Topic: A film's senistometric density curve and determining ISO

[mcduff]

This started as a MASSIVELY OT comment about Eirik’s question about a film's speed. As it is a very nerdy question I thought I would do folks the favor and make it a side post, instead of killing this thread!  That way people could ignore it at their leisure (francois, you cant tho!). It has to do with understanding a film’s ‘sensitometric' curve and the relationship between it and the films rated ISO.

Looking at a typical film curve, showing density and Log exposure...


The flat section of the curve is generally regarded as the film latitude, which makes sense as you want the film density behaving in a predictable manner. My understanding is that the rated film speed will for generally in the middle of this flat part of the curve^*.

I found this site pretty informative (hopefully it is also correct): http://filmshooterscollective.com/analog-film-photography-blog/a-practical-guide-to-using-film-characteristic-curves-12-25

From it gave an equation to convert Log Exposure in lux-seconds into ISO:
LogLuxSeconds = log10 (10/ISO)

If I remember my math correctly, that would mean that
ISO = 10^(1-LogLuxSeconds)  (That is intended to read as “Ten to the power of, bracket, one minus logluxseconds, close bracket.”)

The second equation is correct, i believe, if the first equation is correct!

These values correspond for the sample values that the above info gives in a table and the examples given in the above link. They also seem to work when I go look at various curves (as long as the X axis is using lux-seconds).

SO I HAVE 2 BIG QUESTIONS:

#1 Is this more or less correct???
Am I just fully cooking with horseshit or is this more or less the relationship of ISO to lux-seconds? My OCD is hitting me bad over this, haha, and I really wanna know if I am viewing these bloody densitometric curves in any way that makes sense!!!


#2 What does it mean for Eirik’s situation?

Eirik, I know that you supplied a product data sheet for kodak vision 3 intermediate film, but is this not the film you are wondering about? http://motion.kodak.com/motion/uploadedFiles/kodakVision52427242.pdf

If that is the case then I would eyeball that that ISO could be calculated at somewhere around the point 1.2 lux-seconds which according to the about formula would would out to be 10^(1-1.2) = 10^-.2 = .6 ISO

That vision three film that eirik was comparing it to looks like it is much more sensitive, however in that it looks like its ISO would be somewhere in the range of 400 ISO: 10^(1- -1.6) = 10^2.6=398 ISO



I also found this site, pretty good: http://www.olympusmicro.com/primer/photomicrography/filmexposure.html


^*This is a bit of a gross exaggeration: You can eyeball it or a more specific method from wikipedia follows: Film speed is found from a plot of optical density vs. log of exposure for the film, known as the D–log H curve or Hurter–Driffield curve. There typically are five regions in the curve: the base + fog, the toe, the linear region, the shoulder, and the overexposed region. For black-and-white negative film, the “speed point” m is the point on the curve where density exceeds the base + fog density by 0.1 when the negative is developed so that a point n where the log of exposure is 1.3 units greater than the exposure at point m has a density 0.8 greater than the density at point m. The exposure Hm, in lux-s, is that for point m when the specified contrast condition is satisfied.

[Francois]

Boy it's been a while since I last read-up on stuff like that!
The speed you can calculate from the info given by Kodak could be misleading in many ways. First, they calculated their curves for tungsten light, not daylight. Many films loose a lot of sensitivity under tungsten light. By looking at the curve, you can clearly see that this has affected the red emulsion layer in a big way. The red layer looks a lot less dense than the other colors. And then, they processed it "correctly" using ECN-2 chemistry. That will also affect film speed and color balance.

But since we usually disregard sensitometric data around here, the easy way would be to just eyeball it. But that wouldn't be any fun would it?
Reading the wikipedia info and using coreldraw, I calculated that the log exposure for the blue layer (I really should have taken the green one) is of 1.4 lux/sec.
So using your calculation, that should make a whopping 0.4 ISO... now lest check if this is right... yes it is.
It should be pretty close to the real speed since the value hits the center of the curve's straight part.

I hope I'm right...

[Indofunk]

Hi. I like stuff. Do you like stuff too?

[jharr]

Oh man. I have been looking at sigmoidal IC50 curves at work all day, then I take a break to relax a little and what do I get??? More sigmoids! This thread needs a picture of a pretty flower.


pink by James Harr's Photos, on Flickr

[Bryan]

I like stuff too, but this makes my head hurt.  What exactly is Optical Density?  In English please.

[Indofunk]

Yes! More flowers, less math!! 

[sapata]

man...

[jojonas~]

no! let the math and curves and shit commence!! this is exciting and I yearn to learn this haha but getting past the basics is a bit hard I admit

[jojonas~]

kicker: anyone here tried using their scanner as a densitometer? http://www.mrecord.talktalk.net/density_measurement/index.html

[Photo_Utopia]

"The flat section of the curve is generally regarded as the film latitude, which makes sense as you want the film density behaving in a predictable manner. My understanding is that the rated film speed will for generally in the middle of this flat part of the curve*"

This is completely wrong! Latitude is worked out over the entire curve (toe and shoulder) in other words the range of tones you wish to record should sit within the whole curve.
 The rated film speed is worked out as the speed being at film base plug fog (0.1) the most important part of the curve is the TOE the average gradient speed is worked out from there. (I can see a better explanation in the footnote of the OP)
If you wish to know how a gradient based speed is calculated and the difference between CI and Gamma and how to calculate the latitude.
n.b colour and B&W have different speed calculation methods.
Look here
http://motion.kodak.com/motion/uploadedFiles/US_plugins_acrobat_en_motion_education_sensitometry_workbook.pdf

Kind regards
Have fun

[Ezzie]

Sorry folks, my fault for getting Don started.

[Photo_Utopia]

I like stuff too, but this makes my head hurt.  What exactly is Optical Density?  In English please.

OK You asked for it

The degree of darkening of film is called Density. Density is a measure of the light-stopping ability of the film. To determine the meaning of density, we must first get familiar with the terms transmission and opacity.
Transmission is how much of the light gets through the film, and opacity is how much of the light doesn't get through the film.
 
Thus, opacity is the reciprocal of transmission (O = 1/T). For example, if 100 lux of light is falling on one side of the film, but only 25 lux gets through, then the transmission is 0.25. The opacity would then be 4. Notice that, like all reciprocal relationships, transmission multiplied by opacity equals 1.

Now, back to density. Density is the logarithm to the base 10 of opacity (D = log10O). Since the only logs we will be talking about are to the base 10, we will omit the subscript from now on (log = log10).

The density of your film can be measured by an instrument called a densitometer.
If you use a densitometer, you do not need to figure transmission, opacity, or the log of the opacity, because the instrument will indicate the density on its scale (or readout)

So basically the optical density is how dark the film is the more light that hits a negative the darker it gets, a curve is a pictorial representation of that.

[Ezzie]

I may understand something if this continues, but then again I may not.

[jojonas~]

I like stuff too, but this makes my head hurt.  What exactly is Optical Density?  In English please.

OK You asked for it

The degree of darkening of film is called Density. Density is a measure of the light-stopping ability of the film. To determine the meaning of density, we must first get familiar with the terms transmission and opacity.
Transmission is how much of the light gets through the film, and opacity is how much of the light doesn't get through the film.
 
Thus, opacity is the reciprocal of transmission (O = 1/T). For example, if 100 lux of light is falling on one side of the film, but only 25 lux gets through, then the transmission is 0.25. The opacity would then be 4. Notice that, like all reciprocal relationships, transmission multiplied by opacity equals 1.

Now, back to density. Density is the logarithm to the base 10 of opacity (D = log10O). Since the only logs we will be talking about are to the base 10, we will omit the subscript from now on (log = log10).

The density of your film can be measured by an instrument called a densitometer.
If you use a densitometer, you do not need to figure transmission, opacity, or the log of the opacity, because the instrument will indicate the density on its scale (or readout)

So basically the optical density is how dark the film is the more light that hits a negative the darker it gets, a curve is a pictorial representation of that.

great! I kind of got that part already but putting word on it like that really helps.
I'm definitely printing out that workbook you linked. this will be fun!

[charles binns]

Fascinating stuff.  Now where did I put my Holga?

[Bryan]

I like stuff too, but this makes my head hurt.  What exactly is Optical Density?  In English please.

OK You asked for it

So basically the optical density is how dark the film is the more light that hits a negative the darker it gets, a curve is a pictorial representation of that.

Thank you.  So the curve for reversal or slide film will slope in the opposite direction?

[mcduff]

haha i am sorry to have inflicted this on folks - BUT I LOVE IT!

PhotoUtopia, I am busy reading that great kodak PDF, I will get back to you. But I have already understood that I was incorrect about the flat part of the curve being the latitude. I thought the toe and heel were too crazy to use but it sounds like that is not the case! back to reading...

EDIT: This PDF is exactly what I needed, it has a great description of determining film speed. Still processing...

[Francois]

I must admit that this has all got me scratching my head last night. So I popped out from my bookshelf a small but heavy volume on photographic sensitometry. You know, the kind of book you don't keep on you bedside table unless you want to get a chance to say "not now honey, I have a headache." !
But it at least was quite interesting. I did learn that the reason the old films has a different speed than the new ones was because of the way that the tests were designed was different. From 1943 to 1960, the ASA rating was based on both curves and user input. But in 1960, the method was changed as to better match the easier to calculate European technique.
And then I also learned how to calculate the value properly... And it's quite different from what Don found on wikipedia... I knew it couldn't be that simple.... I'll have another go at it when I'm in front of my computer. And also try and give clear instructions on how to do it simply... Which is very unlike what I read last night...

[Photo_Utopia]

haha i am sorry to have inflicted this on folks - BUT I LOVE IT!

PhotoUtopia, I am busy reading that great kodak PDF, I will get back to you. But I have already understood that I was incorrect about the flat part of the curve being the latitude. I thought the toe and heel were too crazy to use but it sounds like that is not the case! back to reading...

A famous photographic scientist said "all the magic happens in the toe" (of negative film) that is where you put your shadow detail and has a bigger effect on overall quality than I can possibly state.
Most people forget the toe and shoulder and think the range of film is just the straight line. I suppose being brought up on digital that's what they expect...

[Terry]

When Kodak debuted 5219 they showed a test scene that Daryn Okada shot in a barn near Rochester.  It was all about the detail in the interior shadows and in the exterior highlights.  The vidiots in the room were entirely focused on the highlights, as you might expect.

PS: I have to disagree with Mcduff--this thread is not OT at all! 

[Photo_Utopia]

I
But it at least was quite interesting. I did learn that the reason the old films has a different speed than the new ones was because of the way that the tests were designed was different. From 1943 to 1960, the ASA rating was based on both curves and user input. But in 1960, the method was changed as to better match the easier to calculate European technique.

Which volume was that? I ask because the info might not be current. The original 1930's method was evaluated with a Dlog 1.5 E units and in the 1958-63 period they changed it to DLog 1.3 which gave a stop of extra speed overnight (Tri x went from 200 to 400ASA)
This was done for a number of reasons, lens coatings giving better transmission right though to less safety margin needed for exposure as lightmeters became more common.

The Science of Photography-Baines and Bomback and the tome by C.E.K Mees "The Theory of the photographic process" give a more detailed info.

Failing that the linked workbook has the correct methodology in an easy to follow manner

[jharr]

Pretty flower.


Purple Magic by James Harr's Photos, on Flickr

[mcduff]

First of all, MANY THANKS PhotoUtopia!

I think I have it and that PDF workbook you linked me to is most valuable. It is actually pretty awesome as it has exercises that you can work through. I know some folks find this stuff as repugnant as talking Digital, haha, but my poor little brain loves it.

I like two things about the description of determining film speed that this workbook gives. It explains where to determine the calculation point (tho I have one minor question still about this), and then it gives the proper equation for calculating film speed. My equation was wrong (I think my math was sound, I was just basing it on an incorrect equation).

I have gone through and determined the film speed of TRI-X, as an example, and it WORKED. (I have scientifically determined that tri-x is 400 ISO for those of you that havel always wondered what its speed is, haha.)

So here is the description for finding the point on the X-axis (the Log E scale) which we use to then determine film speed (from that PDF):
1) Find a point on the curve that is 0.10 density units above D-min and label it A. Make a note of the log exposure at this point.
2) Count over 1.30 log E units, and mark this point. Draw a line from this point up to the curve, make a note of the density at this point, and label it point B.
3) The film is properly developed if the density at this point is 0.80 ( ± 0.05) more than the density at point A.
4) We then use point A's value on the X scale to determine film speed using this formula: Film Speed = 0.8/(Antilog of A)

The one question I have is that it does not really explain what to do if the density difference is not within .75 and .85. That is the case in the example below where the density difference was only .66. Now I was just goofing around in pshop from a screencapture of a PDF, etc. So there is lots of places for the error to pop up, but clearly there could be situations where the "Gamma" (to now use a term that I learned from the PDF) is going to be too high or too low to give the desired density range of .75 to .85. Photo-Utopia, what do they do then? I am presuming they would say that the film was not developed correctly for the test?

I can live if that last question remains unanswered. 

Anyhow, here is my example using Tri-X with my now, hopefully correct notions!!!!!!


I can now sleep at nite! And James, don't make me take some of your beloved sunflowers and draw "Golden Spirals" on them -- which is driven by the fibonacci sequence:
0, 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144, 233, 377, 610, 987, 1597, 2584, 4181, 6765, 10946, 17711, 28657, 46368, 75025, 121393, 196418, 317811... (sorry)

[jharr]

Don, this is interesting (that's why I keep coming back to this thread). I am just messing with you guys with the flower pics.

I do have one question. It is the question I always ask myself when I start getting bogged down with math, geometry, chemistry, etc. "So what?" Okay, that's a little terse, so let me put it this way. "What does this bring to the table as far as my (or anyone's) practice of the hobby goes?" That isn't a rhetorical or facetious question. It is quite sincere because if there is something that I can get out of this that will help me achieve my film photography goals, I really want to know it. This may apply very soon since I have 400' of expired Plus-X Aerecon II coming and I don't really know how I am going to use it. How does one construct a curve? Do you get a different curve for each developer/time combination? Do I need a proper densitometer or will my V600 work well enough. I'm not making masterpieces and I'm not against wasting a dozen feet of film with brute force experimentation, but there does seem to be in here a method for those of us enamored with expired copy film and the like.

[mcduff]

OH, noes, now my OT is going really OT!
James I am teasing you back because I always have a left brain/right brain battle going on. Not to say you have a left brain/ right brain battle, as I do, but I am pretty certain that you are a kindred spirit that has a nerdy part (hence you dropping the term 'sigmoidal' in an earlier posting  ). For me this battle frequently pops up, and I often my initial reaction is tell that part of my brain to 'shut up, sit down, and hold on' and quit thinking stuff through. And certainly when I thinking about shit like this, my creative side is always in the corner, tapping their foot and saying "are we through, yet??"

So when I am doing creative things sometimes I tell this part to shut up, but now I am trying more to engage it this part of me (without letting it fully take over as it wants to!) I go through periods where I am a fairly active painter/drawer, and some of my most personally satsifying work has both parts of me engaged.

But something like photography pulls the scientific part out a lot, and I think that this is partly because it is a very technical art or hobby. Much more so than when I am drawing and I am really not thinking too much about the physical parameters of the oil stick I am mushing into parchment! This is a hackneyed saying but photography is one of those experiences that does somewhat straddle arts and science. (I know, you could say that for a lot of activities, hence, its hackneyed.)

But in instances like this, specifically, there are a few reasons these sorts of things matter to me. (And I say "matter" in a totally silly use of the word. For example when I say "now I can sleep at night" my creative part of me is mocking the analyical OCD part of me for ironically being such a drama queen!). Here are some reasons why I wanted to figure this out:

One reason it matters to me is that my analytical part of my brain, begs me to 'toss it a bone' every now and then. And I have learned that if I do that, it will go away in the corner and chew on it a bit and keep it out of trouble. It is a way I keep peace with it by letting it nerd out now and then! 

A second reason it matters to me is I hate being presented with information and not understanding it. So it has always bugged me when I am looking at data sheets on film that I did not understand what these curves mean. Now why I was looking at them in the first place may be a good question!

A third reason is sometimes, such as the case for Eirik's film, one sees these charts but that is all the info you can find. You cannot find the rated film speed. If I was in Eirik's case where I had only three rolls of film, I would want to have an idea as to where to start. Now I can look at a chart and quickly determine if the film is fast or slow (his is slow).

[Ezzie]

Fourth reason is that this explains why a film may be rated differently according to developer. Density will build the longer you develop the film, but not necessarily at the same rate (with different developers). Don's curves could have been tagged with developers instead of minutes developed (in the one brand of developer) and worked.

[Terry]

The simplest explanation, in real-world terms, is that this film (as all motion picture interneg stocks) was never intended to be put into a camera.  These are all made for use in optical printers whose operators calibrate exposure in terms of "printer lights" rather than f-stops.  If you look at the Kodak sensitometric curves for their camera stocks (say 5219), you will notice that the exposure scale is given in divergence from optimal exposure in f-stop increments.  This is the most useful information for a camera operator--it tells you how many stops below your meter-reading and above it you can go without losing too much detail.  The lab techs get into denistometry and hair-splitting (because they can) but to someone shooting in the field, knowing how much under or over the "ideal" exposure you can go and still get useable information is the critical matter.  5219, for example, can reasonably claim to give 16 stops of useable latitude and an ISO of 500 in tungsten light; an interneg stock is designed for attributes that pay off in the transfer process, where detail and certain color balance differences may be more important than sensitivity, as their respective curves show. 

[Indofunk]

I am just messing with you guys with the flower pics.

I however am not kidding with the flower pics ... they're all I understand in this damn thread! 

[Terry]

With you there, Satish: as I see it there's no substitute for "shoot some and see" as the best solution, especially when you're using films for purposes other than they were intended for.  James has got some nice flower shots with interneg stocks...(I think)

[mcduff]

One thing I found out when I started reading this stuff is that Ilford (who I am usually a #1 fan of) does not provide us with as much info as folks like Kodak.

For example, as soon as I figured out how to calculate film speed, I thought it would be fun to look at Delta 3200 -- you know the film that has '3200' in its name but everyone says it is really just a 1600 film. What is disappointing is that ilford's charts to not give a unit of measure for the 'log exposure' scale, just calling it a 'relative log exposure' scale. That means we can get a general understanding of how the film works, but we can't look at, do a little math and say "yup that is 1600 ISO film"

You could probably reverse engineer the scale they are using, but a) who knows if it is consistent between their films, and b) that is too much, even for me!!!

[mcduff]

I don't know if I fully buy satish's story, whenever I hear musicians starting to talk about music and explaining how things work, it gets pretty math-y sounding pretty fast. I think he is a 'denier'! 

[Indofunk]

With you there, Satish: as I see it there's no substitute for "shoot some and see" as the best solution, especially when you're using films for purposes other than they were intended for.  James has got some nice flower shots with interneg stocks...

Yeah, that's exactly what I meant! 

But seriously, my vision of old school CN filmstock developers is them just shooting rolls at different EIs, looking at them and saying "ayup, that one there's gonna be what we rate it at"

[mcduff]

But seriously, my vision of old school CN filmstock developers is them just shooting rolls at different EIs, looking at them and saying "ayup, that one there's gonna be what we rate it at"

Well, ya I think that likely is the case. I mean if nothing else they are likely rounding things to the nearest stop. But I think Delta 3200 is a case of what you are talking about because Ilford says 3200 and world+dog says 1600

[Terry]

I've heard a lot of people bad-mouthing Kodak, sometimes with good reason, but they certainly did a lot of research and education work that couldn't have been profit-motivated.  I once visited the research labs at Kodak Park and was blown away by the stuff they were doing there that seemed entirely driven by a desire to make imaging systems better.  (Like paying physicists to study the chatter of film going through the gate of a camera or projector so as to get it to lie flatter and improve the image.)  The sensitometry workbook is an example, as is a whole range of educational material about production organization, lighting, shooting and handling film and postproduction that I and many other teachers have been using in our film classes for a long long time.  The video equipment manufacturers have done nothing comparable to this.

[Terry]

Satish, you just perfectly described a color-timing session.

[Indofunk]

I don't know if I fully buy satish's story, whenever I hear musicians starting to talk about music and explaining how things work, it gets pretty math-y sounding pretty fast. I think he is a 'denier'! 

I think you are giving me far too much credit, Mr McDuff! Plus, talking in terms of notes and chords and even beats is a few layers of abstraction removed from all this sensitometry jibber-jabber

[mcduff]

The sensitometry workbook is an example, as is a whole range of educational material about production organization, lighting, shooting and handling film and postproduction that I and many other teachers have been using in our film classes for a long long time.

No kidding, it is a pretty awesome resource. I was really surprised when I found out the answers were included, I really expected it was a document that was part of a paid course (where the answerbook was separate). They did provide valuable information and like I said, they are really up front about the data. That is really good.

[mcduff]

this sensitometry jibber-jabber

I hope you understand that part of me totally agrees with you!!

[Adam Doe]

My thoughts:
https://www.youtube.com/watch?v=X8CrHrx2k5M

[jharr]

My thoughts:
https://www.youtube.com/watch?v=X8CrHrx2k5M

Now not only is my Youtube suggestions list polluted, but I'm going to have that song stuck in my head for the weekend. 

[Francois]

OK!
I did slave over the book I have and managed to write something for the articles section which I hope is good enough... and which I hope actually works!
Now after reading this book, I feel like I've just been through a lobotomy... or that my brain was just put through the wringer or something...
Here's the link
http://www.filmwasters.com/forum/index.php?topic=7614.msg99609#msg99609

[Francois]

The one question I have is that it does not really explain what to do if the density difference is not within .75 and .85. That is the case in the example below where the density difference was only .66. Now I was just goofing around in pshop from a screencapture of a PDF, etc. So there is lots of places for the error to pop up, but clearly there could be situations where the "Gamma" (to now use a term that I learned from the PDF) is going to be too high or too low to give the desired density range of .75 to .85. Photo-Utopia, what do they do then? I am presuming they would say that the film was not developed correctly for the test?

Sensitomtry is just a lab based value. The density slope indicates what quality image can be expected from the negative. It is often an important value that photographers fix depending on the type of enlarger they use. Kodak recommends aiming for a density range of 0.8 for condenser enlargers and 1.05 for diffusion types.

If the density is too low, you get a flat negative. Too high, it's too contrasty.

[moominsean]

huh?

[Francois]

Don't blame you... I'm starting to feel the same!

[Indofunk]

I think this calls for .... another flower pic!

[mcduff]

haha i love the diversity of responses to this thread! I tried to warn folks that a scary part of me was needing some MATH! 

[Francois]

Well at least it's not "bad math" or overly complex math.
I have a remote cousin who got a masters in algebra. If you want complex math I can ask her to drop by

Actually, I'll spare you all... I haven't seen her in over five years.

[Francois]

I just found this which could be of some interest
http://motion.kodak.com/motion/uploadedFiles/US_plugins_acrobat_en_motion_newsletters_filmEss_06_Characteristics_of_Film.pdf

[jharr]

Any thoughts on how the curves change as film ages? Does the iso decrease in a linear fashion or maybe an inverse log?

[Francois]

It all depends on processing.
What I know is that the base+fog level will increase in density, but the maximum density possible should remain the same. So the curve gets flatter.

But the question is by how much and at what rate?