Film density range

[ChrisR]

I've been researching the bit about density range for the big scanner thread. This is hampered by the fact that nearly everything I read contradicts something else I've read, and nothing seems particularly authoritative. However, I think I've struggled through enough to write something reasonably short and (I hope) comprehensible for density range of scanners.

To make sense of it, though, I think I need to talk about the density range of film. Here I don't find as many sources. Remember that film density is measured on a log base 10 scale, from zero (absolutely transparent) upwards; it appears that a value of 5 is effectively completely opaque.

So Ken Rockwell (yeah, don't @ me!) says (2/3 down the page) that Velvia 50 goes to a maximum density of almost 4.0D (his notation); other slide film is mostly around 3.5D. Since the clear substrate has a density of 0.05D, that gives Velvia a density range of about 3.95D. He then claims that colour and black and white negatives only go to about 1.4D for perfect negatives, and even heavily over-exposed negatives only go to 2.0D (so the density range of scanners is rarely an issue for these films).

On the other hand, filmscanner.info write in a badly translated English version of their German original, that slides have a density range of 2.4 (or 8 stops), while negative film has a density range of 3.6, or 12 stops.

That's quite a difference. Now to me, it seems likely that the density range of negative film is less than slide film. For a start, colour negative film has an orange substrate, so the minimum density must be significantly higher than for slide film. And secondly, I would guess that most of us have experienced more trouble with shadows on slides than with highlights on negatives. So I'm tending to the Rockwell view.

I also wonder if the filmscanner.info author is confusing the film density range with the scene dynamic range, where bizarrely negative film can capture a wider dynamic range than slide film (due to some chemical wizardry I don't pretend to understand).

Does anyone have any better or more authoritative sources than Rockwell? Or alternatively a bit of Velvia 50 and a densitometer?

 

[StephenM]

This will be brief, but the quick answer is to take a look at info sheets film makers issue. You should always find a characteristic curve (= density v exposure) and perhaps even one with the third variable of developement added in.

My understanding is that because film makers traditionally expect negatives to be printed in a darkroom, they don't bother recording results that would fall outside the range of darkroom paper to handle. Or put it another way, you are potentially able to get more out a negative if you don't use a darkroom.

Solid info in some of Michael Langford's books, Coote's Monochrome Darkroom Practice, Way beyond Monochrome, Ansel Adams and any of the myriad books on sensitometry, the older Ilford Manual, Gevaert manual and umpteen more. Several of these can be found as pdfs online.

Depending on who else replies, I may add more later. Including hard figures.

 

[ChrisR]

This will be brief, but the quick answer is to take a look at info sheets film makers issue. You should always find a characteristic curve (= density v exposure) and perhaps even one with the third variable of developement added in.

So if I'm interpreting the curve for Velvia 50 in https://www.fujifilm.com/products/professional_films/pdf/velvia_50_datasheet.pdf correctly, there's a reasonably straight line from a density of 0.5 to 3.5, so a Drange of around 3? No way could I get a Drange of 4 from that chart... (Taking the whole curve rather than the straight line part you could maybe get to 3.5.)

In contrast the curve for Superia 400 in https://www.fujifilm.com/products/consumer_film/pdf/superia_xtra400_datasheet.pdf has a pretty straight line for the green (middle) line from a density of 0.5 to 2.5, so a Drange of 2. The curve for Kodak Ektar 100 looked pretty similar.

Interestingly, the data sheet for Tri-X at https://imaging.kodakalaris.com/sites/prod/files/files/resources/f4017_TriX.pdf shows a range of Dmax values, so the Drange seems to vary from about 1.5 when developed for 6 minutes with Tmax developer, to nearly 2.5 when developed for 11 minutes!

If someone else could confirm or correct my reading of the first two, I'd be grateful!

PS I've just checked out Ektachrome 100 at https://imaging.kodakalaris.com/sites/prod/files/files/products/e4000_ektachrome_100.pdf, and the Drange looks like 3.5 too.

 

[StephenM]

If no-one else steps in (I'll be busy for the next few hours) I'll see what I can do. As a black and white photographer, I seized on the Tri-X data. I note that it's dated 2016, and is only 2/3 the size of the 2004 one I had. I don't know if that indicates that they give less information away. But mine (at least) shows the DMax going to 3 before they stop plotting (figures above that are outside the range of printing papers, if not scanners, so they stop there).

 

[srichards]

Might also be worth emailing Ilford as I'd bet they'd like to answer a techy question like that.

 

[AndrewFlannigan]

I think that, while it might be fun to get your head around dMax and the mysteries of densitometry, all anyone needs is to follow the developer manufacturer's recommendation for time, temperature and film speed to get entirely useable negatives or slides. I've developed several thousand films over the years, many for commercial purposes and I've had difficulty printing from no more than 2 or 3.

When it comes to printing: I never found a better way to get a decent print than by makng test strips. (Actually, I did use a Melico enlarging meter for a while in the 1970s when I was doing a lot of high volume printing). Since digital became the thing I've found no better way of using a scanner than putting the film in the scanner and pressing the button. Using a camera to digitise film is just the same, only easier: Press the shutter release, check the screen and if the image looks OK, go to the next one.

On the other hand, don't let me put you off digging into this stuff if you do find it fun.

 

[Peter B]

So if I'm interpreting the curve for Velvia 50 in https://www.fujifilm.com/products/professional_films/pdf/velvia_50_datasheet.pdf correctly, there's a reasonably straight line from a density of 0.5 to 3.5, so a Drange of around 3? No way could I get a Drange of 4 from that chart... (Taking the whole curve rather than the straight line part you could maybe get to 3.5.)

In contrast the curve for Superia 400 in https://www.fujifilm.com/products/consumer_film/pdf/superia_xtra400_datasheet.pdf has a pretty straight line for the green (middle) line from a density of 0.5 to 2.5, so a Drange of 2. The curve for Kodak Ektar 100 looked pretty similar.

Interestingly, the data sheet for Tri-X at https://imaging.kodakalaris.com/sites/prod/files/files/resources/f4017_TriX.pdf shows a range of Dmax values, so the Drange seems to vary from about 1.5 when developed for 6 minutes with Tmax developer, to nearly 2.5 when developed for 11 minutes!

If someone else could confirm or correct my reading of the first two, I'd be grateful!

PS I've just checked out Ektachrome 100 at https://imaging.kodakalaris.com/sites/prod/files/files/products/e4000_ektachrome_100.pdf, and the Drange looks like 3.5 too.

I don't know if this is of any help at all Chris, but there are a number of web pieces on using Vuescan and other software as a densitometer. As I don't really understaand it all, I heven't gone very deeply into it, I'm afraid. https://www.google.com/search?client=firefox-b-d&q=vuescan+as+densitometer

 

[ChrisR]

I think that, while it might be fun to get your head around dMax and the mysteries of densitometry, all anyone needs is to follow the developer manufacturer's recommendation for time, temperature and film speed to get entirely useable negatives or slides. I've developed several thousand films over the years, many for commercial purposes and I've had difficulty printing from no more than 2 or 3.

When it comes to printing: I never found a better way to get a decent print than by makng test strips. (Actually, I did use a Melico enlarging meter for a while in the 1970s when I was doing a lot of high volume printing). Since digital became the thing I've found no better way of using a scanner than putting the film in the scanner and pressing the button. Using a camera to digitise film is just the same, only easier: Press the shutter release, check the screen and if the image looks OK, go to the next one.

On the other hand, don't let me put you off digging into this stuff if you do find it fun.

Hi Andrew, normally I wouldn't care, but since I'm writing the components for the thread about scanners, I reckoned scanner Drange was something worth looking into. Writing the (relatively short) draft post, I thought I needed one more paragraph on why we should care about scanner Dmax, and film Dmax/Drange was the obvious answer. Hence the question!

 

[StephenM]

I've read this query as being a necessary follow on to the scanners thread, along the lines of "given that scanners state the DMax they can handle, how important is this in practice? What are real life film DMax values?".

On that basis, it's a question as to whether a given scanner could actually extract detail, and whether a darkroom print can be successfully made isn't relevant. My own feeling is that with a scanner with a low DMax, some details will remain unextracted that would be easily handled in a darkroom. On the other hand, many negatives that could NOT be printed successfully in a darkroom produce excellent prints via scanning.

I found last night that reading reference books on a tablet is slow going - no easy switch to index and back to a page on my tablet software - and I'm continuing by using real books. My recollection (which might be faulty) is that the highest DMax is found with Kodachrome, and they do need a scanner that can reach 4.0 or above.

Meanwhile, I'm still checking.

 

[ChrisR]

I don't know if this is of any help at all Chris, but there are a number of web pieces on using Vuescan and other software as a densitometer. As I don't really understaand it all, I heven't gone very deeply into it, I'm afraid. https://www.google.com/search?client=firefox-b-d&q=vuescan+as+densitometer

That's interesting Peter, thanks... my first thought was that the approach could suffer from a slight issue given the scanner itself has limited Dmax! Then I remembered I don't really care too much about Dmax, provided my scanner can cope. This part of the big scanner thread is really about arming folk with information to help them when choosing a new scanner.

 

[StephenM]

Chris' 2 posts above made while I was laboriously tapping on a tablet...

 

[StephenM]

The simplest quick and dirty test I can think of without a densitometer wouldbe a reflected lightreading of an evenly back illuminated surface, with and without a piece of fully exposed and developed film in front. In my case probably a full screen blank Notepad window.

 

[AndrewFlannigan]

Here's an odd factoid: The Encyclopedia of Photography (2 volumes published in 1937) devotes 9 pages to the subject of Negative Density and the calculation of response curves but The Technique of Photomicrography (published in 1960) which is full of graphs and curves about all sorts of other things, doesn't mention the subject at all.

 

[StephenM]

Finally, a piece of flaccid information (meaning semi-hard; the hard part is from a Kodak data sheet on Kodachrome 25; the soft part is that I had to visually interpolate because the graph is in 0.5 units): Kodachrome has a DMax of 3.8.

I've so far checked
Focal Encyclopedia of Photography
Ilford Manual (latest and 1956 ed)
Kodak Black and White film guide
Kodak Professional .... vol 1
Langford's Advanced Photography 6th ed
Rolls et al Applied Photography
Mckay Photographic Negative vol 3

and umpteen others.... Getting slightly jaded.

 

[AndrewFlannigan]

and umpteen others.... Getting slightly jaded.

Attempting to integrate soft data will do that every time. I had to remind myself that someone was paying me to do it lest I run screaming from the building!

 

[ChrisR]

Finally, a piece of flaccid information (meaning semi-hard; the hard part is from a Kodak data sheet on Kodachrome 25; the soft part is that I had to visually interpolate because the graph is in 0.5 units): Kodachrome has a DMax of 3.8.

I think the way I was reading the previous charts I'd have made that a little lower (I was taking the middle of the RGB curves, but maybe taking the top is right). Looks like the Dmin is about 0.2, so the Drange is 3.4-3.6-ish...

Couple of things I hadn't noticed from the colour negative characteristics curves: blue is the highest density and red the lowest (the opposite from slide film), also, the Dmins of the 3 colours are greatly different, but so are the Dmaxes, so the Drange for each colour is pretty consistent. Slightly different from slide where it seemed like the Dmin for each colour was very close, but the Dmax varied quite a bit (red densest).

Looking back at the values from post 3 of this thread, and the underlying data sheets, I reckon it's reasonable to say the Drange of slide film is around 3.5 and colour negative film is about 2. Which is what I wanted to find out... so, I thank you all!

 

[StephenM]

I'm reminded irresistably of the introduction to "Advanced Inorganic Chemistry" by Cotton and Wilkinson. I first met it in the 2nd edition; it's now in the 6th edition with the relevant part still there. They made the point that students were happily learning how atomic properties changed as the periodic table was traversed; that they were happy to say that, for example, melting points tended to increase as you went across the table. But actually didn't know what any melting points were.

Looking at all the sections on sensitometry, all the characteristic curves, and all the discussions on density, left me feeling the same way. People explain trends, but fail to quantify.

One big change I have noticed is that characteristic curves are all truncated. The graph plotting stops at a certain level of exposure, and fails to show what happens after that. If density suddenly flattens, and the curve shows the peak, all I can say is that I would not personally have expected it from where the curve was heading.

Old curves showed a toe, where you needed a big increase in exposure to produce a small increase in density (in other words, low contrast in the shadows), a straight line part where contrast was usable, and a shoulder where the slope eased off (low contrast in the highlights). After that, the curve might start to drop, an effect used by Ansel Adams in his "black sun" photograph. The old advice was not to use the shoulder to enable highlight tonal differentiation to be made.

The other practical point was best (to my knowledge) demonstrated in one of Michael Langford's books; the the subject tonal range was compressed in getting it onto film, but that compression was nothing compared to that required to make a print. Going beyond the exposure where the curves now end would simply record things on the film that darkroom printing couldn't extract from the negative.

 

[StephenM]

Are we comparing apples and oranges? When you say Drange do you mean density range (if you like, the y axis of the characteristic curve) or dynamic range (a term I dislike because it has so many meanings as to be almost meaningless) which is the x axis of the curve?

 

[StephenM]

I've just revisited post 3, and actually checked what you said against the reference. I can see that the density increases with extra development, and the contrast as well (steeper slope). That translates in darkroom terms into tonal compression in the print, or a lower range of tonal values that can be printed. I think that Drange might mean in your terminology the increase in the maximim density. Am I hopelessly confused?

 

[ChrisR]

Are we comparing apples and oranges? When you say Drange do you mean density range (if you like, the y axis of the characteristic curve) or dynamic range (a term I dislike because it has so many meanings as to be almost meaningless) which is the x axis of the curve?

Drange is the density range, so in this logarithmic world it is Dmax minus Dmin. In post 1 I speculated that filmscanner.info had got confused between density range and scene dynamic range in their calculation, but of course there's no way of knowing that. By the way, in the same table they suggest the Drange of printing paper is 1.5, or 5 stops, which ties in with your comment above, Stephen. I don't know if the Drange of inkjet paper with the best pigment inks is the same as the best wet printing paper, but I'm not going there...

I am really confused by the results we've found... it appears that slide film has a density range of 3.5, or over 11 stops, but it's always suggested it can only encode a scene dynamic range of around 5 stops. Meanwhile colour negative film has a density range of 2, or around 6 stops, but can encode a scene dynamic range of 8-10 stops! However, I'm inclined at this point to put it down to chemical wizardry, and leave it at that! EDIT: As @StephenM points out, of course this is EXACTLY what these charts actually show, with the dynamic range effectively on the x-axis and the density range on the y-axis!

 

[ChrisR]

I don't know if this is of any help at all Chris, but there are a number of web pieces on using Vuescan and other software as a densitometer. As I don't really understaand it all, I heven't gone very deeply into it, I'm afraid. https://www.google.com/search?client=firefox-b-d&q=vuescan+as+densitometer

I finally had a look at that... several of the links say, this doesn't really work. Eventually I found this site: https://sites.google.com/site/negfix/scan_dens . He says, after showing the approach using an IT8 target "As you can see, my scanner (Epson V750) is pretty accurate in the range of 0.2 - 2.0 D, which is enough for the intended purpose. You should note, however, that this is not a standard densitometer, so you cannot compare those results..."