+INDEX($B$2:$B$9,MATCH(D2,$A$2:$A$9)) at http://www.mrexcel.com/board2/viewtopic.php?t=128440

Which doesn't work... And isn't it a lovely formula to edit for each entry in a table - Brief took care of that, though.

 

[Photo Engineer]

The curve for an older grade MG paper, used by Ctein in his book shows a much better straight line. The method that Ilford uses is different than that of what EK used, IIRC.

The actual curve that represents an H&D response by analog photographs is called a cubic spline. In film, it is D Log E and in Digital it is V Log E. (Voltage).

Both respond similarly due to the laws of physics.

PE

 

[Nicholas Lindan]

The actual curve that represents an H&D response by analog photographs is called a cubic spline.

Is that

- one cubic equation representing the entire HD curve

or

- a set of, say 10, cubic equations splined together to represent one HD curve

?

If it is the first then I now understand your observation that the center of the HD curve should be ‘linear’.

 

[Photo Engineer]

Due to the physics and chemistry involved, the spline results from the inherent properties of both digital and chemical photography. One curve represents a complete H&D response of the film or paper.

The film curve is multiplied by the paper curve on a point slope by point slope basis (using the derivative of the curve to obtain the slope) to give the final point by point density and slope of the resulting print.

This is why neg-pos yields better tone scale results than pos-pos, as there is only data compression in the print of neg-pos, but there is compression of data in both stages of the pos-pos. It is also how the bumps in either negative or paper turn up in the final print.

PE