Wow - read this re: Film versus Digital debate!

[vijay_nebhrajani]

<i>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. </i><p>

You, in simple words that even a layman could understand, made the exact same argument that I was making in more

complex terms: "To give you an information theoretic analysis, if the silver specks act like they are sampling

the light/dark tones of a 1.6:1 contrast target, and storing the information only as 1s or 0s, the sampling

frequency has to be much higher and the quantization step much smaller; or the Nyquist frequency much lower. "

<p>

<i>Rishi: You still couldn't admit that you were wrong earlier in saying they're absolutely unrelated because,

well, you're Vijay.</i>

<p>

I only said that you were confusing tonality and resolution - but that was as far as finding out a resolution

limit was concerned. For the purposes of that, I chose the size of the largest crystal, <i>effectively assuming a

halftone process</i>. I didn't need to take tonality into account for this calculation at all.

<p>

Now if I said that tonality and resolution were absolutely unrelated the that was a mistake - just point out

where I said that, and I'll happily apologize to you and admit my mistake. Happy?

<p>

Now what technical jargon did I hide behind? I compared the situation with digital - the largest crystal size is

what you have to use if you want to calculate "pixel size", but if you want to calculate "bit depth" you have to

use the smallest crystal size. I said there was no way to use silver speck sizes for this calculation because

there is no predictability/controllability about how the silver forms.

<p><i>

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.

</i><p>

Yes, but this is the sum total of all crystals, large and small, and by using that number, you were coming up

with the argument "Film should resolve in the gigapixel range" - and that is why I kept saying that you are

confusing resolution and tonality. If you only had small crystals, perhaps, but the large crystals work to kill

information, so don't worry about total image forming elements even; just worry about the largest crystal size

for the purpose of calculating a resolution limit.

<p>

<i>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.</i>

<p>

Sorry man, I do apologize - I don't want to challenge your knowledge. I am sure you know a lot about a lot of

things. But I sincerely hope that your understanding of binary and resolution/tonality has deepened after our

discussions. My knowledge of chemistry sure has.

<p>

And still, do point out where I said that resolution and tonality were unrelated (out of the context of

calculating a limiting resolution) and you will have shown that I contradicted myself; if I did that, I would

have made a mistake, and you have my apologies.

<p>

At this point, do you see <i>how</i> Reichmann went astray? He used the size of the largest crystals and then

took 40-60 of these for the purpose of tonality. He should have taken 40-60 of the smallest crystals for

tonality, and the largest for resolution. Then he would have come up with some sensible figures - like maybe 100

cpmm at 1000:1 contrast target, and somewhat less, maybe 50-80 cpmm for a 1.6:1 contrast target. By taking the

largest crystals as the "bit depth" or quantization step, at 1.6:1, his theory would require resolution to drop

down to 10 cpmm, something not observed to happen.

<p>

And finally, about not being able to admit I'm wrong - dude, I'm married. I can admit I'm wrong and apologize

even when I'm actually not. Just kidding.

[vijay_nebhrajani]

And Rishi, to give credit where it is due - you were one of the only people in this debate who not only admitted he was wrong, but also course corrected and actually tried to find what the truth was, rather than merely arguing.

 

Thank you for that. If not anything, that should serve as a model for others to emulate.

 

About the diagonal resolution thing; if we wanted to calculate the resolution of a digital sensor independent of subject orientation, we'd have to use the formula (number of x pixels * number of y pixels)/1.4 - so a 21 MP camera is actually 15 MP. I'm just saying, digital still has some distance to go before it catches up with film in terms of raw resolution (even at 1.6:1) - but as far as tonality is concerned, nothing beats digital.

 

If I had to take sides, it would be the side of digital. I love film too, but I design digital stuff for a living - it pays the bills, so if I had to shill for something, I'd shill for digital. It is funny - you said you wanted film to win; and I am saying I want digital to win.

 

If I could buy say a 30-35 MP camera for about $2000 I would abandon 35mm and perhaps even MF altogether. I still would use film for LF though.

[rishij]

FYI:

 

This entire thread takes up 373 pages at size 12 font, single-spaced, without any of the pictures.

 

There are approx. 160,000 words. At 20 words/minute typing speed (OK I can do like 100-120 words/minute, but not when typing this sorta thread posts), that's approx. 133 hours worth of typing. This thread has been active for 15 days. That yields about 9 hours worth of typing going into this thread EVERY DAY since its inception.

 

And that doesn't even take into account the countless more hours that went into thinking & researching the posts before typing them.

 

This is beyond a thesis, at this point. And really should set some sort of record.

 

Rishi

[stephen_pitts]

RIch and Vijay,

 

Thanks "a heap (or grain)" for spending so much time explaining some of the basics here. You guys have made this

thread a vary valuable reference for those that truly want to be able to compare the two. Not to disparage the

opposition here too much, but I am frankly surprised at some of their commentary (particularly poor Daniel going on

about binary -- every system in the world has finite states of existence ala quantum theory but that does not mean

they are all binary). But let me just add one last element to the puzzle of image quality here. Many scientists have

found that noise is a powerful tool in signal processing -- i.e. the addition on noise actually enhances one's ability to

process a signal (Scientific American did a great article on this). The power of digital is its reliable repeatability; the

power of analog is the incredible fidelity to the information content of the source material. If the sampling rate is high

enough (and it needs to be really high), digital can approximate analog and, of course, is much more replicable.

 

Size for size, analog will actually have greater information content than digital despite the opposite being preached

fairly regularly -- but the visual "cleaniness" (I mean this in a positive and not pejurative way) of digital is appealing to

many; for others, the richness of the analog reproduction dominates other considerations.

 

The intersection of the two technologies is today limited by the size and cost of the lens. The beauty of traditional

analog lenses is that they could perform very well at a reasonable size (and cost); to try to get a digital reproduction

to match its analog counterpart, a huge stress is placed upon the lens. The light rays from the back of the lens

must strike the surface of the digital receptor as perpendicular as possible; in contrast, analog has many layers and

so can accomodate a fairly oblique ray. To satisfy the requirements of the digitial sensor to match an analog

recording medium, you must therefore have quite a bit larger lenses. Of course, once you do, you are free of the

noise or "grain" of the analog system.

 

My true suspicion is that most photographers enjoy the convenience of digital over analog much the same way as in

audio. Convenience for its own sake should be admired but to then somehow magically claim that it is both more

convenient and better in image quality -- well thats whats called a free lunch. I'm hugely into both systems (more

ditigal lately) and still think nothing still comes close to large format analog photography (which, of course, is the one

type of photography I just never seem to have time for).

 

Go figure.

 

Cheers,

 

Steve

[bernardwest]

<i>He should have taken 40-60 of the smallest crystals for tonality, and the largest for resolution.</i><p>

 

I'm not following your reasoning for this. Can you put in it layman's terms?

[rishij]

Thanks Vijay. Yeah, I've learned a lot. So have you, along with a lot of other folks. If I remember correctly, earlier on we were talking about 'grains' until I pointed out that that is nonsensical for developed film because crystals do not remain; only filamentous growths do (as you yourself pointed out, or guessed, from Rich's beautiful image). And then you switched over your terminology from 'grain' to 'filaments'. So we've all learned; let's not pretend like one of us has been right all along.

 

As for being wrong -- I don't mind admitting when I'm wrong. Shamefully, glaringly so, in some cases (see http://www.photo.net/digital-darkroom-forum/00RTYM). Don't really know if I'm been 'wrong' per se in this thread re: the analog vs. binary nature of grain, because I've been careful to admit along the way, at each step, that I don't know which it is, but here are arguments for A and here are arguments against B, and vice versa. But you're right that while I started off defending the 'full reduction of a grain' camp, I didn't hesitate to jump ship when evidence mounted to the contrary -- that reduction works from the outside of the crystal inward, from sensitivity sites, and can stop at any point along the way. The more the sensitivity sites 'latent' to begin with, the more exposed the particular crystal is, and the more likely that, given a certain development time, more of the silver within the grain will be reduced.

 

Anyway, all that aside, I'm still not entirely convinced that the halftone process plays no part in the photographic process of film. As for Reichmann's argument, I don't know that we should be talking about 'his' argument as I seriously doubt he put in anywhere near the amount of time & thought as we are.

 

Vijay, I like your argument that resolution is limited by the size of the largest grain. So, what is the largest size grain/crystal in high resolution Velvia or Provia film?

 

Vijay, re: the smallest crystals limiting tone -- that's assuming that the largest crystal can assume the largest tonal range of any grain. Which, duh, of course makes sense. But, can the largest crystal assume the largest tonal range possible for one layer of film? Definitely not for the entire film, because grain stacking across multiple layers increases the total tonal range coverable by any given surface area of film. But, what I'm asking is, if we were to just get rid of all the other layers of film and just work with ONE layer, would one fully exposed LARGE grain leave behind a silver deposit that, when viewed from a distance, looked like the darkest spot across the entire layer of film? Or would a clump of deposits from 100 such fully exposed grains look darker?

 

I'm just trying to figure out if a halftone process is involved at all. I.E. what IS the total tonal range representable by the myriad of silver deposits that can be left behind by the largest grain in the film?

 

How do we approach this problem & solve it?? I'm at a loss.

 

Rishi

[bernardwest]

I'm confused now. All along, I have been basing my position on the 400x microscope image, and the literature which Rishi pointed to at the beginnings of this debate. Unfortunately I have been caught up in arguing about things at the atomic/molecular level, which I am not qualified enough in, and I believe Vijay is not qualified enough in either going by some of his contradictory lines of reasoning. But now we have other literature saying that a grain is not neccessarily converted to fully black or left clear, ie. not binary. We also have an image which purports to show grey grains. Although, I'm not totally convinced by this image because of the, ironically, halftone printing process, interfering with interpreting the image.

 

Ultimately for me, this gets back to the need for some more microscope images showing grey grains. This shouldn't be hard to find if grains are tonal after development. Surely some of you geeks can arrange this, couldn't you?

 

Rishi, I haven't been taking in all the nuances of your hypothesis, but on initial consideration, I can't see how this can be BOTH a halftone and continous tone process. If you agree that a grain (or more precisely what's left in the space it occupied) after development can be a shade of grey, this tone and the surrounding ones should be able to be represented in the print (giving due consideration to the resolution of the enlarger and paper, of course).

 

As for the confusion I talked about, I don't think it is really necessary to debate this out at the atomic/molecular scale. Someone, somewhere, must have some convincing photomicrographs which show grey grains. If not, then surely amongst us, we can arrange this. Rich... duty calls. Start taking some more images... ;)

[bernardwest]

<i>But, can the largest crystal assume the largest tonal range possible for one layer of film? Definitely not for the entire film, because grain stacking across multiple layers increases the total tonal range coverable by any given surface area of film</i><p>

 

I'm not following your reasoning here. If one crystal can assume a range of tones from clear to black, then how can that range be extended, as you seem to be suggesting? Also, don't forget that while the interpretation of tone on any given area of film will be dependant on the 3 dimensional arrangement of grains, so will the exposure of those same grains.

[rishij]

<i>"We also have an image which purports to show grey grains. Although, I'm not totally convinced by this image because of the, ironically, halftone printing process"</i>

<p>

HAHAHAHA! Easily the best line in this thread... sums it all up. The unending confusion... :)

<p>

Bernie, my catering to both camps is just my way of dealing with cognitive dissonance, haha :)

<p>

No ok seriously. It could surely be both analog & a pseudo-halftone process in which the sizes of the dots (silver deposits) can be variable (which we've already all agreed to, including DLT, which, incidentally, sounds like the name of a sandwich), AS WELL AS the opacity of the dots.

<p>

I mean, in the end, isn't it just a game of semantics? At the level of the silver atom, it IS halftone... Sort of. Not really. BAH! I have to go make myself useful.

<p>

Rishi :)

[bernardwest]

<i>It could surely be both analog & a pseudo-halftone process in which the sizes of the dots (silver deposits) can be variable (which we've already all agreed to, including DLT, which, incidentally, sounds like the name of a sandwich), AS WELL AS the opacity of the dots.</i>

<p>

I see what you saying, but I can't visualise how this would work. Rishi, you are the master of diagrams... well, get cracking then... ;)

[vijay_nebhrajani]

<i>"We also have an image which purports to show grey grains. Although, I'm not totally convinced by this image

because of the, ironically, halftone printing process"

<p>

HAHAHAHA! Easily the best line in this thread... sums it all up. The unending confusion... :)</i>

<p>

Dude, Bernie - the images to the left and right, same printing process, same book, same pages, have no problem

showing other tones - such as black and darker gray.

<p>

That is a picture showing a state of silver specks between transparent and fully opaque. Is this not proof enough?

[vijay_nebhrajani]

<i>Rishi: "No ok seriously. It could surely be both analog & a pseudo-halftone process in which the sizes of the

dots (silver deposits) can be variable (which we've already all agreed to, including DLT, which, incidentally,

sounds like the name of a sandwich), AS WELL AS the opacity of the dots." </i>

<p>

Actually Rishi's pictures of the maple leaf help visualize it perfectly. Dude, those images are sheer genius.

<p>

Let me explain.

<p>

First visualize separating the silver halide crystals (that in real film may be dispersed randomly) into layers by

size. Thus, visualize one layer that has the largest crystals, then another layer that has medium ones, then

another than has the small ones, then one that has the extra small ones and so on.

<p>

Now imagine the image being formed by all layers. Then overlay these images on top of one another. The layers

with the finest crystals will give you the best, the finest tones, but the layers with the largest crystals will

kill resolution the most, by obscuring and fudging any detail that may be contained in the finer layers.

Unfortunately you need those big crystals to get you speed, otherwise you'd still be asking your subjects to stay

immobile for minutes.

<p>

Thus you get how it works - the largest crystals limit resolution, and the smallest crystals limit tone.

<p>

By the latter part of the statement I mean that you can't get any better tonality than what the finest crystals

afford you.

<p>

Or the digital analogy - the largest crystals act like pixels, but the smallest crystals act like "bits".

<p>

And Rishi, your images have given me the insight to prove Reichmann wrong <i>even if film were a true halftone

process</i> - it is organized like I described, so he can't go about taking the largest crystal as the resolution

limiting factor, and <i>also taking the largest crystal as the tone limiting factor</i>. I know that Reichmann

doesn't matter at this point, but it is so much fun to debunk his pseudo scientific nonsense that I can't resist.

The fun part is that disproving him doesn't even require us to change the opacity of the dots - so actually I

have no argument about that even - Reichmann is wrong in either case. I'm loving this.

<p>

(Why my obsession with disproving Reichmann, you might ask - man it is because of his article that this whole

thread even exists. Were it not for him, I'd be doing something productive, like sleeping.)

[vijay_nebhrajani]

<i>Rishi: I'm just trying to figure out if a halftone process is involved at all. I.E. what IS the total tonal

range representable by the myriad of silver deposits that can be left behind by the largest grain in the film?</i>

<p>

It is theoretically infinite. Or more precisely, it would be theoretically infinite if you could control the

process such that the silver formed as a "disk" of silver, growing only in thickness as exposure/development time

increased.

<p>

In practice, this is not the case, of course, because you have information compression - so there is no guarantee

and no control over the final shape of the deposit. (I guess film makers want some control over this, hence

tablet grains with dopants on the sides and whatnot, but for our discussion, ignore that). It could be disk

shaped (like platelets), or it could be like a

spherical ball with filaments poking out - like that toy rubber ball (I forget what its called).

<p>

Because of this, the tonal range that that silver deposit represents has no relation to the light that fell on

it. But, and here is the important part - the total amount of silver in that deposit is proportional to the light

that originally fell on the crystal. More light - more silver. This <i>is</i> like varying the size of the dots;

but with a difference: halftone dots have no thickness; the only way to make a bigger dot is to increase its area

- but with silver, it is three dimensional: a silver speck can grow in thickness as well, so "more light = more

silver" may translate to "more opaque silver", rather than "same thickness but greater area".

[bernardwest]

<i>That is a picture showing a state of silver specks between transparent and fully opaque. Is this not proof enough?</i><p>

 

If I grant that one, then we have one for each side of the debate, plus a whole lot of anecdotal evidence from Daniel and Adams. The scale is still tipped towards the binary camp. At the very least we need one more incontravertable image to play the role of tie-breaker.<p>

 

<i>....Thus you get how it works - the largest crystals limit resolution, and the smallest crystals limit tone.</i><p>

 

Ok, that makes some sense. But is it not conceivable that a silver filament (by this I mean a clump of silver that has formed within ONE grain) from a large grain could actually be larger, and let's say more opaque, than the largest silver clump that could be possibly formed by a very small grain? If this is the case, then in the developed negative there would be no way to determine what sized crystal was responsible for a particular silver clump. Does that make sense?

[vijay_nebhrajani]

<i>"If I grant that one, then we have one for each side of the debate, plus a whole lot of anecdotal evidence

from Daniel and Adams. The scale is still tipped towards the binary camp. At the very least we need one more

incontravertable image to play the role of tie-breaker."</i>

<p>

It um, doesn't exactly work that way. If you insist that something is one way, then a single exception disproves

the rule. Read the entire text alongwith the images - and also see the other page where the text clearly talks

about the thickness of silver deposits - how twice the thickness allows 1/100th the light and this is density 2

on a log scale etc.

<p>

<i>"It has been mentioned that if development of an emulsion grain is observed under a microscope it is seen to

start at one or more points on the surface, and to spread therefrom throughout the grain (p.112). Under these

conditions, once development has started, it proceeds with rapidity, and the whole grain is almost immediately

converted to metallic silver. It was, therefore, concluded that a partially developed emulsion must consist

largely of completely developed and completely undeveloped grains, with very few partially developed grains.

<b>This conclusion was erroneous</b> and based on the assumption that the conditions on a microscope slide are

similar to those in the body of an emulsion layer. <b>A partly developed film or plate does in fact contain a

high proportion of partly developed grains --- if it did not do so, there would be no such thing as fine grain

development</b>."</i>

<p>

The boldface is mine.

<p>

Essentially the exact same thing I've been saying - not a binary process because it takes time, hence you can

stop in the middle, hence you can get partly transmissive grains etc. Then there is a picture to finally add some

icing to the cake.

<p>

What say, Bernie?

[vijay_nebhrajani]

"<i>Bernie: Ok, that makes some sense. But is it not conceivable that a silver filament (by this I mean a clump

of silver that has formed within ONE grain) from a large grain could actually be larger, and let's say more

opaque, than the largest silver clump that could be possibly formed by a very small grain? If this is the case,

then in the developed negative there would be no way to determine what sized crystal was responsible for a

particular silver clump. Does that make sense?"</i>

<p>

Sure, if you have a larger speck of silver, regardless of where it came from, it would occlude a smaller one - if

the larger one were completely opaque, the smaller one would make no difference. If the larger one was "thin and

spread out", then the effect would be like a small ND filter placed on a large ND filter, and so on.

<p>

What matters is the symmetry of the process - first it is a macroscopic process - so we postulate that it is

symmetric with respect to size - so if the exposure and development was such that 50% of the halide in a crystal

gets converted to silver, then this should hold true regardless of size of the crystal.

<p>

I'm sure that in practice, there will be issues with that, since there is gelatin involved, and the local

micro-concentration of the developer is not exactly constant (hence the agitation in the development process) and

so on; so yes, you will have random fluctuations in practice.

<p>

The overall concept is still the same though.

[bernardwest]

<i>It um, doesn't exactly work that way. If you insist that something is one way, then a single exception disproves the rule. Read the entire text alongwith the images </i><p>

 

I could reflect the same thing back onto you. In fact, I will. There is "credible" evidence from both sides. We need either our own controlled experiment, or someone above reproach who is clearly a master in this field.

[bernardwest]

<i>Sure, if you have a larger speck of silver, regardless of where it came from, it would occlude a smaller one </i><p>

 

That's not my point. I'm not talking about occlusion of small grains. What I mean is that when we look at a developed negative, there is no way of telling the initial grain/crystal size, is there? (I am assuming that all evidence gets washed away). If this is the case, then there is no way to categorise 'tonal' silver deposits vs 'resolution' silver deposits. In this case, you minimum resolution (in respect to tones as well) would have to default to the largest initial crystal size.

[vijay_nebhrajani]

"In this case, you minimum resolution (in respect to tones as well) would have to default to the largest initial crystal size."

<p>

Not sure what you mean. Minimum resolution is zero - don't expose the film. If you mean maximum resolution, then yes, that has an <b>upper</b> limit - you can't get any better than what you could get with the biggest deposits alone. Since there is no way to know what the size of the biggest deposit is, I suggested we go with the size of the biggest grain of silver halide, since the silver formed from this is smaller, or equal to the crystal size.

<p>

That's exactly what I've been saying - "largest silver halide crystals limit the resolution". You can't get any better resolution than what the largest crystals afford you.

<p>

Imagine the overlaying of images and you'll be able to visualize it. If you overlay Rishi's maple leaf images, resolution can be no better than the leftmost one (with large dots) and tonality can be no better than the rightmost one (with grayscale or really fine dots).

<p>

I postulated <b>upper</b> limits - you can surely get worse than that. Try underexposing by 3 stops and pushing 3 stops your favorite fine grained film.

<p>

Thus the original point I made was - "If you can get no better than <b>this</b>, then as soon as digital matches <b>this</b>, its game over. No more debate about which is better - we can then go be happy."

[rich_evans]

One possible explanation that I can propose re: different development times and level of Ag conversion is that the smaller the film grain (or any grain/particle matrix), ie. Pan-X < than Plus-X < than Tri-X, is that as particle size decreases, area increases exponentially. Thus the increased area of the logorithmically increased number of smaller particles pull more of the active developer agent from a soup - all other things being equal. Just sayin' - does any of this make sense? Is it relevant? Any chemists?

 

I'll be only too happy to take another stab at finding some way to more clearly define the nature of the grains - its going to have to wait as I'll be away for the next few days. When I get back to the lab on Monday, I'll try to dig up some fresh B&W film and try to make cross-sections through the emulsion/base to see what I can see.

 

Everyone have a good weekend,

 

--Rich

[vijay_nebhrajani]

<i>"I could reflect the same thing back onto you. In fact, I will. There is "credible" evidence from both sides.

We need either our own controlled experiment, or someone above reproach who is clearly a master in this field."</i>

<p>

We could argue semantics if you want at this point, but its like this:

<p>

I see black, so I postulate that black is the only color. One single exception where you show me gray disproves

what I am seeing.

<p>

as opposed to:

<p>

I postulate that there are all sorts of colors, from white to black, including all shades of gray. You showing me

black doesn't disprove me.

<p>

As for controlled experiments or masters, I'd say go for the controlled experiment, and the Baines book is pretty

authoritative.

[vijay_nebhrajani]

I said: "I see black, so I postulate that black is the only color. One single exception where you show me gray disproves what I am <i>seeing</i>."

<p>

I meant: "I see black, so I postulate that black is the only color. One single exception where you show me gray disproves what I am <b>postulating</b>."

<p>

Sorry for the error.

[bernardwest]

<i>I postulate that there are all sorts of colors, from white to black, including all shades of gray. You showing me

black doesn't disprove me. </i><p>

 

Well it does if there SHOULD be grey in there. Your reasoning is silly anyway. If we followed that sort of rationale in

the scientific process there would be very few things we COULD prove. As we keep trying to tell you, the impetus is

on YOU to show that grey exists, as it is <b>impossible</b> for us to prove that it does not exist.<p>

 

Regardless, as you say, a semantic argument is not what's needed here. I think the controlled experiment is the

way to go. Who's this Baines character anyway? And why should I defer to him?

[vijay_nebhrajani]

<i>Bernie: "That's not my point. I'm not talking about occlusion of small grains. What I mean is that when we

look at a developed negative, there is no way of telling the initial grain/crystal size, is there? (I am assuming

that all evidence gets washed away). If this is the case, then there is no way to categorise 'tonal' silver

deposits vs 'resolution' silver deposits. In this case, you minimum resolution (in respect to tones as well)

would have to default to the largest initial crystal size."</i>

<p>

Oh - I see what you mean now. You're basically saying - "if I assumed that the largest crystal size is "binary"

to calculate resolution for high contrast targets, why would the same not hold for low contrast targets?"

<p>

Well, because the amount of silver from that large crystal is proportionally smaller as well. It has either

become a smaller speck, or a spread out speck; not interfering that much with the specks from smaller crystals,

which also

"resolve", of course. But being a random process, it still will interfere to some extent - hence you will see

drop off of resolution as contrast is decreased, but not as much as if you assumed it could only be entirely opaque.

<p>

Reichmann says that it would take 40-60 specks to form a tone. Even if you only agreed with the halftone process,

it'd still be an "overlaid" halftone process, so 40-60 of the <i>smallest</i> specks would form a tone by

themselves. You can't then use the size of the <i>largest crystal</i> for the purpose of a resolution

computation. Like you said, who knows where a speck came from?

[bernardwest]

<i>Imagine the overlaying of images and you'll be able to visualize it. If you overlay Rishi's maple leaf images, resolution can be no better than the leftmost one (with large dots) and tonality can be no better than the rightmost one (with grayscale or really fine dots). </i><p>

 

Sorry, 'Minimum' wasn't the right word. That should have been 'maximum', or what I really meant - 'smallest' resolvable feature.<P>

 

What I am trying to get at is that the largest grain size might ASLO be the best resolution for tones as well. You say the smallest grains can represent tones. But the term 'grain' has no meaning after the development process as they are all washed away (is this correct?). Hence you can't distinguish silver deposits belonging to small grains or large grains. You have to assume for resolution (including tones) that they came from the largest grains. But I am thinking this might not be the correct approach to this, as with a test chart you can map the image on the print to the image on the chart, and clearly determine resolution that way. I am quite possibly rambling. Just thinking out loud.