Adventures in film characteristic analysis

radiant said:

I've read somewhere else that over-development doesn't really affect to toe and doesn't increase contrast in midtones. That is visible from the previous graph.

Click to expand...

This is mostly correct in respect to the toe.
Over-development usually has some affect in the mid-tones - otherwise "push" processing would have no real apparent benefit.

radiant said:

Anyone? This bugs me, it cant be this way ..

Click to expand...

Don't worry. It's not. Bill had some good advice to start with.

Determining development times using percentages, especially in a one size fits all manor is crap. In fact, the field of photography is full of this kind of rule of thumb crap. Sure, there might some kernel of reason for it, but even if it is derived from some actual conditions, no one ever communicates what those conditions are and the assumption tends to be the rule is universally applies. One of my favorite examples is how to determine Normal processing time. The rule is to take the manufacturer's recommended minus something like15%. Why? It could be from the development times from the manufacturers are based on the the statistically average Scene Luminance Range that will contact print onto a grade 2 paper. Contacting requires a similar negative as printing with a diffusion enlarger. Many photographers use a condenser enlarger which requires a negative with a lower density range. So they found they needed to develop the negative less than the published time; however, their conditions aren't communicated when they give advice. It becomes a universal matter of fact and is no longer just a situational adjustment.

Different film / developer combinations result in different average gradients for different development times. Good practice is to test an emulsion in a developer over a range of times. This produces a family of curves. You take the development time from each curve and the average gradient and create what is commonly known as a Time-Gamma curve, but can also be referred to by the method the gradient was determined: Time-CI Curve or Time-Average Gradient Curve. The curve plots the development time against the resulting contrast.



By determining a development model for your printing conditions, you can determine the printing time for any required contrast. For the record, determining a development model isn't as straight forward as it might first appear. As I said before, the curve is dependent on the film / developer / development combination. Here is the same film developed in two different developers.



The next example surprised me. APX 25 barely moved in Xtol.



And here are a bunch of different examples superimposed on top of each other. The horizontal line represents CI 0.58.



Determining a development time based on a gradient value is just the basic function of this type of a curve. Think about the curve as having two constants: velocity constant, which defines this rate of increase in the line and the other constant, gamma infinity, which s the maximum contrast obtainable from a given film/developer combination. it's rate of increase over time and where and if it reaches Gamma Infinity can be applied to which film developer combination is best for different needs and conditions.

To obtain the velocity constant, find the average gradient for either the entire curve or for smaller sections (for more precise calculations). The higher the film gradient, the higher the velocity. The higher the velocity, the less development time it takes to produce a change in the contrast. The velocity constant can have a higher or lower rate and same with Gamma infinity. Comparing the four possible combinations illustrates certain characteristics.

Velocity Constant ------- Gamma Infinity
High -------------------------------- High

Contrast builds quickly, proceeds at a high rate and contrast continues to build up to a high value.

Velocity Constant ------- Gamma Infinity
High -------------------------------- Low

Contrast builds quickly, but soon slows reaching its limit at a low value.

Velocity Constant ------- Gamma Infinity
Low -------------------------------- High

Contrast builds slowly and increases at a slow rate, but with time, it is able to build to a high value.

Velocity Constant ------- Gamma Infinity
Low -------------------------------- Low

Contrast builds slowly, increases slowly and levels off at a low value.

Films with high velocities reach normal contrast fairly quickly. The developer may have insufficient time to adequately develop the sub latent image in the halide crystal which can result in lower effective film speeds. A 400 ISO film could have an EFS of 100 in a high velocity film/developer situation.

A lower velocity combination would reach normal contrast with sufficient time to allow for more complete development of the shadow values. This will more than likely result in a film with an EFS closer to the ISO.

A film with a high gamma infinity makes for a versatile film. It is able to handle a greater range of pushing. However, when pushing for speed, a low velocity/ low gamma infinity film would work best. The longer development times will give greater speed potential and the lower gamma will restrict excessive contrast.

Films with a low velocity are good for less precise systems or for an amateur just learning to process film. The allowable error in development is higher for the same level of precision control than with a high velocity film/developer.
The changes in development times do not have as great an affect on the contrast of a low velocity film/developer as with a high velocity film/developer. A person can be sloppier with a low velocity film/developer without it necessarily showing.

But to determine what CI is required for what condition, you will first need a developmental model.

MattKing said:

This is mostly correct in respect to the toe.
Over-development usually has some affect in the mid-tones - otherwise "push" processing would have no real apparent benefit.

Click to expand...

Yes. I didn't write any. And I think these effects are visible in my previous graphs, I think.

Stephen Benskin said:

But to determine what CI is required for what condition, you will first need a developmental model.

Click to expand...

Thanks Stephen, a great write-up again. I completely understand. I actually haven't understood that film+dev combos have such a big variances on the CI. And mostly the developer. You need to have the correct developer to achieve wanted CI. It would really great to have database of film+dev combos and the velocity constants calculated.

I start to understand also why everyone should stick to maybe one film + developer combo because making developmental models is quite a job. Once one finds good combo, then just stick with it. I've been doing combo-search for years without understanding the effects behind all these and have been puzzled many times.

Also my suggestion to everyone to overdevelop a bit is based on my small experience on just few developers (and one enlarger). I've already seen that nothing crazy happens when I overdevelop heavily but that is probably because of low velocity combo I'm using, perhaps. Or maybe I'm just a contrast junkie

Actually now when I start thinking, our community darkroom enlarger has different look on prints than what my hope enlarger has. I've realized this before but haven't thought about it. And maybe by just a feeling I have found out that my home enlarger needs contrast in the negatives. Maybe positive thing here is that I have been able to notice good (or one that suits me) pattern from my work.

This discussion has opened my eyes a lot.

p.s. It will help me (I like) graphs of film characteristics in Log scale, based on readings taken to the hundredths precision (providing numbers in a text file helps too, loose numbers would clutter the forum) on both X and Y axes. Where you place the absolute meter candle seconds does not matter very much on the X-axis, if a range covering 3.0 can fit, great. If unknown, just put the toe to the far left… absolute mcs reveals itself on fresh major manufacturer (Kodak/Ilford) film when you meet the ASA parameters with a standard developer (ID-11/D-76). Not a standard but when 1 meter candle second is at the far right, then the toe of a 400 speed film fits the page (landscape letter). I like to know if base +fog has been subtracted, or if you provide base+fog separately.

Bill Burk said:

p.s. It will help me (I like) graphs of film characteristics in Log scale, based on readings taken to the hundredths precision (providing numbers in a text file helps too, loose numbers would clutter the forum) on both X and Y axes. Where you place the absolute meter candle seconds does not matter very much on the X-axis, if a range covering 3.0 can fit, great. If unknown, just put the toe to the far left… absolute mcs reveals itself on fresh major manufacturer (Kodak/Ilford) film when you meet the ASA parameters with a standard developer (ID-11/D-76). Not a standard but when 1 meter candle second is at the far right, then the toe of a 400 speed film fits the page (landscape letter). I like to know if base +fog has been subtracted, or if you provide base+fog separately.

Click to expand...

Good tips. I start to understand why characteristic curves are done with standard; otherwise you have no clue what you are looking at.

Bill, could this be a good graph?

The densities are translated from Stouffer (Scanned together with strips and then grayscale mapped to T2115 density).

X-axis is translated by multiplying stops 0.3.

I think it starts to look maybe.. correct? The mapping to actual densities was probably the key.

The toe is now clearly steeper on 12 min, as it should be. The 6 minute one is very linear actually. Highlights compression is enough far away. If exposing EI 1600 and developing 12 minutes the compression is actually two stops more up, so no problemos with that either.

And yes you can see why people push HP5, it gives that nice CI boost. 12 minutes is a bit heavy of course but it doesn't ruin anything.

And previous graph data zipped.

radiant said:

Bill, could this be a good graph?

The densities are translated from Stouffer (Scanned together with strips and then grayscale mapped to T2115 density).

X-axis is translated by multiplying stops 0.3.




View attachment 297574

Click to expand...

Thanks! I was trying to do that in my head so I will try to follow from that. I’ll estimate the x-axis and Time/CI after I get a look at it.

radiant said:

And previous graph data zipped.

Click to expand...

I did this quick and in Excel. I used a relative log-H and the density is the supplied density minus film base + fog. This is looking to me like a film curve.



One thing I found doing some simple coding is that while you will get an answer, it may not be the correct answer. When developing a program, you need to use examples where you know what the results should be and how it's supposed to look.

I think you might be trying to reinvent the wheel before completely understanding the concept. It appears to me you are at a point where you want to move to a new level but are can't let quite yet let go of old ideas (18%, pushing for speed). You might want to consider Craig's suggestion and check into Phil Davis' Beyond the Zone System. While I have the same reservations I have with any system and in particular I disagree some of Davis' and find his approach a little restrictive. Still, it covers the basics which are enough for most and his plotting program saves a lot of time and effort instead of working it all out on your own.

I drew up a graph from your numbers, I had to leave off a couple readings because of the scale of my paper.

You had a CI of 0.60 for 6 minutes and 0.77 for 12 minutes Rodinal 1+25

You are "just under" the tolerance of the ASA triangle with 6 minutes. So I cannot properly judge film speed but for now I dropped the scale at 400 right at where the curve crosses 0.10 above base plus fog.

Try 15 or 30 seconds more next time to try to hit the triangle.

The Time-CI curve is just an example, the parts of the Time-CI curve outside the two data points is completely made up.

You can get a more complete family by developing several tests at different developing times, some people like a series that tends to give a broad set of contrasts such as 4, 6, 9, 13, 17 minutes.

https://www.beefalobill.com/images/RadiantTime-CI.pdf

Here’s how the Contrast Index Meter looks adjusted to fit your 6 minute curve. Right at 0.60

Stephen Benskin said:

I did this quick and in Excel. I used a relative log-H and the density is the supplied density minus film base + fog. This is looking to me like a film curve.

Click to expand...

Hmm, how did you come up with that? And I have to show my stupidity byt what is log-H and why that is used, since the density is logaritmic itself?

Stephen Benskin said:

I think you might be trying to reinvent the wheel before completely understanding the concept. It appears to me you are at a point where you want to move to a new level but are can't let quite yet let go of old ideas (18%, pushing for speed).

Click to expand...

I'm just keen to see first hand how things work. And by doing that I have learned a ton already, which is pretty exciting. Also I wanted to see if I can build densitometer from just simple accessible things (if you consider 3D printing and using WS2812 accessible, but those are pretty normal today). So far it seems it is possible (and why not, really).

What comes to sticking to ideas - I don't know any other way to calculate lux-seconds for exposure that this "18% method". I believe that is pretty standard. I'm using 18% term as reference for you so you know how the film has been exposed. Also I've never used pushing to gain speed, actually. Because of my methods; salt prints and overall love in punchy photos, I've found that I need contrast in my negatives. Maybe this is just proving my self that overdeveloping is OK I know it gains contrast but what else horrible happens? That has been my question.

Bill Burk said:

I drew up a graph from your numbers, I had to leave off a couple readings because of the scale of my paper.

Click to expand...

Wow, this was nice!

Bill Burk said:

You had a CI of 0.60 for 6 minutes and 0.77 for 12 minutes Rodinal 1+25
You are "just under" the tolerance of the ASA triangle with 6 minutes. So I cannot properly judge film speed but for now I dropped the scale at 400 right at where the curve crosses 0.10 above base plus fog.

Try 15 or 30 seconds more next time to try to hit the triangle.

Click to expand...

Actually, I have something to remind from previous posts. The exposure is a bit off because of old tests. I made some tests when I was looking for correct exposure with 100,200,400 ms exposure for the 18% slot. Until I found out that I had mistake in my code and the correct time should be 125 ms. That is 1/3 stops underexposure. That means these strips are exposed at EI 318. Could this be the reason being "just under" ?

I didn't want to correct the time because I wanted to use the previous ones but now is a good point to fix the system.

I'm sorry I didn't remind of this, but I was in "does this even work" stage before this so my I didn't care about that. But I think this is even more great this way, you probably found this issue by plotting? Amazing.

Bill Burk said:

The Time-CI curve is just an example, the parts of the Time-CI curve outside the two data points is completely made up.
You can get a more complete family by developing several tests at different developing times, some people like a series that tends to give a broad set of contrasts such as 4, 6, 9, 13, 17 minutes.
https://www.beefalobill.com/images/RadiantTime-CI.pdf

Click to expand...

Yes, of course with two sample points the Time-CI is a bit useless.

But thanks again for making such graphs (by hand!!)

radiant said:

Hmm, how did you come up with that? And I have to show my stupidity byt what is log-H and why that is used, since the density is logaritmic itself?



I'm just keen to see first hand how things work. And by doing that I have learned a ton already, which is pretty exciting. Also I wanted to see if I can build densitometer from just simple accessible things (if you consider 3D printing and using WS2812 accessible, but those are pretty normal today). So far it seems it is possible (and why not, really).

What comes to sticking to ideas - I don't know any other way to calculate lux-seconds for exposure that this "18% method". I believe that is pretty standard. I'm using 18% term as reference for you so you know how the film has been exposed. Also I've never used pushing to gain speed, actually. Because of my methods; salt prints and overall love in punchy photos, I've found that I need contrast in my negatives. Maybe this is just proving my self that overdeveloping is OK I know it gains contrast but what else horrible happens? That has been my question.

Click to expand...

Maybe I'm just misinterpreting the terms you use like with this example. It appears to me that you are indicating two tests were made with the same film type with one being developed to 400 and the other to 1600. To me this indicates developing for speed.



As to 18%, it refers to Reflectance. Reflectance has to do with Luminance and what strikes the film is Illuminance. Basically, 18% doesn't communicate anything about the sensitometric exposure. Now, you could say this is the metered exposure for 100 speed film, but that doesn't mean the meter object is 18% Reflectance. You can also show where the 18% Reflectance value will fall on the film curve as with the following example, but as you can see, it's relevance is related to the camera exposure. If you use of 18% refers to some method you use, then you can see how easy the use of the term can be misinterpreted. Maybe you could explain your 18% method. I may have missed it if you defined it earlier in the thread. I may have also missed how you confirmed the actual exposure for each of the steps. For my EG&G, I sent it to EG&G for calibration testing, and used a calibrated step tablet.



log-H is used because H is also used and H is arithmetic exposure in lux seconds.

I'm working on an explanation of what I call a developmental model, which is just another way of saying what one considers Normal, plus, and minus development. It's an important part of interpreting sensitometric results and correctly integrating sensitometry into practice. Maybe we can start out by asking what is your values for -2, -1, N, +1, and +2 and why and how they were determnined.

The slot you marked 1,5 is so close to 1 meter candle second that if you made it give less exposure by 0.03 units it would be 1 meter candle second. That could be a very useful aim for a calibration of your sensitometer, to have it give a whole unit of light

The reason you are “just under” ASA is nothing to do with exposure. It is everything to do with contrast. Developing just thirty seconds more will get you there. Then after graphing we may find the tip of the triangle moves a little. If so, my estimate of your meter candle seconds will also move. Maybe your 1,5 tube really needs to be reduced in exposure by 0.04 log units (as it stands now cutting it 0.03 will probably hit 1 mcs). Move them all the same amount to keep the 0.3 units apart from each other.

And I think since some of your tubes made no mark on film you can increase them all. I was totally wrong about you needing dimmer low exposures… you are wasting a few tubes not getting anything on 400 speed film

Stephen Benskin said:

Maybe I'm just misinterpreting the terms you use like with this example. It appears to me that you are indicating two tests were made with the same film type with one being developed to 400 and the other to 1600. To me this indicates developing for speed.

View attachment 297622

As to 18%, it refers to Reflectance. Reflectance has to do with Luminance and what strikes the film is Illuminance. Basically, 18% doesn't communicate anything about the sensitometric exposure. Now, you could say this is the metered exposure for 100 speed film, but that doesn't mean the meter object is 18% Reflectance. You can also show where the 18% Reflectance value will fall on the film curve as with the following example, but as you can see, it's relevance is related to the camera exposure. If you use of 18% refers to some method you use, then you can see how easy the use of the term can be misinterpreted. Maybe you could explain your 18% method. I may have missed it if you defined it earlier in the thread. I may have also missed how you confirmed the actual exposure for each of the steps.

Click to expand...

Ah yes, miscommunication here. Yes I'm developing for speed of course - because there is no point varying exposure with such exposure latitude that I have been using (it is basically moving slots up/down in analysis process). Lux seconds for Zone V or 18% gray = 10/ISO.

The slot times around this 18% are just one stop increase/decrease in time, of course logaritmic (as how stops work).


(X-axis exposure, Y-axis time) - I cut some of the longest times to increase resolution. Last slot gives 25.6 seconds of exposure at 0.2 lux-seconds.

But as noted before, these times are a bit off, underexposed by 1/3 stop.

Family life demands, so I will continue later ..

Stephen Benskin said:

Maybe we can start out by asking what is your values for -2, -1, N, +1, and +2 and why and how they were determnined.

Click to expand...

I haven't determined those. I've done some zone system photography with large format but as I was shooting negatives for salt prints, I just needed to max out the contrast. I've also made some iOS apps to help in Zone measurement (https://apps.apple.com/fi/app/zone-tool/id1489637620 and https://apps.apple.com/fi/app/zoneview/id1555845301).

Bill Burk said:

The slot you marked 1,5 is so close to 1 meter candle second that if you made it give less exposure by 0.03 units it would be 1 meter candle second. That could be a very useful aim for a calibration of your sensitometer, to have it give a whole unit of light

Click to expand...

How did you measure that? Do you assume that the film ISO is 400 and EI is 400?

Bill Burk said:

The reason you are “just under” ASA is nothing to do with exposure. It is everything to do with contrast. Developing just thirty seconds more will get you there. Then after graphing we may find the tip of the triangle moves a little. If so, my estimate of your meter candle seconds will also move. Maybe your 1,5 tube really needs to be reduced in exposure by 0.04 log units (as it stands now cutting it 0.03 will probably hit 1 mcs). Move them all the same amount to keep the 0.3 units apart from each other.

And I think since some of your tubes made no mark on film you can increase them all. I was totally wrong about you needing dimmer low exposures… you are wasting a few tubes not getting anything on 400 speed film

Click to expand...

The 1.5 is 5 stops from 18% point, right?

My device works so that it calculates the exposures for slots based on the "18% point" time and increases/decreases time in logaritmic way for each step. Now in these strips I'm going one step in both directions, so it causes the low density side from 18% to be exposed at -8 stops (or -2.1 in log) which is quite a demand from film. That is why the lowest slots do not have density as those are just getting too little light.

I think dimming the lowest slots would be good idea for faster films. But if I change the system to go in half steps down from 18%, it would mean the lowest one will get -3.5 stops and lowest slot should be totally visible. I could even aim for the 5 stops range that would make 0.63 stops between and therefore increase the resolution for toe analysis.

I've also thought of changing the offset of exposure so that there would be less slots over 18% position, maybe by 2 (so 6 slots used). That would increase the toe range even more. The beauty of this design is that I can freely adjust all exposures on each slots as I want.

What comes to calibrating times for separate slots .. If we expect that the system exposes all slots with good accuracy, which it seems to be doing, I wouldn't calibrate the slots at all. It would be interesting to know how in the earth you can see the needed calibration.

I think Bill you are a magician as you can read so much from just so little of data!

Numbers tell more, so this would be my next exposure plan (where 0 = 18% exposure of course):

radiant said:

Lux seconds for Zone V or 18% gray = 10/ISO.

Click to expand...

I think I need to get a better understanding from your perspective. So you are using 0.025 lxs for 400 speed film? This is your base and you are adjusting / setting the other exposures up and down the range according to this point?

How do you know you are getting 0.025 lxs?

The ASA triangle (ISO ) Bill talks about refers to the contrast parameters in determining the ISO film speed for black and white negative films.



What he didn't mention is that you can determine your EI without it by using the Delta-X equation. The CIs are a form of average gradient to determine the slope of the film curve.

Stephen Benskin said:

I think I need to get a better understanding from your perspective. So you are using 0.025 lxs for 400 speed film? This is your base and you are adjusting / setting the other exposures up and down the range according to this point?

How do you know you are getting 0.025 lxs?

Click to expand...

It's good that you ask. Yes 0.025lxs for 400. I've measured the lux coming out from the slots and it is pretty accurately 0.2 lux.

And correct again with that I use that as "root" (slot 8) and run different exposures around that with the stop stepping that I want.

What is strange is how you can use CI for determining speed.. That is strange. You must need to know something else of the exposure?

I've been reading about determining speed. Of course if you know lux-seconds of the sampled data, it is really easy to find out the speed. Just find the speed point and that's it.

Is there some other characteristics rather than how long the film has been exposed to determine the speed?

I screwed up the drawing but guess what… you hit the ASA triangle exactly with six minutes. Don’t change the developing time, forget what I said about 30 seconds more! This is how the triangle should look…


Some sensitometer designs include feedback for intensity / time with a light sensor / accumulator circuit. I assume you don’t have feedback but have some kind of voltage regulator that assures the light will be the same each time?

You have the ability to tweak the time individually per tube. That’s a unique (patentable!) design feature.

Your scanner seems to be serving very adequately as densitometer!

I calibrate my sensitometer by the ASA triangle and on the assumption that Kodak works very hard to assure that TMY2 is exactly 400. (As Ilford does with HP5+). Using that triangle as a guide, the closer your curve gets to hitting the points of (1,3 run 0,8 rise, from 0,1).

Since you hit the triangle exactly, you can take my paste-up top scale as your calibration (until another test suggests a different calibration, or maybe a discovery that something is “out of control”)

If you can reduce all the exposures by 0.03 log meter candle seconds, then your calibration will coincide with 1 MCS at your “#3 from the top” tube (1,5).

To sum: You hit ASA exactly so now my top scale is your calibration.

I mean your #4 tube is 0.03 log MCS and if you reduce all the exposures 0.03, your #4 tube will ‘conveniently’ be 0 log MCS

And you have readings down to the #14 tube.

You could make a few tubes (e.g., #16, #15, #14 and #13) be 0,15 log units apart. It’s good to “waste” one tube such that it makes “so little exposure it can’t be detected”. But you have a few that you’re “not using” so you can shift everything down (or experiment with varying intervals). Guaranteed if you vary the intervals I will make a mistake or two graphing…

If you can make the total run be -3,1 log MCS to 0,45 log MCS, your whole curves will sit pretty on my graph paper