Wow - read this re: Film versus Digital debate!

[James G. Dainis]

What can I say. My regular tripod carrier refuses to go out in the snow.<P>

<center><img src=http://www.geocities.com/dainisjg/bogen_jen9.jpg></center>

[rishij]

I can't believe you employ child labor.

<p>

If that's the cost of medium/large format, I'm sticking with my principles & morals.

<p>

And the small backpack of mine that carries all my equipment (minus my tripod):

<br>

<img src="http://staff.washington.edu/rjsanyal/Photography/Backpack.jpg">

<p>

Rishi

[James G. Dainis]

Rishi,

 

Not to worry. I don't pay her a dime.

 

What have we decided yet vis a vis the gray grains vs. only all black grains? Using the Zone system, extended development increases contrast or darkens highlight on the negative. In a low latitude overcast scene, if I develop normally, the highlights will appear as dark gray on the film. If I give 50% more development time the highlights will appear as black on the film.

 

Did the grains go from gray to black, which is my working way of thinking of it, or did the grain fibers just get thicker letting less light through so it appears darker?

[vijay_nebhrajani]

That's not child labor - that is simply quality bonding time with family - which even I do aplenty.

 

Grains - you got that right, and both statements are true in their own way. Any partially transmissive silver

speck is "gray" - whether this partial transmission is because you have a 5 micron grain with a 2 micron hole in

it or whether there's so little silver that photons simply sail through it, or because it is a solid lump but is

tiny enough that you get is a diffractive light distribution (and of course photon absorption).

 

Silver specks may be of any size, so all these effects are possible, and so there are "gray" grains. Indeed,

there is no reason to assume that silver specs can only be of a size large enough to be fully opaque, failing

which, they must be absent altogether, or their effect must be zero. Such an assumption requires some threshold

for speck size - and nobody can say what such a threshold might be - and why would it exist in the first place.

If such a threshold be defined on the basis of human vision through a microscope, then this means that the laws

of physics bend to human perception, which is absurd. In other words, if a speck of silver absorbs photons, it

does this regardless of its size, regardless of whether or not that speck is part of a photographic print and

regardless

of whether or not it can be perceived by humans.

 

If such a speck exists on a photographic negative and absorbs some of the light that is incident upon it, then it

is a "gray grain" by definition, regardless of whether or not humans can see it through an optical microscope.

There was

the argument that there can't be gray grains in any significant number, but this argument requires that silver

halide crystals be immediately converted to silver on contact with a developing agent, which leads to the

conclusion that you can't get finer grain, improved tonality and reduced contrast if you underdevelop (pull), or

the opposite effect if you overdevelop (push) which is of course untrue.

 

The natural process is that silver deposits increase in size from a few atoms to several billion as film spends

time in the developer. This implies that if you stop development early, you'll find silver in various stages of

deposition, and this would include "gray grains".

 

Nobody has come forward with a counterargument to this yet, and I continue to hope that Bernie will come around

soon. After what may be the longest thread in photo.net history, my open challenge still stands. Someone please

come forward and explain the above (namely, getting finer grain, improved tonality and reduced contrast with

reduced development based on a binary grain hypothesis). Nobody (other than Rishi) has come forward to admit that

they were mistaken

either, which is really sad. Admitting a mistake is a wonderful thing, simply because it means you really learned

something new, i.e., increased your level of knowledge and that you acknowledge this learning.

 

Sticking to ones ego for fear of appearing ignorant only guarantees continued ignorance.

[rishij]

Vijay:<i>"Any partially transmissive silver speck is "gray" - whether this partial transmission is because you have a 5 micron grain with a 2 micron hole in it or whether there's so little silver that photons simply sail through it, or because it is a solid lump but is tiny enough that you get is a diffractive light distribution (and of course photon absorption)."</i>

<p>

This is one of the most important summaries/paragraphs in this thread, I would have to say. Well put, Vijay.

<p>

But first, let me remind you all: in the end, it's not 'grains' that we're looking at at all... in the traditional definition of AgBr crystals being 'grains'. These crystals are dissolved away in the development process. But anyway, that's a technicality.

<p>

Vijay's description above begs a very good question: what does a 'gray grain' mean anyway? Gray when viewed from lower magnification? Gray when viewed up close at the level of a silver metal atom (not possible by the human eye!)? If we assume that one reduced 'black' silver metal atom absorbs a certain amount of the light incident upon it, then two such atoms will absorb more, three even more, and so on. So it becomes silly to ask if it's 'black' or 'grey'. The larger & more dense the deposit of silver, the more light it absorbs, & the blacker it appears from a distance. Now, silver clumps themselves probably consist of rather dense filaments (as EM images show us), because once you get the reduction reaction going at any sensitivity speck, it continues along many paths from the surface of the crystal towards the inside (that is, in most EM images, we don't see isolated long silver filaments isolated from other filaments, but that's not to say you *can't* grow such filaments under specialized conditions). But, light should be able to pass, by diffraction for example, through very small gaps. As well as large gaps between silver clumps.

<p>

So it's a number of things that determine how much light passes through any given spot in the film. But, most importantly, it's dependent upon the density of silver at that site. Importantly, though, as we've hammered out in this discussion, an entire grain is not ALL reduced to silver just because it had like one exposed 'latent' site on the crystal. I wish the more non-esoteric texts that people are used to reading would admit this and cover this a little better, because, as we've found out, they're entirely misleading.

<p>

Rishi

[vijay_nebhrajani]

<p>Well, I've been putting "gray grains" in quotes, and I keep using the rather long but proper term "partially transmissive silver specks" exactly because of this very reason.</p>

<p>Baines clearly shows a mechanism - just underdevelop your film - and you'll be able to see these partially transmissive silver specks, or loosely, "gray grains" under a microscope.</p>

<p>Now that you have Rishi's summary (hopefully this was it, Rishi - otherwise we are still waiting on it) Bernie, anything you want to say?</p>

[rishij]

<p>I've still gotta plow through Mees, but, generally I think we've come to an understanding, yes. With enough magnification, it's probably hard to see 'gray grains' if you're really zooming into silver specks. Because individual silver specks against film base may appear black... zoom out enough, though, and an adjacent silver speck that is more dense may appear blacker. Zoom out more and you start to see more and more tones due to a semi- sort of halftone process.</p>

<p>But in the end, I have to develop a roll of B&W and check myself.</p>

<p>I'll eventually post back with what Mees has to say... but don't hold your breath :)</p>

[vijay_nebhrajani]

<p><em>But in the end, I have to develop a roll of B&W and check myself.</em></p>

<p>Underdevelop if you do it yourself, or of you send to a lab, ask for pull processing. Expose normally for film speed though. This should exaggerate the effect, hopefully making it easier to see.</p>

[rishij]

<p><em>"Underdevelop if you do it yourself"</em></p>

<p>I'm 28, not of that generation, buddy :)</p>

<p>Anyway, I should have a number of highly unexposed regions (shadows) in my images of resolution charts, yes? So why would I need to underexpose to see underdeveloped silver deposits?</p>

<p>Rishi</p>

[vijay_nebhrajani]

<p><em>So why would I need to underexpose to see underdeveloped silver deposits?</em></p>

<p>Well, the process doesn't seem to be symmetric: I mean this - take mid gray; if you underexpose by a stop and then push by a stop you don't get the same results as when you overexpose by a stop and then pull by a stop. Both will get you the same density, but one will have finer grain, the other coarser.</p>

<p>I am not sure of exactly why this lack of symmetry exists - and I'm just saying that if we are to see "gray grains" the likelihood is higher if we underdevelop, so that by and large we don't have big "black grains" masking the effect of the "gray" ones.</p>

<p>It is almost as if the overexpose/pull will form the mid gray tone with "gray grains" and the underexpose/push forms them as a halftone process does. The former will make the elusive "gray grains" easier to catch. Then once you know what you are looking for, you can repeat the same experiment with normal development or even pushing.</p>

<p>Here's where your expertise with chemistry can help: why does this lack of symmetry exist in the first place? I don't know the answer, but this probably has to do with the chemistry of the process, at which I'm no expert. Perhaps you can explain it.</p>

[rishij]

<p>Sure, I have an explanation for this.</p>

<p>When you overxpose, and then pull, you get finer grain, yes?</p>

<p>When you underexpose, and push, you get coarser grain, yes?</p>

<p>Well, here's the mechanistic explanation as I would guess it to be. When you overexpose, you will either:</p>

<ol>

<li>Form more 'latent' sensitivity specks that previously had no clustered reduced metallic silver atoms</li>

<li>Grow larger 'latent' silver deposits at sensitivity specks that previously already had >3-4 clustered reduced metallic silver atoms</li>

</ol>

<p>I suspect that for any well-saturated site (well saturated as in: receiving lots of photons), method 1.) will start to prevail once silver specks have already grown to a certain size -- that is, beyond a certain size, it's likely that no more reduced silver will deposit at a given sensitivity site because further growth would be energetically unfavorable due to disruption of the lattice structure of the crystal. So let's say method 1.) starts to prevail as you overexpose. Even if it doesn't 'prevail', it at least <em>exists</em> as a method for increasing exposure.<br>

Even when you 'pull' process, I would assume that most 'latent' sites do begin development; however, you stop the development process early, so the reduced silver deposits that form are, in the end, <strong>smaller</strong> .</p>

<p>Now, when you underexpose, you get the opposite of the two possibilities above, namely:</p>

<ol>

<li>Form less 'latent' sensitivity specks</li>

<li>Grow smaller 'latent' silver deposits at sensitivity specks</li>

</ol>

<p>When you then 'push' process, perhaps you start development at a lesser number of sites, BUT, you prolong the development of any silver deposition (via reduction of silver ions) process for a 'latent' sensitivity speck, thereby, in the end, having <strong>larger</strong> but <strong>fewer</strong> silver filamentous deposits.</p>

<p>The resultant density for either the overexpose/pull or underexpose/push processes will be (can be) the same when viewed at lower magnification; however, the latter will consist of fewer but larger filamentous growths, whereas the former will consist of more but smaller filamentous growths.</p>

<p>Make sense?</p>

<p>-Rishi</p>

[rishij]

<p>Henning,</p>

<p>Have you seen this thread where Edward Brinker claims 32 lpmm resolution of lens+35mm film?</p>

<p>http://www.photo.net/canon-eos-digital-camera-forum/00Qvfs</p>

<p>How does this jive with your analysis and, if you disagree, where is the error in his calculation(s)?</p>

<p>Rishi</p>

[vijay_nebhrajani]

<p>Rishi, I hope Henning answers, but in the meanwhile, if I may point out a few things:</p>

<p>Edward calculates 32 cpmm resolution for lens + film; which isn't too unreasonable for many real situations, like handheld with a tele lens and so on. However, he uses 50-70 cpmm for film which is rather low; and 70 cpmm for lenses, which is also rather low. Controlled tests by Zeiss and others do show 100 cpmm on-film resolution with good lenses. Henning did point to several articles and papers from Zeiss in Mauro's resolution post.</p>

<p>And a couple of posts down, see Alan Rockwood's post. Essentially he also uses 100 cpmm for his calculations.</p>

<p>And as you yourself pointed out 14.4 cpmm is way too low a number for film + lens + scanner - Mauro posted many images that show much higher resolution than this after scanning. And you're right - Nyquist's theorem requires twice the sampling frequency, not thrice; but he may be considering three color dots in a triangular arrangement to form black - this would need two dot widths per black line and a dot width for the white line? Anyway - let him explain, I'm just guessing here.</p>

<p>See, this is the fallacy with trying to calculate "typical" or "average" numbers - one can truly only calculate limits with any manner of resemblance to reality.  Edward takes "typical" numbers and eventually uses them for a comparison. This is inherently flawed because one can so easily upgrade a lens, film or scanner and get a different set of numbers. Or I could take a 21 MP digital camera but put a crappy lens on it; I'll get numbers that can't directly be compared.</p>

<p>Meaningful comparisons will rely on limiting cases: like "This is the best I can do with MF film and this is the best I can do with an EOS 1 Ds Mk III; so in the limiting case, this is better or that is better". But in real world situations, you could worsen the results from either, so you could make MF look terrible by hand holding it, or you could make a 21 MP digital camera look terrible by smearing the lens with Vaseline. (That's not so far fetched - many photogs used a cheap UV filter for this purpose, to give special softening effects. A long time ago.) As you can imagine, such comparisons are rather useless.</p>

[luca_stramare2]

<p>Frankly speaking, those comparisons are interesting to me, but I am just wondering how much they will be useful. Why? Because we can still buy film and thanks god we will be able to do for a while, but from an R&D point of view film was dead in 2000. Apart from Agfa (went bankrupcy and resurrected as a consumer brand), Kodak and Fujifilm have cut their R&D budgets. The quality of film is set in stone, we are basically stuck to 2000 levels, with very few exceptions. The reason is simple, because the market went digital, there was no reason to further invest in film development. Actually, some of the films we were used to, like Kodachrome 25 and 200, Agfa RSX, Kodak Elite Chrome 50, Agfa Scala, ... simply disappeared.<br>

A lot of money is thrown into the development of digital stuff, instead. Therefore, even assuming that yesteryear film is better than today digital, it will be only a matter of time before things change because no film company is really looking to develop films that take advantage in the advances made in chemistry and materials over the last ten years. We might have digital sensors based on nanomaterials, we will never have a film based on nanomaterials.</p>

[rishij]

<p>Vijay, Bernie, DLT, Mark, whomever (wait a minute, whatever happened to DLT anyway??):</p>

<p>Did my last post above regarding overexpose/pull vs. underexpose/push make logical/chemical/whatever sense?</p>

<p>Would value your opinion,<br>

Rishi</p>

[bernardwest]

<p>Wow, there's been a bit of discussion since I was last in the room.  Rishi, something didn't quite sit right with my first reading of your summary.  I'm not going to give it the time now, as I am concentrating on 'that other thread'.  When I have time, I will give it more consideration and also catch up with what else has been said (I see Vijay has been naming me a lot.  Better have a look at that later as well...).</p>

[vijay_nebhrajani]

<p><em>Vijay, Bernie, DLT, Mark, whomever (wait a minute, whatever happened to DLT anyway??):</em></p>

<p>You forgot Rodeo Joe.</p>

<p><em>Did my last post above regarding overexpose/pull vs. underexpose/push make logical/chemical/whatever sense?</em></p>

<p>Made sense to me. I think thats a valid explanation - thanks for putting it forward.</p>

<p> </p>

[bernardwest]

<p>Rishi, I just had a good read of what you wrote, and it's making good sense to me now.  On first reading I was focussing on an idea (that I probably imagined you were talking about) concerning the larger crystals having more sensitivity sites available.  I was kind of thinking that if seperate sensitivity site 'share' the development of the crystal, it might lead to a greater number of filaments each with a smaller number of silver atoms in them.  But on thinking about that, this idea should hold for larger grained (ie. faster) vs finer grained (ie. slower) films.  Of course that doesn't happen, or we would all be shooting iso 3200! </p>

<p>So where are we now with black vs gray grains?  Does this explaination prove or disprove one or the other?  I'm thinking it doesn't, is that right?</p>

[rishij]

<p>I wasn't talking about larger or smaller crystals at all -- it's irrelevant to comparing overexposing/pull vs. underexposing/push (I think).</p>

<p>But since we're on the subject, we don't shoot ISO 3200 because the larger grains limit resolution. A photon striking a larger grain will knock an electron off a bromide, and then this electron could travel to any of the 100s or thousands of sensitivity sites on the surfaces (primarily) of this crystal. Right there you lose resolution, because there's no correlation between where the photon hit and which sensitivity site this electron will travel to within the crystal (of course, it won't travel to a different crystal... in the limiting case, if it could travel to any crystal, then film would have no resolution :)</p>

<p>Anyway, as for black vs. gray grains... I mean, there's no such thing as a 'grain' anyway after film's developed. Just silver deposits. I think the entire question/premise needs to be reworded!</p>

[bernardwest]

<blockquote>

<p>I wasn't talking about larger or smaller crystals at all -- it's irrelevant to comparing overexposing/pull vs. underexposing/push (I think).</p>

 

</blockquote>

<p>Yep, I figured.  I was dragged down the line of thought to do with larger crystals exposing first and how that might affect it, but I can see that probably doesn't play a part (i think...).</p>

<p>When I say black vs. grey 'grains', I of course mean 'silver deposits'.  It seems to me that the question still remains.  Can a clump of silver form a grey area on film (yes it can, coz we see grey in the negative at eye resolution)?  But, most importantly, what resolution is this greyness visible?  Is it at the few micron level, or is it at the 10's to 100's of micron area?  It seems your explaination doesn't exclude the idea that silver could be totally opaque (i think)... (I can hear Vijay excercising his typing fingers already.... ;)</p>

[rishij]

<blockquote>

<p>But, most importantly, what resolution is this greyness visible?  Is it at the few micron level, or is it at the 10's to 100's of micron area?  It seems your explaination doesn't exclude the idea that silver could be totally opaque</p>

 

</blockquote>

<p>You're right, this is the relevant question. But it's an uninteresting question to me, at this point. Because how 'opaque' is 'opaque'? If we define it as 'as opaque as a deposit that small can get', then what becomes the interesting question? Whether the gaps between filaments or the gaps between silver specks confer more 'translucency' (less opacity)? How many layers of dense 'silver filamentous growths' have to be stacked on top of one another to be 'truly opaque' since diffraction dictates that some light can pass thru anyway?</p>

<p>See, it just becomes one complicated mess... it's just not that simple. It's not as simple as 'binary' and it's not as simple as 'gray grains capable of representing thousands of tones'. It still takes an area of film of formidable size (what is this size? that's an interesting question) to represent thousands of tones, via what I believe to be a halftone process that also varies the size and density (though the latter, density, I'm not so sure about because most EM images I've seen show rather dense filaments as opposed to filaments growing independent of one another like the rays of the sun in a 5 year-old's drawing of the sun) of the silver specks (or filamentous growths).</p>

<p>That will be what limits real-world resolution. But this bottom-up approach won't give absolute numbers for resolution. On the other hand, Mauro's 'top-down' approach will :)</p>

<p>Rishi</p>

[vijay_nebhrajani]

<p><em>It seems your explaination doesn't exclude the idea that silver could be totally opaque (i think)... (I can hear Vijay excercising his typing fingers already.... ;)</em><br>

<em><br /> </em><br>

Ha, ha. Sure, some grains could be large enough to be "totally opaque" - but there is some difficulty in testing the opacity. I mean sure, it appears opaque under a microscope, but what if you held that thickness of pure silver foil up to the sun? Would it let some light through? No idea, and no way to test it either, so "totally opaque" just means "opaque" through an optical microscope.</p>

<p>What is really important is the understanding that silver deposits can be of any size from a few atoms to several billion, and there are sizes that won't be totally opaque even to the light from an enlarger.</p>

<p>But no matter, you're on the right track, Bernie; so I trust you'll eventually hit the truth - you don't need me to argue anymore. Which is really good, because I already have an appointment with an orthopedist for my wrist.</p>

 

[danielleetaylor]

<p>I can't catch up with or comment on all the posts since I left the thread, but I will toss in this tidbit.</p>

<p>From this link: http://www.newton.dep.anl.gov/askasci/phy00/phy00906.htm</p>

<p><em>You can see through metals if they are less than 0.1 micrometer thick. One-way mirrors are this. But if thicker, the part of light which is not reflected at the front surface will be completely absorbed. Metallic opacity.</em></p>

<p>I argued that if a silver deposit in B&W film were large enough to be observed using visible light, it would be observed as opaque. This quote (if accurate) tells us that a metal thicker than 100 nm will be completely opaque. It does not specify behavior just below this boundary, whether or not the metal shifts through a range of grays as it becomes thinner or simply goes transparent (except for reflection at the surface, i.e. the one way mirror example).</p>

<p>Regardless, I think that practically this information confirms my original argument. Silver is not going to develop out at a precise, fixed thickness right at or just below 100 nm. If a silver grain or deposit is large enough to see under an optical microscope, then you can bet its thickness is comparable to its height and width, and that all three dimensions measure much greater than 100 nm. Individual silver grains or deposits on film are opaque, black. They are not like individual pixels, which can record any level of tone.</p>

<p>This should not be a surprise as this is what we observe under a microscope, through an enlarger, and in print.</p>

[danielleetaylor]

<p>From what little I did skim of recent posts, it looks like there's also a side discussion on resolution. All I'll add is this: target contrast is everything when it comes to film. You can easily push film over 100 lpmm with high enough contrast. But that doesn't mean you are actually achieving that number in the real world, nor does it mean that your real world images will have more detail than they would have on a DSLR.</p>

<p>Case in point: Henning Serger, in his post, claims 100 lpmm for Fuji Neopan Acros. I'm very familiar with this film. For months I carried a DSLR and a film body loaded with Acros, and most of what I shot on digital I shot on Acros for comparison. I made prints in the darkroom, prints from scanned film, and prints from digital files converted to B&W. I have no doubt that Acros can out resolve a Canon 10D or 20D on a high contrast resolution chart. Never the less, the DSLRs consistently delivered real world images with just as much detail, resolution, and sharpness as Acros. Sometimes they would even exhibit a bit more detail in areas of low contrast and in shadows.</p>

<p>The relative performance between film and digital cannot be predicted using resolution charts. Or comparisons of pixels and grains. There's just more to it.</p>

[mauro_franic]

<p>Daniel,"DSLRs consistently delivered real world images with just as much detail, resolution, and sharpness as Acros. Sometimes they would even exhibit a bit more detail in areas of low contrast and in shadows"</p>

<p>Not true. In a different thread I asked you to post the examples you said you had. I am eager to see them.<br>

What is true: This is the 40D (which resolves more than the 20D and 30D) next to negative color film and black and white film. Look at the cans and spices next to the resolution chart and compare film to the 40D.<br>

<br /> http://www.shutterclick.smugmug.com/gallery/6616619_YJEwK#421902416_ibL4V-O-LB</p>