Film Testing (WBM) Curves 320TXP in XTol 1:1 Too Contrasty—How Did I Mess Up?

Just like the image contrast at the film plane will be lower due to flare, so will the image contrast on the paper be lower due to enlarger flare. Although the paper LER for grade 2 is right on 1.05, an NDR of 1.05 will not at all compensate for any loss due to enlarging flare. So, bumping the aim NDR up to 1.2 compensates for at least a half stop loss of contrast due to flare losses at the enlarging stage---------a full stop loss of contrast at the enlarging stage would require an NDR of 1.35. But I've no idea exactly what the flare loss is from the enlarger, but bumping the aim NDR up by some amount to help compensate for it is probably wise.

Rafal Lukawiecki said:

I am puzzled by how the flare/non-flare computations arrive at almost identical aim CIs: Stephen's practical flare model, and the WBM gradients, mentioned on page 212, especially as there is one other major, I think, difference in approaches. Stephen uses the generally accepted LER of 1.05 for Grade 2 paper. This, by ISO definition, only uses a portion of the available scale of a paper, notably not exceeding 90% of the available DMax. WBM carefully points out that "log exposure range of grade-2 paper is limited to 1.05, but this ignores extreme low and high reflection densities." and it suggests "We have no problem fitting a negative density range of 1.20 onto grade-2 paper, if we allow the low end of Zone II and the high end of VIII to occupy these paper extremes." WBM uses the figure of 1.2 in the calculation of aim average gradients on page 213.

Click to expand...

Numbers don't mean anything on their own. To clarify everything above, all you need to do is put it all into context, and this is done by applying a little theory. First let's look at the paper LER. As with the film speed point, the 90% of D-max of the paper is just a point of measurement. Remember what I said about needing to define the limiting points of quality on the film curve? It's the same principle here. Loyd Jones found that after 90% of a paper's D-max perception of tonal separation drops off (it's never about density. It's about tonal separation). He never suggested that this is the usable range of the paper. In fact, the graph below is from the Theory of the Photographic Process and it illustrates this concept using a tone reproduction curve. The curve is generated by comparing the resulting print with the original subject.



The average luminance range is 2.20 or 7 1/3 stops, but this doesn't represent the entire luminance range of the average subject. There usually are small areas of specular highlights and "cavity" shadows. The graph brings this out to a luminance range of 3.0 or 10 stops. So it's easy to understand how the ISO LER for a grade two paper is 1.05 even though a paper can handle a negative with a greater negative density range. The ISO LER of 1.05 only represents the 2.20 part of the luminance range.

Now if the paper can handle a 1.20 density range (paper D-max and D-min), and the film is processed so that the 2.20 range equals a density range of 1.20, then there's nothing left over for the specular highlights and cavity shadows. So how can Ralph suggest a negative density range of 1.20 for a 7 stop luminance range? The reason has to do with how the curves are measured.

Gradient is defined as rise over run. The luminance range is the run and the negative density range is the rise. 1.20 / 2.10 = 0.57. The standard model uses a luminance range of 2.20 and a negative density range of 1.05. The difference is that it factors in a flare factor. The subject luminance range is reduced at the film plane by veiling flare. Whatever the actual range of the subject might be, the determination of the degree of processing is based on the optical image at the film plane. Average flare is around 1 to 1 1/3 stops. Kodak uses 0.40 or 1 1/3 stops of flare in their calculations. 1.05 / (2.20 - 0.40) = 0.58.

Both methods produce almost identical gradients. The gradients of 0.57 and 0.58 are ideal for producing quality negatives printed on grade two glossy paper using a diffusion enlarger. As gradient is input to output, if the gradient values are the same, the same input will produce the same output. How then can the theoretical output numbers not be equal? Simple. One uses flare in it's calculation and one doesn't. Mathematically, In order for the no flare method to produce the target gradient, the rise must be higher. For a gradient of 0.57 / 0.58, a density range of 1.20 is where a seven stop illuminance range (note: not luminance range) will fall, but we are not processing the film for a 7 stop illuminance range but for around a 6 stop illuminance range. At around the 6 stop illuminance range, the negative density range falls around 1.05.

And I have not even thought much yet about the impact of enlarger flare.

Click to expand...

There is a chart in Photographic Materials and Processes that has the aim LERs for printing with a diffusion enlarger and a condenser enlarger. For printing on a grade two paper with a diffusion enlarger, it has the aim LER as 1.05. For a condenser enlarger it has it as around 0.80. Mind you this is for the same paper. As the aim numbers are for negatives printing with enlargers, we can reasonably conclude that the aims take into consideration some degree of flare. The reason why the diffusion enlarger LER aim is the same as that derived from a contacted test is because there is a slight Q-factor with the diffusion enlarger that tends to compensate for average enlarger flare.

Stephen's post illustrates why I say it is important to follow your teacher.

He shows the two different calculations that come up with almost the same result.

1.2 NDR is the right target when working with the techniques described in "Way Beyond Monochrome," while I found 1.0 NDR to be the right target the way I work.

---

The Kodak Data Book, Kodak Papers (first 1955 printing), says you should select a paper with scale index 0.2 greater than the Density Scale of Negative and offers an explanation: "The reason for this appears to be that the method of measuring scale index takes into account the total available scale of the paper, while in practice it is desirable to use less of the extreme shoulder portion of the paper curve."

In the data book, Density Scale of Negative is effective density scale at the exposure plane so there is no doubt that you are left to figure out your own enlarger flare.

The data book gives a chart, keep in mind this is effective density scale:

Grade 2 paper, Scale Index 1.3, Density Scale of Negative 1.0 to 1.2

Further the data book says the densitometer would match a contact print. Diffuse enlarger lowers scale 10% or more, condenser enlarger with a diffuse bulb raises scale 15-25%, point source condenser maybe raises scale 60%

I think just a few sheets with your new setup will get the sensitometry exposure in a useful place. With the 10-stop range on the Stouffer wedge, really any exposure placement that gives you 7 stops' worth is useful. I like to get the toe down near base+fog on one of the last few steps, that gives a long run and graphs the toe nicely. Absolute accuracy is only required for the ISO tests the manufacturers have to follow. For home and pro photographer, consistency such as 1/10th stop is remarkable - 1/3 stop would be better than we need.

Instead of trying to home in on a full statistically valid test with 25 sheets, I think you can get by with a single 5 sheet test. Maybe put two or three exposures on each piece of film - then you can average the readings. Spot check yourself occasionally for consistency. Run a test sheet every once in a while with your regular film runs. I update my Time/CI chart with the results (dots and date stamps creates a "scatter" diagram that tells how good the predictions are).

Then run another 5 sheet test using a compensating developer. You will then be armed with two curve families, one family using a standard developer and one family by compensating developer. You will be able to look at the compensating developer family curves to choose times to develop N-3 or N-2. Even N-1. Look at what shape you would like the whole curve. As Michael R 1974 reminds us, you may be better off developing longer and printing on Grade 0 - choose the most attractive curve - not necessarily the one that fits Grade 2.

Flare is an issue we debate, that's why there are 300 posts on it. I'd focus on identifying "it" and checking where my procedures correct for "it" so that I know I correct exactly once. If a correction for flare is buried in your methods, it's still a correction. So the results can be identical. Just don't want you to correct for flare twice.

The negative density range as derived from the paper LER is more of an aim than an absolute. There isn’t a perfect correlation between the two. Popular photo books aren't big at caveats and tend to simplify this relationship which can lead to the conclusion there is. Serious sources have never claimed a perfect correlation. Jones, who developed the method, in one of my favorites of his quotes wrote, “The procedure followed in obtaining a relationship between DR and LER may seem forced and artificial. This we grant, and it must be borne in mind that the print quality obtained by its use will not be the highest quality. But what other course is there to follow? Either we must make the best of a somewhat imperfect relationship or face the prospect of having no criterion whatever for choosing the paper contrast grade.”

His conclusion shouldn’t surprise anyone. It’s not that psychophysics isn’t an effective science. It’s just that there are too many variables involved with the determination of print quality to fully express in a few digits. Here’s a graph from some of Jones' tests.



The graph shows the DR of the negatives, horizontal lines, printed on a grade two paper that were judged as high quality prints. They don’t all fit neatly within the defined limits of the print. The type of subject matter as well as the amount and distribution of the tones are a couple of factors that makes it impossible to predetermine print quality with any precision.

But of all the various methods Jones explored, this one tends to produce better results in the greatest number of circumstances. It gets you close enough to work with. It's statistics. I don’t know about everyone else, I only develop in one stop increments, and I don’t believe most scenes are considerate enough to have luminance ranges in perfect one stop increments. Even if there was a perfect correlation between DR and LER, it would be unrealistic to believe it could be achieved in most cases because of all the variables involved. Flare along would make this impossible.

There is another inconsistency with the DR / LER relationship that most books don’t discuss and that Michael has observed. The DR / LER relationship chances with the scene’s luminance range. Jones writes, “for the soft papers, the density scales of the negative (DR) should in most cases exceed the sensitometric exposure scale of the paper (LER), whereas, for the hard papers, the density scales of the negatives should in most cases be less than the sensitometric exposure scale of the paper (LER).” In other words, it’s not only acceptable, but preferable to have a slightly contrastier negative for contrasty scenes and a slightly flatter negative for flatter scenes.

Sure it's not perfect, but what other course is there to follow?

New Test, New Data for Kodak 320TXP and Ilford HP5+

I have just completed a second test of 320TXP and a first one for HP5+, a film I am quite familiar with, yet which I have never tested formally. I am attaching WBM spreadsheets for both of them. As I have been reading about sensitometry, both here, thanks to your detailed posts, and in books, I am getting a better feel for the tone reproduction process, and the data that supports it. I am, at last, starting to gain a clear understanding of my materials, which is a wonderful feeling.

I will plot the curves myself, and I will try to arrive the the gradients/CIs by hand, as I am not sure if the 0.17–1.37 DR method is the one I can easily relate with my book reading. I am also calibrating my enlarger light source-paper-developer combination, and soon I will be able to use an target DR based on that analysis (plus 10-15% for average flare, I suppose).

In the meantime, have a look, if you are still following this thread, at the numbers below. These tests were performed using a contact technique, with a flash above an opal plexi diffuser (shown on (there was a url link here which no longer exists)). I've exposed 3 sheets for each film and developing time, and I am amazed that the consistency of the flash exposure was so good. Sheets do not vary by more than 0.02 at any point (sheet-to-sheet), and, in fact, they usually read the same density for most steps (sheet-to-sheet), except for the steps 1-2, where they may vary by 0.04 at most. A test sheet exposed with an ND filter varies only 0.04 across extreme corners, so the illumination was OK for the purpose of this test.

It has not all gone very well, though. I have clearly underexposed 320TXP by a 1.3 stops, pity... Also, I have mistimed the 5.5 min development batch, as I had an incident with my timer. I think that batch is closer to 6 min, and I have tried to correct for that on the "Summary" page of the spreadsheets.

Overall, it would seem, using the WBM spreadsheet method only, that the dev times for HP5+, within 30 sec of what I am used to, are a very reasonable: N-1 at 8.5 min, N at 12 min, and N+1 at 16 min, using XTol 1+1 20˚C. I am very pleased to see those results, even though I have reduced agitation from 4 to 3 full inversions every 30 sec (plus 30 sec at the beginning). I suppose that accounts for the difference.

As for 320TXP, I do not have experience of that film, so I am not sure how to trust the dev times, bearing in mind the lack of full data due to underexposure. The spreadsheet seems to suggest: N-1 at 5 min, N at 7 min, and N+1 at 11 min. All of those use the target gradients suggested by WBM, which I remember Stephen agrees with, too, concerning the aspect of flare.

I am curious what aim DRs will I get when I measure my paper ES.

Many thanks to all for guiding me.

Hi Rafal,in those excel files you have omitted all the other worksheets which plot the curves. I am curious to see them.

I went through exactly what you did with WBM and spent ages getting it worked out and understood too ! I also had big flare issues that I solved . My last spreadsheet and film speed curve can be found in this thread (there was a url link here which no longer exists)on APUG

Hi Peter,

I just downloaded the attachment, and it shows 8 sheets, with the individual curves. I'm not sure why they hid from you. I can PM the files to you, if you would be so kind.

I'll review your thread now.

Many thanks,
Rafal

thanks Rafal. Feel free to pm them to me. I'll check them out in about 18-24 hours.

Thanks Rafal, I tried again and managed to see the curves. I note you don't have any data in the film speed curve for higher than N+1. I initially thought it was because you only developed for 16 minutes but I see if you added another point out at N+2 on the summary plots that you would have hardly any difference in film speed between N+1 and N+2. This is a consequence of your diml speed curve asymptoting. Additionally your film speed curve is concave up whereas mine is concave down for nearly the same combination of film and dev (I used XTOL 1+2 rather than 1+1).

Thanks, Peter. I was also a little surprised by the concave up, when I saw it, and I have no real explanation why this happened. I was wondering if, perhaps, some lack of evenness in illumination could be the reason for it, but a test sheet, exposed with ND, is pretty even.

Michael R 1974 said:

I guess what I don't understand about things like CI is, why "summarize" the curve with a straight line of best fit when you already have the full curve?? Why not simply plot the curves and compare them?

Click to expand...

In parusing this thread, I found this to be a striking observation...............this is precisely the sentiment made clear in The Negative.....points along a curve connected by a straight line are not touted in the The Negative as having any significance toward any ZS analysis of characteristic curves. It's the comparison of two curves in their entirety that provides the best evaluation of a film in a developer, so says the author. Gamma, which applies to "true" straightnline portions of any curve, is given more merit than points connected by a straightline. Whether this is agreed with or not is kind of beside the point, it's just an observation. I don't bother with CI myself, I just compare the curves, I find that to be most illuminating and the most straight forward.

Could somebody post the curves?

Stephen Benskin said:

Could somebody post the curves?

Click to expand...

Stephen, in post #68 of this thread I attached the full spreadsheet with the data points and the curves for both HP5+ and 320TXP. If the file does not open for you, however, I would be happy to extract the curves, please let me know.

CPorter said:

I don't bother with CI myself, I just compare the curves, I find that to be most illuminating and the most straight forward.

Click to expand...

I don't have your experience of reading the curves, as I have only started sensitometric analysis of my materials. I have no doubt, however, that you are right. Re-reading BTZS, plus many good points raised on this, and on other related threads, have opened my eyes to the holistic nature of the tone reproduction cycle. It is clear to me that subtleties of the cycle cannot be summarised by just a single gradient number.

Nonetheless, I found it necessary to figure out developing times for my N, N-1, and N+1 negatives. I found that knowing an aim DR, as described by BTZS, or a gradient, G-bar, or CI, all are perhaps crude, but practical ways to summarise the requirements of the printing process for the film development output. This was my main goal of this calibration exercise.

Rafal Lukawiecki said:

Stephen, in post #68 of this thread I attached the full spreadsheet with the data points and the curves for both HP5+ and 320TXP. If the file does not open for you, however, I would be happy to extract the curves, please let me know.

Click to expand...

Rafal, I'm not able to plot/find the curves, so if you can extract the curves, I would appreciate it.

Rafal Lukawiecki said:

It is clear to me that subtleties of the cycle cannot be summarized by just a single gradient number.

Click to expand...

And it was never meant to be. Just like the average driving speed in a trip doesn't specify the time spent at stop lights. It's an average. The average gradient, which includes the testing conditions, offers a general picture of the the way the film responds to exposure and development.

For any film curve, the average gradient for a 1.30 log-H range might be different for the average gradient if measured using a 2.20 log-H range. How it's measured is part of the answer. Think of the arguments for and against the various methods such as Gamma, Ilford's Average gradient, and Kodak's Contrast Index. And while we're at it, how much information about the film curve does a simple density range offer as some methods suggest using?

Once again, nothing's perfect especially when you are attempting to quantify a subjective response. And as with the NDR / LER relationship, what better course is there to follow.

Hi Rafal,

Haven't had time to plot your curves manually, but here's the graph paper I would use. The green lines correspond to your step tablet actual densities... I find graphing goes faster when I have green lines for each x-axis step...

http://www.beefalobill.com/images/sensitometry-RL1.pdf

The top "film speed" scale is not calibrated to your setup. Cut it off and slide it left or right to get an idea of your relative film speeds.

Found the other pages.