Wow - read this re: Film versus Digital debate!

[bernardwest]

<i>Nothing to do with whether you can see it absorbing light or not, right?</i><p>

 

Of course this matters! If you can't <b>see</b> it absorbing light (ie. at printable levels of magnification), then how can that be represented in a print?

[bernardwest]

I should point out also, that the critical term used in my 10.55pm post is <i>measurement</i>. Because after all,

the negative and the print are just a forms of measurement of light.

[vijay_nebhrajani]

But you can't see individual atoms absorbing light, can you? Now I'm confused again.

<p>

<i>"If you can't see it absorbing light (ie. at printable levels of magnification), then how can that be

represented in a print?"</i>

<p>

I don't know - you tell me - if you try to put a neutral density filter under a microscope, you won't see any

grains - any particles of anything at all - at any magnification - so you can't see it absorbing light, can you?

But if you put it on an enlarger lens, the effect will be seen won't it?

[vijay_nebhrajani]

Re: your mail at 10:55 "Individual ones are most certainly invisible to light based measurement, but when you clump a 'gazillion' of them together, you can visually measure them."

 

Not sure what you mean - there are photon detectors that can measure just one photon striking them, so you could presumably make a sensitive enough device to measure the attenuation of a 150 nm particle of light if you wanted.

[bernardwest]

<i>But you can't see individual atoms absorbing light, can you? Now I'm confused again.</i><p>

 

That's correct. But that's an individual atom. But you sure can see a couple of trillion trillion trillion of them absorbing light. Do you not agree with this description?<p>

 

<i>if you try to put a neutral density filter under a microscope, you won't see any grains - any particles of anything at all - at any magnification - so you can't see it absorbing light, can you?</i><p>

 

I strongly suspect you would see particles with enough magnification. How do you think the 'darkness' gets onto the filter? Clearly there is <i>something</i> on there. What's your explaination for why a ND filter blocks light?

[bernardwest]

<i>there are photon detectors that can measure just one photon striking them, so you could presumably make a sensitive enough device to measure the attenuation of a 150 nm particle of light if you wanted.</i><p>

 

I feel like you are leading me into a trap here... but i'll take the bait anyway. There is no doubt you could make a sensitive enough device to measure the attenuation of light through a 150nm particle. They already exist. But the camera/film/enlarger/print isn't one of these.

[j._d._griggs]

At any given site that you might observe in a developed B&W film emulsion, you will see a mixture of silver particles and no silver particles. If you magnify further and further, you will reach a point where you see only silver or NO silver. At that level of magnification, the film is binary. There is no "middle gray" silver... only black. At this scale, there is either black or not-black.

 

The photographic process produces the formation of large variable density clusters of these tiny black regions. These clusters, luckily, relate in their form and position to the way in which light falls on the film. In this way, these variable density clusters do a wonderful job of simulating continuous tone (analog). I believe that is what Adams was saying, oh so long ago.

 

There is no good or bad to this. It just is.

[bernardwest]

There may be a problem with my last statement. And that relates to the particle size. I have already stated that I am unsure what the minimum size particle is that can be measured (in respect of light), and until a physicist pops in and clears this up, we are all just speculating (although, clearly one viewpoint is right, and one is wrong - at least in the supra-quantum realm).

[bernardwest]

J.D... welcome to the debate. This is something some of us have been saying. We have presented photographic evidence to support this, but the 'silver is not opaque' camp refuse to accept this, even though they have no evidence themselves.

[j._d._griggs]

"Can't see the forest for the trees."

 

I think, in this case, we have lots of trees!!!

 

 

:-)

[vijay_nebhrajani]

<i>That's correct. But that's an individual atom. But you sure can see a couple of trillion trillion trillion of them absorbing

light. Do you not agree with this description?</i>

<p>

Yes, I do.

<p>

<i>I strongly suspect you would see particles with enough magnification. How do you think the 'darkness' gets onto the

filter? Clearly there is something on there. What's your explaination for why a ND filter blocks light?</i>

<p>

Hmm, you're playing the role of educator at this moment, and I the student. But here goes, "my" explanation:

<p>

They work as dyes do - with no particles.

Now you could make a filter with carbon particles - but I'm referring to dye based filters, whose "darkness" comes from

the fact that the molecules of the dye absorb photons. There are no particles at all - you don't need "particles" at all, other than the

atoms or molecules of the substance.

<p>

You could make a thin film of gold (on astronauts helmet visors, for instance) and it is the atoms that would do the

absorbing. You wouldn't see individual gold particles at all - not unless you count individual atoms as the "particles".

There may be statistical coating variations due to the manufacturing process, but it is otherwise a uniform film of the

metal, with no "holes" and no "particles".

<p>

Certainly to absorb enough light to make it visible to the human eye, you need a vast quantity of atoms, but those atoms don't have to be

clumped together as specks or particles that are visible (or "image-able" with visible light) with clear spaces between them.

<p>

Does this make sense, or have I gone off the deep end again?

[bernardwest]

<i>Hmm, you're playing the role of educator at this moment</i><p>

 

Not at all. I'm playing the role of 'devil's advocate' to your devil's advocacy.<p>

 

I'm leaving work and will have to get back to this later. Quickly though, there are holes between the atoms in a thin film of gold. I don't know what the size is, or how that interacts with light, but it's wrong to say there are no 'holes'. And the reality is, that a thin film of gold, if it was able to maintain it's structure without having to stick it to a visor, could very well be considered a particle (just a very big one.). The other thing, you say atoms don't need to be clumped together as specks or particles to be visible. No one is arguing that, but that is one arrangement that makes them visible, just like clumps of silver are visible.<p>

 

More later.

[vijay_nebhrajani]

Bernie, of course there are holes between the atoms. Even the atoms themselves are vast empty space - a tiny nucleus

surrounded by distant electrons - if the nucleus were the size of a golf ball, then the first electron would be a kilometer

away. Holes? It's actually tiny bits of "matter" in vast vacuums.

 

But at a macroscopic level, or at a level of the wavelength of visible light, there are no holes of any sort in that film.

None. Whatsoever.

 

Now, a film like that probably wouldn't be able to maintain its structure if it were not stuck to glass, but thats because of

terrestrial conditions - in a vacuum with zero gravity and no other external influence, why not? Besides, if we are talking

about silver in a film, it is stuck to the gelatin, so the issue is not relevant.

 

If you want to call a quantity of such film a particle, you can - nothing says that a "particle" has to be tiny. But then a

similar particle, but tiny, and stuck inside a gelatin film would behave the same way; absorbing light and would be a "gray grain".

 

You realize that every single particle of silver in a photographic film grew from zero silver, right? That at one point in

time, it was all silver halide and no silver? Which means that depending on exposure and development, there would be a

miniscule amount of silver formed, or a large amount - as large as the entire crystal.

 

Now all silver deposits will behave identically, at least as far as absorption goes. Once they are there, they will absorb

photons that strike them. Larger silver deposits will absorb more light and smaller deposits will absorb less light. This

has nothing to do with whether we can "see" them absorbing light. If you wanted, you could build a device sensitive

enough to measure this absorption - but that is immaterial.

 

Once they absorb light, they contribute to the formation of tone. Tone, by its very definition is variation in the intensity of

light. All silver specks, large or small, visible of not, will contribute to this process.

[photo_smith]

J.D... welcome to the debate. This is something some of us have been saying. We have presented photographic evidence

to support this, but the 'silver is not opaque' camp refuse to accept this, even though they have no evidence themselves.

 

Well all the photographic texts on the subject, say silver is not opaque, CEK Mees, H Baines, etc

 

Ron who worked in Kodaks Research lab making emulsions, says they're not he has 15 patents inc Kodak gold film tech....

[bernardwest]

<i>But at a macroscopic level, or at a level of the wavelength of visible light, there are no holes of any sort in that film. None. Whatsoever.</i><p>

 

Firstly I am assuming by 'film' you mean the thin gold/silver film? Secondly, how do you know this? I'm not saying you are neccessarily wrong, but what are you basing this statement on?<p>

 

This argument basically gets down to whether a silver atom(s) on a negative is binary, or continuously toned. This is something I can't answer, so I can't really debate this point with you. There is no doubt that electrons in atoms 'intercept' photons, but I'm not qualified enough to say what this means in terms of <i>visible</i> light. If they are continously toned, as you suggest they are, then this would mean that b&w film should have resolution down to the atom level. If this was the case then, b&w print resolution would be only limited by 'resolution' of the enlarger lens and the photographic paper. I'm not sure what that translates to, but I suspect it would be far greater than what is achieved in reality. I guess it might act something like this: silver atom(s) precipitated onto film; while they may or may not transmit some light, that transmission would be overwhelmed by diffraction effects, making it moot.<p>

 

Then we get to the hypothesis stating that silver (particles?) is binary. If that is the case, then the observation we see under the microscope nicely explains that. So far you haven't given a valid reason why that microscope image might be lying to us.

[vijay_nebhrajani]

Bernie: "Firstly I am assuming by 'film' you mean the thin gold/silver film? Secondly, how do you know this? I'm not

saying you are neccessarily wrong, but what are you basing this statement on?"

 

Thin film creation processes like vacuum deposition and sputter coating are generally designed to create as uniform a

coating as possible. Yes, I meant the gold film.

 

Bernie: "This argument basically gets down to whether a silver atom(s) on a negative is binary, or continuously toned.

This is something I can't answer, so I can't really debate this point with you. There is no doubt that electrons in atoms

'intercept' photons, but I'm not qualified enough to say what this means in terms of visible light."

 

Atoms have no property of opacity or tone by themselves. But they do have the property that if a photon strikes them,

that photon can be absorbed. At this level, a photon is just a packet of energy, and it can give this energy to the atom.

 

In that sense, atoms themselves form tone, but only when there are lots of them together, and then too, depending on

arrangements and so on (for instance, you may take any number of carbon atoms, but if they are arranged as in

diamond, they will not absorb light). But the important point is, atoms are atoms - just protons, neutrons and electrons

(and maybe a hundred or more subatomic particles, I don't remember) - in fact it is from the arrangement and quantities

of these subatomic particles that things like silver or gold come about. Weird but true - the constituent particles have no

properties such as color or opacity or even whether they are metallic or not, hard or not - but arrange them in different

ways and you get metals, non-metals, gases, liquids, plants, flesh, you and me.

 

I already explained in depth why resolution of film is limited by the largest crystal size. Not silver specks and their size,

mind you, but the size of the largest silver halide crystals.

 

But yes, if you keep lowering the size of these crystals, then resolution would increase. But of course, you would run

into diffraction effects - the source content itself wouldn't have more information.

 

Also it is somewhat fallacious to think of silver deposits as having a uniform "gray" color - the process of silver forming

from silver halide is random and leads to filamentary structures. This is quite unlike a gold thin film; but the same

absorption mechanism holds. Because of the random shape, size and thickness of a single speck of silver, it could be

thick enough to absorb all light or absorb where it is thick enough and attenuate (reduce) where it is not thick enough.

 

As for the microscope, I've explained why the observation could be faulty. But here, Mark Smith sent me some scans

from Baines, "The Science of Photography", which clearly show what I've been saying.<div></div>

[vijay_nebhrajani]

Second scan from book, clearly showing gray grains.

[bernardwest]

I'm clocking off for the night, so will look at your scans tomorrow. I'm not sure about your microscope explainations. Unless I'm mistaken, we debunked all(?) of them.

 

I wanted to leave you with a thought: We both agree a single atom or, in your case, a tiny speck of silver, can absorb and transmit some light, right? Can you see a single atom or tiny speck of silver?... No. The moral of this story is that photon absorption/transmission is not neccessarily related directly to visible resolution. This is at the core of our argument.

[rishij]

Vijay,

<p>

Please see image below re: my assertion that the real-world resolution perceived for a certain imaging format is

inherently dictated by the tonal range that the image-forming element (grain or pixel) can represent:

<p>

<img src="http://staff.washington.edu/rjsanyal/Photography/FilmTonalityVsResolution.jpg" width=800>

<br>

<a href="http://staff.washington.edu/rjsanyal/Photography/FilmTonalityVsResolution.jpg">Link to Full-Size Image</a>

<p>

You're definitely gonna wanna view that at 100%, so follow the link to the full-size image.

<p>

On the left, you have a 4 megapixel image, viewed at 100%, containing a halftone image of a leaf generated with a

brush size of 600 pixels in Adobe Photoshop (generated in grayscale, then converted to halftone).

<p>

In the middle, you have a 16 megapixel image, viewed at 25% (like viewing from a distance), containing a halftone

image of a leaf generated with a brush size of 600x4, or, 2400 pixels in Adobe Photoshop. Think of it this way:

in this middle image, the 'grain' or 'pixel' sizes making up the halftone are effectively smaller, and more in

number.

<p>

To the right, you have a 4 megapixel image, viewed at 100%, containing a <i>grayscale</i> image of a leaf

generated with a brush size of 600 pixels (as for the first halftone image) in Adobe Photoshop.

<p>

This should be pretty self explanatory at this point, but, let me spell it out for you: In the left & right

images, the 'tonal range' of each 'image-forming element' (pixel in this case) is 2... black or white. In the

grayscale image on the right, the 'tonal range' of each 'image-forming element' (pixel in this case) is 256

(8-bit). The resulting image on the right is much better than (read: appears to be higher resolution when viewed

at a suitable distance) the halftone image, of the same number of 'image-forming elements', on the left. The

halftone image in the middle required 4 times as many 'image-forming elements' or pixels to generate something

that looked the equivalent of the grayscale image on the right.

<p>

SO, given the, AND THIS IS IMPORTANT, <b>same number of image-forming elements</b> to begin with (e.g. the same

number of grains per 35mm frame of film), tell me again how whether or not the image-forming elements are binary

(halftone) or analog has no bearing on viewed resolution at lower magnifications. Tell me, Vijay.

<p>

If you're talking about black and white lines, sure, fine, tonality and perceived resolution are not related

(because you only need 2 tones to represent said black & white lines). But if you're talking about real IMAGES,

explain to me how the tonal range achievable by each image-forming element is NOT at all related to the perceived

resolution... when my example above clearly shows otherwise.

<p>

Rishi

[vijay_nebhrajani]

Dude, Rishi, buddy - I will be eternally grateful - for what you have done is proved what I was saying - that film can't be a

halftone process, precisely because it does resolve so high at a contrast ratio of 1.6:1 - I posted twice from Norman

Koren's site an image showing 20:1 contrast. Take a look at it and see how finely differentiated the gray tones are at

20:1; now imagine 1.6 : 1; and tell me how film could resolve two tones, so close apart with such high resolution (50-80

cpmm as quoted by Daniel) if it were a halftone process. I presented an approximate information theoretic analysis, including numbers,

which I am quite confident are in the ballpark; but a picture speaks a thousand words.

 

Now to calculate the resolution limit - if you overlaid the rightmost image on the leftmost, you'd see what I mean; the big

dots all but kill all the fine information on the grayscale image, so for the purposes of calculating the resolution limit, all

you care about is the size of the largest crystal.

 

There is a subtle aspect of film that everyone who does such analysis misses - that is that resolution is limited by the

size of the largest crystal, and that tonality by the smallest.

 

The problem with film is that within a single crystal, the silver formation process is uncontrolled, and random; a

consequence of the information compression (the nature of the process), so even though the light intensity that fell on a

crystal and the development should have resulted in say 50% of the halide being converted to silver, you can't control

how those 50% atoms will form - they could be a thin film of filaments, or they could be a single, solid clump of silver.

So for the purposes of resolution, you have to go with the worst case - that you can't resolve finer than the largest

crystal.

 

On the other hand, smaller crystals form smaller (and even invisible under a microscope) specks of silver that improve

the tonality. Naturally they do this by "filling out the spaces" between larger deposits of silver (like a finer halftone

process overlaid on a coarser one); but they also do this by diffracting and by absorption, and by variable attenuation of

light. If they are thin enough, they behave like thin films (grayscale) and if not like halftone dots.

 

This is why I kept saying that you're confusing resolution and tonality. Of course resolution and tonality are related. They

are related in the exact same sense that sampling a signal is - you can't resolve smaller than 1 bit difference. Meaning

that if two tones differ by less than 1/65,636 of the dynamic range, both will be represented by the same number. To

differentiate even smaller differences in tone, you have to increase the bit depth. This has little to do with how many

pixels you put in a given area.

 

Resolution and tone are related as the size of the pixel and the change a single bit can represent, or in information theoretic terms, the

sampling frequency and the quantization step.

 

Once again, thank you for proving my case.

 

P.S. Rishi there is still one question for you to answer - the diagonal resolution thing for digital.

[rich_evans]

First off - I'm not taking any 'sides' in this debate - just trying to 'watchdog' any technical points which from my perspective

may need some clarification or the addition of relavent information in my area of expertise. I gotta tell you thought, there's

lots of valid points being made on both sides.

 

Vijay: "...if the silver deposit is thin enough, light will pass right through the silver, won't it? .... it will let light through

anyway, correct? .... You're the PhD in Optics - could we have a definitive answer please?..."

 

Correct - let's go back to my earlier post about the nature of the vapor deposited coatings on glass lenses. Those

coatings allow probably 99.999% of all the light falling on the first surface to pass through, because the lattice spacings or

gaps between the crystals of coating are large enough to attenuate the light ever so slightly (diffraction) and alter or

enhance the performance of the glass - ie to either increase refractive index AT THAT SURFACE or otherwise alter the

path of the light. In addition to that, many of these coatings are not the elements themselves, rather they are oxides or

other compounds which when grown to large enough sizes will be 'visible' as transparent particles.

 

ANY material of sufficient thin-ness is transparent to some degree. Take a piece of paper and hold it up to the light. Does

light pass through? Of course, despite the fact that there's a very thick layer of cellulose fibers usually doped with large

quantities of organic binders and other crystalline compounds such as Zinc Oxide, Titanium Dioxide - both of which are

excellent 'opaquing' compounds. You even asked why does the sunlight penetrate through two layers of film. At the

levels were talking about, few if any materials are TOTALLY LIGHT ABSORBING - maybe Carbon Black.

 

Silver, chromium, gold, or any other metal that can be used as a coating works the same way, thickness and

transparency depend primarily upon thickness of the coating (among other things). Remember that crystals have very

specific geometric form. Light (or any radiation) is altered by its interaction with the molecular structure of the crystal and

is altered in some way - slowed, bent, reflected. But at the molecular level, even these seemingly opaque materials

absorb and attenuate light.

 

Now as the crystals grow larger, they will become more opaque, finally reaching a size where light can no longer escape.

At this point, even sputtered or vapor deposited films have some 'dimension' to the crystals but the spacing between

crystals now allows light to travel through the substrate. Crystal growth as a result of almost any technique is a matter of

how the first atoms form on a substrate, usually as a result of some molecular dislocation on the substrate which is used

as a nucleation site for crystal growth. Over time, the crystals will grow in a somewhat controlled manner - not

necessarily uniform.

 

Anyway - I think I got off the topic, sorry. Below is a SEM image of evaporated gold on a carbon substrate taken earlier

this year for a resolution check of my instrument. The magnification is 400,000x and you can see the scale mark

indicating size. Note the significant amount of space still remaining between the individual crystals. This is a pretty thick

coating as films go, and its a pure metal - but its still visually transparent at this level. Oxide films look a bit different and

are much thinner.

 

"...Holes between atoms..."

 

Hmmmmmmmmm........as I recall, atoms are bound to each other by 'sharing' electrons - if by 'holes' you mean 'less

electron dense' than that's OK, but atoms don't just hang out by themselves.

 

AFAIK, atoms don't generally absorb photons - or if they do, they alter them or are altered by them. A photon striking an

atom will cause the atom to: A) release additional photons, B) release electrons, C) become deflected/accellerated -

changing its wavelength.

 

(Caveat - Rich is not an nuclear physicist)

 

Bernie: "...I am unsure what the minimum size particle is that can be measured (in respect of light),..."

 

The minimum particle size that can be ACCURATELY measured with light obscuration devices and/or light microscopes

is limited by the wavelength of the light used. Some laser-based devices can resolve and measure down to 0.01 um,

while the 'white light' (actually tungsten or halogen) typically used in general microscopy will be able to resolve no better

than 0.1um with the proper optics (high NA objectives, all infinity corrected optics). You'll see some digital/video

enhancement software that may be able to deconvolute the signal somewhat but for accurate measurements beyond

these limits, its all smoke and mirrors.

 

Is any of this helpful? I hope so.

 

--Rich<div></div>

[cameraart]

Here's my $.02

 

I saw an outdoor exhibit in Amsterdam in 2005, "The World From Above".

There were about 100 prints, each 4 x 6 feet. they were shot with Canon cameras on Fuji Velvia and printed on Fuji paper.

They were pretty impressive!

I had used Velvia since it appeared on the market.

Then I went into the inflatable building on the site, where they were selling books, etc. There was a 4 x 6 foot print on the wall with

the caption, "Can you believe this is digital?". I looked at it closely. It was better than the prints in the exhibition; finer grain and

sharper.

It was shot with a Canon 14Mpx camera.

 

That was it for me. No more film.

 

A friend said to me when i went digital, "But I love to look at those chromes with a loupe and see all that detail".

Yes, I agree, but it's detail you'll never need.

The largest I print is on 17 x 22 inch paper, and the prints are certainly good enough to look wonderful on the wall for a long time.

I don't need more than that.<div></div>

[rishij]

Vijay,

 

Thanks for taking my post entirely off-topic from its original purpose. It's purpose was not to argue whether film is halftone or not. It was to argue that IF film were strictly a halftone process THEN it'd require many more times the amount of 'image-resolving elements' packed into the same amount of space to represent the same element/object of an image with the reasonable contrast/perceived resolution that grayscale (256 levels) could produce.

 

That's why tonal range of the image-forming element and perceived resolution ARE related, WHEN the total number of image-forming elements in the 35mm frame is kept constant.

 

That was my only point. Because throughout this thread you keep telling me that tonality and resolution aren't related, that I confuse the two, that I don't understand binary, etc. And you may as well be telling me I can't tell an apple & an orange apart.

 

You still couldn't admit that you were wrong earlier in saying they're absolutely unrelated because, well, you're Vijay. And you hid behind a bunch of technical jargon near the end of our last post, but I'll just take this as your admission: "Resolution and tone are related..." and just ignore the rest of your post.

 

Again, I repeat, if image-forming elements can only be binary, for the same # of such elements within the same given area of the medium, you won't be able to 'resolve' subtle differences in tone, because you'll need a larger area with more such 'image-forming elements' to represent the various tones. So by increasing the area & number of image-forming elements, you're essentially increasing the number of bits that can represent any given tone... as long as you concomitantly step further back so that the spot of image-forming elements of increased area/density looks the same, in size, as your previous spot of image-forming elements of decreased area/density. In a halftone process.

 

Get it??

 

Rishi

 

P.S. Sure, your argument about resolution dropping along diagonals makes sense to me. Who doesn't know about the horrible 'stepping' seen in diagonal lines in digital images? Not sure what you're trying to prove, but, yeah I agree with you. More importantly, I'm man enough to admit when you're right. Unlike somebody I know... hmm who could it be?

[kolaczan]

Wow, just Wow.

 

So, have we figured it out yet folks? No? I didn't think so.

 

Sheesh, go take some photos and stop counting fairies on pinheads.

 

Holy Mackerel.

[rich_evans]

Hey - have we set a record with this thread yet? ;-)