Maximum density

[swmcl]

Hi,

Can someone tell me whether it is the type of film or the developer that generates the maximum density on the film ? I am looking at film spec sheets and most don't advertise the 'max density' but I see that the FP4+ sheet does have a curve that peters out at less than 2 (see page 4 of the pdf). Doesn't this mean that it cannot display more than say 6 or 7 stops ? (Each stop being 0.3 on the density scale).

Can one get greater density out of a film depending on developer or is it fixed in the design of the film ?

Densitometers go to a density of 4 which would again be say 13 or 14 stops yes ? Are there films or combinations that go to beyond 3 ?

The reason I ask these questions is because I picked up a photography book the other day that said B&W film could record beyond 14 stops. I can't see how unless one were to interpret a 'stop' as being say a 0.2 step on the curve rather than a 0.3.

Cheers,

Steve

[Rafal Lukawiecki]

Steve, to answer the question if a film can or cannot record X number of stops you need to look not just at the film density range, but at the relationship between the light being recorded, on the horizontal scale of the curve, and the film's response as density, on the vertical. You count the stops being recorded, just as you wrote, as 0.3 logE, but on the horizontal, the X axis. What you want to see is that for a range of stops there is a meaningful change is density on the vertical, Y axis. Where the curve is flat, you get no separation, therefore you are not registering any "stops" of light additional to those already recorded.

For example, to record 14 stops, which you mentioned, you would like to see a curve that is not flat (or nearly so) over an exposure range of 14 * 0.3, that is over some 4.2 units of exposure (logE). With a typical film contrast, eg. about 0.6, this would produce a density range of as much as 2.4 logD on the negative. Some film/developer combinations can get you such a response, as indeed, it is not the film alone that matters, though it is the key to it.

The rest of the story is how this recorded data, on the density, vertical axis, gets to represent something on paper. This is all a matter of the selected paper's contrast, or if you think of the paper's very own curve, its own useful exposure range.

[N Dhananjay]

The subject luminance range that can be recorded is not a function of the Dmax of the film (unless you were trying to get close to a 1:1 relationship, which you usually do not). Rather, it is a function f the entire chain. Usually, the subject luminance range is compressed in the film density range and then expanded out in the paper density range. In other words, the longer the subject luminance range, the lower the contrast to which you would develop the negative.

Usually, in the chain, the limiting factor is the paper. Paper typically has a Dmax of about 2.1 or thereabouts, which means it can record about 6-7 stops. If you tried to record 14 stops on the negative and then translate that to paper, you are eventually compressing the 14 stops in the subject onto the available 7 - each 'zone' will be poorly separated and the local contrast will be terrible. Now, if there are unused bits in that range, you may be able to make a good print by trying to use various techniques to steal contrast from the unused area and redistribute it in the useful area but it is all usually more trouble than it s worth. So, in general, making photographs with 14 stops is a bit pie in the sky and if one is interested in that, becoming a lighting artist on the stage will be more satisfying.

What you are trying to do in the negative making stage of the chain is to get a negative that matches the scale of your paper. Different grades of paper accommodate different negative density ranges. So, what you are trying is - compress the subject luminance in the negative to the appropriate amount so that when it is matched up with the appropriate paper, it will yield an appropriate print. In reality, a 1:1 relationship is not obtainable because of toes and shoulders on the characteristic curves, flare in lenses etc. There is also a question of aesthetic preference which seems to lean towards somewhat heightened midtone contrast.

Cheers, DJ

[Bruce Watson]

...I picked up a photography book the other day that said B&W film could record beyond 14 stops.

Way back when the first Tmax film was introduced (1980s?) I remember reading an article that said that Tmax was linear for more than 20 stops. Linear to well beyond a density of 4. Basically completely opaque outside of Kodak R&D labs.

But just because it can, doesn't mean you should. What would you hope to accomplish by generating that kind of density? Just the Callier Effect would be killer.

Indeed, if you're going to print on silver gelatin paper or scan, you're much better off with a Dmax less than 1.3. Some of the alt processes want more density, but nothing approaching 4.0 IIRC.

[swmcl]

Guys,

I should mention that I have no darkroom and do not intend to. I'm going digital for the printing. Therefore, my intention is to use the film as a recording device getting a full data-capture as it were. I just need to get things right in my head.

Cheers,

[Lenny Eiger]

Guys,

I should mention that I have no darkroom and do not intend to. I'm going digital for the printing. Therefore, my intention is to use the film as a recording device getting a full data-capture as it were. I just need to get things right in my head.

Cheers,

It doesn't actually work that way. When you "overdevelop" to spread the tones out farther you add a lot of contrast to the image. If you scan this film you will have a print that is too contrasty for most people.

I like to top out at about 1.4-1.6 for scanning film, more than many here like. That's quite a bit less than 4.0. If you want to make contrasty images, however, there's nothing saying you shouldn't.

I still use the zone system, even with the ability to correct things somewhat in the scanner and in PhotoShop, it is still better to get a perfect negative than not. There isn't that much leeway.

Lenny

[Jim Noel]

A simple test for DR and d-max.
Tape a step wedge to the film.
Expose on an evenly lit area like a north facing wall.
Calculate the exposure for Zone X.
Develop as you normally do.
A second sheet developed 40% longer will give you a higher d-Max, and with some film/developer combos, a longer scale.

[Jim Fitzgerald]

Steve, I will tell you my experience and I am by no means one of the scientific types. I just know what I see and have in my hands right now. I shot a scene last year in the Redwoods with 8x10 Efke 25 Developed in Pyrocat-HD 2:2:100 for 12:00 minutes.The readings are: 3.53-.50 for a net of 3.03DR. It makes an amazing carbon transfer print and the highlights are hot but have plenty of detail and carbon transfer can tame it in the print. I have several others and this neg runs the gamut of range from zone I to Zone XI possibly higher. To bad Efke is gone.

[swmcl]

So much for me to think about here. Jim, I assume your densest part is 3.53 and the least dense is 0.50 making a range of 3.03 ? I'm just sticking to normal processes but it does sound good I must admit. That to me is a piece of film able to record 10 stops or thereabouts (10 x 0.3 = 3.0). I'm primarily using FP4 in Pyro Glycol 1.5:1.5:100 (4.57ml per square inch of film) for 11 mins @ 20C rated at 64 and shadows on Zone 3.5 to 4 - which gives me something like a density of 1.6 at Zone 11 (?)

Is the scanner a linear device or a logarthmic one ? Surely it is linear ?

Lenny, in my ignorance I have assumed that a 12 or 14 stop image in front of me can be put on film. If it can be put on film then I can choose how to print it later. The important part is to get a useable negative - be it slightly over or under exposed for the main subject. If I can get every image to say from a 5 stop to 10 or 11 stops on film then I have freedom to change later. Its just that the datasheet says a Dmax of less than 2 ! Lenny I trust your judgements and take your advice seriously - thanks.

What is it when a digital photo is 'expanded' ? Is it a quantisation error - or approximation ? So you shoot a scene with an 8-bit sensor and then stretch the data to fit a 16-bit range you get the comb yes ? I feel it is like that when you shoot onto a film with a range of say 2.0 and then try to get 14 stops (@ 0.3) out of it ... It just feels wrong in a way but then again, it is an analogue medium not a digital one. My tests didn't get beyond 1.7 for FP4 and that was with 2:2:100 and 10 mins at zone 11. The film seems to show a linear response but at around 0.15 per zone (not 0.3).

What I take from this discussion is that I should shoot more and worry less !

I hope to get out this weekend.

Cheers all !

[Emmanuel BIGLER]

I have placed here a copy of the official specs fot the discontinued Kodak technical Pan film
http://cjoint.com/?CDDsfBXWDZK (http://cjoint.com/?CDDsfBXWDZK)

As you can see on page 10/12 of the pdf document, when processed with either Dektol or D-19 developer, you can reach a D-max equal to 4. But the gamma is extremely steep, above 3. On page 11, you can see the density when processed in technidol; with this very special chemistry you can get a value of .65 for the gamma, but the plotted diagram does not show any density values above 2.

Now that Kodak technical pan is discontinued, you can still find microfilms on the market as well as special chemistry to process microfilm as a continuous-tone negative with a low gamma value.
Among pupular European microfilm supplies for the hobbyist, you'll find Agfa Copex Pan Rapid, a classical microfilm; and packaged in rolls and 4x5" sheets, ADOX CMS-20. if you process Copex or ADOX CMS-20 in Dektol or D-19, sure that you'll get a D-max around 4. Low contrast developers for microfilms are available from SPUR or ADOX.

The Kodak T-max data sheet is here
http://www.kodak.com/global/en/professional/support/techPubs/f4016/f4016.pdf
curves shown do not climb above D=2.5 when you keep the gamma around 0.65. Do not forget that if you want to print the negative directly on silver-halide paper, you'll have hard times to compress all this dynamic range on any available printing paper. And scanners capable of a D-max=4 are not that easy to find.

Hence only Kodak people might have seen with technical pan or T-max a useable density range of 3 with a gamma of .65, yielding, in theory, an exposure range of 1 to 10^{3/.65} = 10^{4.6} ~= 40,000 i.e. about 15 f-stops.
20 f-stops = 1 to 1 million = 10^{6} hence the corresponding density range @gamma=0;65 = 6x0.65 ~= 4. A challenge to actually see this, but if Kodak people have achieved it with T-max, we can believe them.

[Emmanuel BIGLER]

Is the scanner a linear device or a logarthmic one ? Surely it is linear ?

A scanner + its software is a complete system, and you only get out of it what the manufacturers actually accept to deliver to the end-user.
Of course a silicon photo-detector has a wide range of linear response vs. incident ligh intensities; and within a range of, say 1:10000 = 10^4 with respect to incident light levels, corresponding to a density range of 4, if you had a direct access to the number of generated photo-electrons, you'll have direct-access to a linear light intensity sensor.
The problem is that most scanners are intended for continuous images aimed at human beings as the end-users, and our vision system demands a gray scale modeled like Saint Ansel's zones, i.e. displayed as a logarithm of the number of photo-electrons. Hence, very soon in the signal processing system, logarithmic output levels are computed from an input number of photo-electrons.
This raises an interesting question, for which I have no idea: is-it possible to have a direct access to a linear scale in any scanner, I mean: does the software, outside the scanner, computes the logarithmic scale, or some is the computation done by some internal processor inside the scanner, before sending the data to the scanning software and the end-user?

[Bruce Watson]

I should mention that I have no darkroom and do not intend to. I'm going digital for the printing. Therefore, my intention is to use the film as a recording device getting a full data-capture as it were.

Then high Dmax is the exact thing you want to avoid. Scanners, including drum scanners like I use, are at least as susceptible to Callier Effect as darkroom enlargers. What this means to you is that if you go higher in Dmax, your highlights will loose local contrast. And you'll have to correct for that in Photoshop. Another step that can be avoided (or at least minimized) if you expose and process your film optimizing for scanning.

You won't believe it. At least, I didn't. So run some tests. If you do it carefully and control your tests, you'll see it and find a "sweet spot" for your hardware and software workflow. I get best results with a Dmax of around 1.0 to 1.2. You're likely find your sweet spot is a Dmax anywhere from 1.0 to 1.5. You can go too low and see loss of tonality. And you can go too high and find all kinds of interesting artifacts, including halos around dark/light transitions, excessive (and really ugly) graininess, loss of resolution, and large loss of local contrast in the highlights that's difficult to impossible to recover in Photoshop.

B&W films are designed for darkroom printing. They give they best performance there. This shouldn't really be much of a surprise.

[Stephen Benskin]

Steve, it sounds like you are confusing negative density range with log-H range. The X-axis represents exposure or log-H which relates to the subject luminance range. The y-axis is the negative density range. A film gradient of just over 1/2 the subject contrast is considered normal. So 0.30 log-H units will translate into a density of around 0.15. When they say a film has a range of X stops, it means that the film can still separate tones (not yet reached gamma infinity) for a subject luminance range of X stops.

[swmcl]

Mr Benskin you are most correct. I am probably most confused ! It won't matter a whit to some extent because I am using FP4+ and some HP5 and that will be all for now. If those films can't give a high latitude then that's that. I think it was probably films of yesteryear that gave DMax's of 3 and above - perhaps the Tech Pans ? No Matter. I have a bit of a routine and I need to get on with it. Given some of my previous results I'd be glad to get an image of any sort !

Cheers.

[Stephen Benskin]

Mr Benskin you are most correct. I am probably most confused ! It won't matter a whit to some extent because I am using FP4+ and some HP5 and that will be all for now. If those films can't give a high latitude then that's that. I think it was probably films of yesteryear that gave DMax's of 3 and above - perhaps the Tech Pans ? No Matter. I have a bit of a routine and I need to get on with it. Given some of my previous results I'd be glad to get an image of any sort !


Here's an example I've been playing with. It's a projection of a seven stop subject luminance range onto the film curve (no flare). Maybe this can help in visualizing log-H vs density.

94292

[swmcl]

Stephen,

Your graph looks very similar to the FP4+ datasheet on page 4. The film would top out at around 9 or 10 stops.

I guess I should be honest with myself and say that I actually like a bit of contrast in my images anyway. The images I see posted with a 'large range' like the two bath Pyro thread are not to my liking I must admit. I think if I were interested in the whole range I'd take two or more photos and goto the HDR software.http://www.largeformatphotography.info/forum/newreply.php?p=1020804&noquote=1

Rgds,

[ic-racer]

I can't see how unless one were to interpret a 'stop' as being say a 0.2 step on the curve rather than a 0.3.

Film is usually developed to a gamma around 0.6 to 0.8. So you are correct. (ie 0.7 x 0.3 = 0.21)

[Larry Gebhardt]

Steve,

You are missing the point that negative film doesn't need to respond in a one to one way, with each stop of light in the subject adding .3 of density on the film. If you develop to a normal contrast for printing or scanning you will add about .15 density units per stop of subject brightness. So a 10 stop subject brightness range would result in about a 1.5 film density range. When you print the paper will expand this; the scanner will do the same. Within most subject ranges the response will be linear, just a lower slope.

Slide films are meant to be viewed by eye. They usually have densities over 3.0 in the black areas, and are difficult to scan. They are also much closer to a 1 to 1 density in the midtones (one stop SBR adds .3 density units to the film).

If you have a negative that captured a full 14 stops in a linear manner, and say was developed with a slope of .5 (.15 density per stop of scene brightness), you would have a 2.1 range on the negative (see Stephen's graph and not the XY relationship). If you scan this and adjust the black and white points to the clearest and densest parts of the negative you will get an image that looks similar to an HDR picture (14 stops squashed into a normal digital file doesn't look natural). I don't think the HDR approach of two images is really useful with negative films and 99% of the subjects out there because the films are able to record such a large range in a linear fashion.

Negatives like this are hard to print in a natural way without dodging an burning, at least in my opinion. Sure you can adjust the contrast of the paper, or use levels to adjust in photoshop, but the tones all get squashed and don't look natural (weak is how I think of them). If you dodge and burn in areas you can keep the contrast higher between adjacent tones, but move the areas with the brighter range down (but still with contrast). For example a scene with strong trees in the shadows but a very bright sky with clouds can be fit so all tones register on the paper as a straight print or scan. But all areas will look weak since there will be very little separation between adjacent tones (say needles of a pine, or two slightly different cloud tones). If you print with higher contrast in the trees to make them snap, the sky will be washed out. The trick is to burn the sky in (mask it in photoshop) and still use a higher contrast. Then you get correct separation between the sky and the clouds and correct contrast in the shadowed foreground areas. This is how your eye sees the scene. The trick of course is to dodge and burn things in such a way that you don't see the manipulations.

The two bath pyro negatives used to fit a large subject range onto the film in such a way that it can be printed on normal darkroom paper. The brighter areas end up being compressed more than the shadows, so the response to the subject isn't linear. That may be what you don't like about it.

Compressing the whole image range is one reason why I don't use the zone system to strongly change my film development in the negative direction (say N-3) even when metering indicates it's needed; I'm just going to need to increase the contrast during printing to make up for it and then burn in the sky. I know the film can hold the range and I'm going to need to dodge/burn it anyway to make it work for. The compensating developers like two bath pyro (which I haven't used yet) should in theory help because the contrast is only reduced in the highlights, leaving a normal range for much of the shadow and midtone range. So instead I just meter for the shadows (as always) and develop normally. If N-3 is indicated I will develop at about N-1 just to not risk rolling off the highlights.

Hope that doesn't confuse the issue more for you.

[swmcl]

Thank you Larry,

Your words tell me what I came to eventually. I was always assuming that one stop was 0.3 and it isn't.

Your response has cleared things up not made them worse.

So much of my learning is from books and posts to forums because I have no way to buy and sustain a darkroom I'm sure.

Sometimes too the explanations need to touch a nerve so to speak to be able to get through thick skulls. Yours has done that too.

Cheers,

[Ben Syverson]

I'm finding that normal B&W development can push highlights into a very high density, which can be difficult to capture with a scanner. I know that my scanner (Epson 4990) struggles with overexposed areas on HP5, which can bunch up and get almost totally opaque.

Because of this, I'm finding that slightly underdeveloped sheets tend to be easier to handle digitally... The shadow detail is still there; it just needs to be brought out with Curves. But the highlights aren't quite as pitch black, so I can reach more of that detail. My normal course from now on will be to cut development by a minute or two for shots with large areas of bright sky.

[Larry Gebhardt]

I'm finding that normal B&W development can push highlights into a very high density, which can be difficult to capture with a scanner. I know that my scanner (Epson 4990) struggles with overexposed areas on HP5, which can bunch up and get almost totally opaque.

Because of this, I'm finding that slightly underdeveloped sheets tend to be easier to handle digitally... The shadow detail is still there; it just needs to be brought out with Curves. But the highlights aren't quite as pitch black, so I can reach more of that detail. My normal course from now on will be to cut development by a minute or two for shots with large areas of bright sky.

I think many people advocate developing to a lower gamma if the intended use is scanning, especially on scanners like the Epson flatbeds. This will give you less pronounced grain, and should help keep the noise down in the highlights. The only downside I see is you may need to reduce your EI by a bit, but if the shadow detail is there then you are probably exposing correctly. For enlarging (diffusion) and scanning use I try to target a grade 2.5 to 3 paper, which tends to be a good balance for both.