Wow - read this re: Film versus Digital debate!

[bernardwest]

By Vijay's logic, the view through a microscope with an ND filter that blocks 99.9999999% of the light will still look white. Absurd? Methinks so.

[vijay_nebhrajani]

Bernie, I thought you were going to keep out of this? Anyway, if you want in, I'll answer your question, I

promise. <b>But I asked you to do the actual experiment before you asked me to do the thought experiment.</b>

Please do

it, come back, lets talk a little about

your experiment and the results you got and then we'll move on to your thought experiment. Or you could do either

of the experiments I asked Daniel to; doesn't matter. Lets go in chronological order please?

<p>

If you answer a question with a question how can we proceed forward?

[vijay_nebhrajani]

<i>By Vijay's logic, the view through a microscope with an ND filter that blocks 99.9999999% of the light will

still look white. Absurd? Methinks so.</i>

<p>

OK, lets go with that. A fully exposed negative is black, right? It blocks a great deal of light, right? It is

quite close to your 99.9999999% ND filter, right?

<p>

Now put one of these over your eyes and go look at the midday sun on a clear day. Do you see a gray sun or a

white sun? (By the way, be careful when doing this, you could go blind, and I refuse to accept any responsibility

if something should happen to you.)

<p>

Over a microscope, sure, the ND filter will block all light, and you won't see a thing; but this is not about

that at all. This is about relative brightness difference, about exceeding the dynamic range of your eyes.

<p>

Obviously you didn't get my original point. So now you have to do any two experiments to compensate - (I

mentioned four - one to you in the beginning, two to Daniel and one in this post). OK, I'm kidding.

[bernardwest]

<i>Over a microscope, sure, the ND filter will block all light, and you won't see a thing; but this is not about that at all. This is about relative brightness difference, about exceeding the dynamic range of your eyes.</i><p>

 

Ok, so what happens with a 60% blocking ND filter? White? What about a 30% blocking filter? Still white? 15%? etc. etc.... So at what point does the light magically change from white to black? Is this a gradual thing or does it happen instantaneously?

[bernardwest]

By the way, it is my lunch break... that's why I am (temporarily) back...

[vijay_nebhrajani]

<i>So at what point does the light magically change from white to black? Is this a gradual thing or does it

happen instantaneously?</i>

<p>

At the point the light has decreased to below the threshold where it excites your rods and cones. It is a

logarithmic response.

<p>

That was also not the point. The point is not about how our eyes see. It is about how our brain processes the

information.

<p>

When the brain processes optical data from the eyes it is relying on two things: absolute intensity of light, and

relative brightness difference; actually the absolute signal response from the retina, and relative signal

response differences.

<p>

Our brain evolved in a way (to maximize the efficiency of the visual recognition process as a survival tool) such

that it tries to maximize the number of different brightness levels visible at the same time. In other words, if

you have a range of brightnesses in a view, it will automatically place the brightest thing at "white" (yes,

exactly like autolevels in photoshop) and try to extend the dynamic range (saccadic motion and all) to the

darkest object.

<p>

This is why in all the experiments I suggested, the brightest thing - computer monitor, or the sun, or the light

table - appear white.

<p>

If the white part is bright enough compared to the dark parts (such as the letters on the screen, or midtone

negative pieces on a light table) the dynamic range is exceeded (the pupils contract enough because of the bright

light) and they appear black, completely opaque, impervious to light.

<p>

When you stop down by one stop in the experiment I suggested to Daniel, you cut down the intensity of the bright

part by 50%. Two stops = 75%. 3 stops = 87.5%.

<p>

Until about 3.5 or so stops you won't see any difference - after that, when the light becomes so dim that the overall

retinal response to light is low enough (some arbitrary threshold in our brain that probably evolved to recognize

night or something) we don't any more perceive white, and our brain recognizes gray.

<p>

Merely this understanding should tell you how fallacy ridden the view through a bright field microscope (bright

light shining at one eye in a dark tube - other eye closed) can be.

[vijay_nebhrajani]

Welcome back, Bernie. You think I don't want you here? Actually I do very much - your questions and arguments

will help shape our overall understanding. But I don't want to be the cause of your wife leaving you, man.

[bernardwest]

<i>Ok, so what happens with a 60% blocking ND filter? White? What about a 30% blocking filter? Still white? 15%?

etc. etc....</i><p>

 

What I meant to say was a <b>transmitting</b> filter, not a blocking filter.

[vijay_nebhrajani]

What I meant to say was a transmitting filter, not a blocking filter.

 

That's OK - 60% blocking is 40% transmitting - so the point was clear. I already explained it, though.

[bernardwest]

<i>if you have a range of brightnesses in a view, it will automatically place the brightest thing at "white" (yes, exactly like autolevels in photoshop)</i><p>

 

So when you go outside at night, everything is really white is it? Not in my backyard it isn't. Too extreme of an example (i.e. reduced too close to absurdity)? What about the storm example I gave. The sky certainly wasn't a blown white when I was driving around? Let's hear it then...

[vijay_nebhrajani]

OK, fair enough. When we have a frame of reference, things work differently. This is especially true with the

outdoors; our brains have evolved to recognize day, night, stormy conditions, sunrises etcetera. When outdoors,

we have a memory based frame of reference about what "bright sunlight" looks like; our brains recognize that this

isn't bright sunlight, hence they auto adjust to make things gray, including the sky.

<p>

In those very same stormy circumstances if you had tried to read a book, you'd still have seen black letters on

white paper, but this time because the brain "remembered" that paper is white. Tricky brains. (If you are

interested in this sort of thing, I recommend a book called "<a

href="http://www.amazon.com/Trick-Magical-Illusions-Activate-Brain/dp/0760766983/ref=sr_1_1?ie=UTF8&s=books&qid=1226988236&sr=1-1">Trick

Eyes</a>".

<p>

You have to create a circumstance where the frame of reference is missing (hence the darkened room for looking at

the monitor or the light table in my experiment) in order to replicate the view through a microscope, <i>where

there is no frame of reference</i>.

[vijay_nebhrajani]

Oh, by the way, this is the same issue at play when I invoked heliocentricity or that the moon is grey but appears white.

 

In the first case, there is no frame of reference so the sun appears to revolve around the earth, and in the second, the brain has evolved to place night at "black" (autolevels again); so the moon, which is actually quite gray, appears white.

[vijay_nebhrajani]

@ Stephen Pitts , Nov 17, 2008; 01:32 p.m.

<p><i>

Been there on the internet ($$$) loss. Now experiencing the banking loss... Its like being in Dark City.</i>

<p>

I'm glad someone feels my pain about that. Man, everything these days is just bad economic news.

<p>

<i>I think your overall estimate of 35 mp for 35mm is probably right on the money.</i>

<p>

Thank you. At that point, diffraction, manufacturing tolerances, film plane flatness, focusing errors, dispersion

through gelatin... so many factors come into play - that there is almost no practical resolution available beyond

that for 35mm. Digital sidesteps some things, like film plane flatness, or gelatin dispersion, but can't sidestep

the fact that diffraction limits lens resolution. So it is 35 MP in either case - digital or film, as an upper

limit. In practice, I think that at 25 MP or so, both are equivalent in terms of resolution, so comparison has to

be based on other factors, like aliasing, Moire patterns, grain etc.

<p>

Long ago in this thread, my point was that digital had actually arrived - technically that is. So choose based on

personal preference. I had expressed a lot of joy at this, since I could now go get that EOS 5D mkII without

feeling that I could be missing on technical quality.

<p>

And by the way, I usually don't shoot anything smaller than medium format. I'm primarily large format.

<p>

<i>Of course, there are lots of other factors in the film versus digital comparison. But despite the length of

the piece, the stongly worded arguments on each side, this has been the most thoughtful and (well) argued debate

on the subject. </i>

<p>

I'll take a small portion of that compliment - of course, everyone deserves praise for civility in this thread.

Amazing, huh? Over 600 posts and no name calling, no abuse hurling!

<p>

<i>Including, of course, Armando's comment of "just look at it to compare the two" (in the world of audio, the

ear is a much more particular and perceptive instrument than lab bench equipment and so an analytical debate of

this nature is almost pointless -- you can only judge by listening).</i>

<p>

I design high fidelity audio amplifiers for a hobby - Class B, complementary push pull, very low THD circuits. If

you've read Doug Self - blameless amplifiers. I know how debates about audio can get. Those audio nuts are freaks!

<p>

<i>Your analysis bears much fruit and one would hope that this thread would be cited to simply establish the

rough level of equivalency in information content between the two. By the way, the 35 mp nicely matches to the

diffraction limit of the very, very best lenses available on the market (you would need to build huge lenses to

get any better). So the different worlds all begin to converge on some common figures and everything begins to

make sense. Anyhow, thanks so much for providing so much to the thread (but please calm down a bit) -- we'd hate

for you to be so frustrated as to give it up. A truly great contribution to this most relevant and important

subject.

</i>

<p>

Many thanks for your compliments.

[vijay_nebhrajani]

Help! Can't turn off italics </i> Help!

[vijay_nebhrajani]

Drat, where's that Rich HTML guy when you need him? </i></i>

[bernardwest]

Ok, good, we're getting a good discussion here based on logic that both of us agree on. So, let's keep going (although, I'll have to return to the other part of my life soon, so hopefully, if I'm absent, DAniel can take up the cause).

 

What happens when we look at a slide (by this I mean a microscope glass slide) that contains some biological thing that is grey? Let's say that the grey stuff takes up about half the field of view. So, we've got a slide that looks half grey and half white. Hang, on.... according to your 'Auto-Levels Brain Hypothesis' (ABLS), we should only see black and white. But, of course, as anyone who has looked through a microscope knows, this is nonsense. We can definately view grey tones. And don't go claiming that the surrounding black around the field of view is the Auto-Levels black. When I last used a microscope, this surround, even if it was ever visible, was so far to the extremes of the field of vision, that I can't accept that it could contribute to your ABLS. But just incase it does, I'll present you with another scenario:

 

Let's say we have said grey stuff on 50% of a glass slide, such that it occupies fully the left half of the slide, and it blocks 50% of the light. So we have 50% grey and 50% blown white (on the right side). What happens when we slip an 10% attenuating ND filter between the blown white side and the view piece? According to your theory it should remain white. What if we slip a 20% ND filter between? The same white, according to you. Now, what if we slip a 49% attenuating ND filter over the white half of the slide? Still white? Yes, says Vijay. Now what if we swap the 49% for a 51% blocking filter? According to your theory, the image in the viewpiece should instantly reverse! We should now have a white left half and a grey right half. That, Vijay, is something I believe you are supposedly familiar with. It's called <i>reductio ad absurdum</i>. But wait, let's take it one more step further. What happens when we put a 50% filter over the right half? What do we see? Is the whole image white, or is it black? Can it be both?

[vijay_nebhrajani]

OK, if you have something that appears like a gray midtone, it is actually blocking between 87% and 83% of the light. (Remember 13% - 18% middle gray?) Not 50%, but over 80%.

<p>

You can easily replicate a situation like this by taking a negative that is slightly darker than mid gray (uniformly so) and place it on a light table. Then view it with a loupe so that the negative occupies half of the area. Try it - you just need a negative that is a little darker than a midtone on a light table. Look at it through a loupe and let me know. (We need slightly darker than mid gray because the light table may not be bright enough for the contrast to exceed the dynamic range of your eyes.)

<p>

Fair enough so far?

<p>

Your 51% and 49% example does not work simply because our eyes work with <i>brightness differences.</i> The difference between 49% and 51% is simply not enough for our eyes to decide that one is black and the other is white. Similarly, if you think that if you put something that is 95% transmissive, it will also appear black, that is not true - you have to exceed the dynamic range of your eyes; a 5% difference is well within your eyes' dynamic range. Dynamic range is exceeded when there is sufficient difference between the dark thing and the bright thing.

<p>

But here's an interesting experiment. Why don't you actually try it? Go back to that light table - where half the viewing area was covered by a middle gray (or slightly darker) negative? Slide the negative so that the loupe covers the entire negate and Voila! - it'll appear dark gray. This is because there is now no contrast information for the brain so it is using the amplitude of the signal from the retina, comparing it to information in your brains memory banks and coming up with dark gray. (That signal amplitude probably corresponds with late evening or something, from your brains evolutionary data bank.)

<p>

Don't believe me? Just try it.

[owen_laprath]

WOW! Looooong thread ... getting back to the original question:

 

For Pro use, you may have an edge, as there are still a few stickler editors for film scans for magazine prints.

 

You certainly have any dSLR beat in resolution for daylight film.

 

For low light shooting though, digital has an edge now with the "full frame" models by the evil Caniksony triumvirate of industry bullies - at 3000 bucks a piece and more :)

 

Serious though - an entry level dSLR may be a good investment for a pro, even if you do film - a sample shot on the dSLR will let you know what you can expect from your shot, and then the Pentax 67 goes in for the Medium Format "kill" to get the ultimate in high resolution and dynamic range on film.

 

By the time the system including the el cheapo sub USD 500 dSLR pays for itself, prices will have come down again, and you can make the next step in your business investments with a "full frame" dSLR - but hang on to that film outfit, you never know when you may need it - also get a stash of film and a mini lab :)

 

There is one more dimension:

 

Environment

Much is written about our voracious chip appetite and wasted high tech equipment, but even if we buy a new digicam every 3 years (don't skimp, and it will last at least that long!), for those of us who would shoot a lot either way, there is still less waste in the ground, water, and air, than shooting ... a few hundred or a couple of thousand rolls of film maybe?

 

Owen

[bernardwest]

<i>Your 51% and 49% example does not work simply because our eyes work with brightness differences. The difference between 49% and 51% is simply not enough for our eyes to decide that one is black and the other is white.</i><p>

 

Do you have any literature to back this up, or are you just pulling this stuff off the top of your head? It doesn't matter anyway, as I agree that in the case of closely related grey tones, you are not going to see black or white. By the way, who said anything about middle grey? I don't care what tone it is, all that matters is the relationship between the two examples.<p>

 

So clearly, you are invoking limited brain/eye dynamic range, to hinge your argument on. But Daniel has already debunked this line of reasoning. Have you ever even looked through a microscope? You do understand it is perfectly reasonable and common to observe grey tones under a microscope. Otherwise, what would be the point. If the light was so bright that it made everything appear either black or white, then there would be no use for microscopes. But the light isn't like that, and there are even filters and/or apeture blades on some which can reduce the intensity of the light. Regradless of all that, limited dynamic range is a furphy anyway, because if it was limited when light was projected through the negative, then it will be limited in the print as well.

[vijay_nebhrajani]

<i>Do you have any literature to back this up, or are you just pulling this stuff off the top of your head?</i>

<p>

Bernie, I referenced a book, you could read it; you could use google to come up with literature, you could go to

your local library.

<p>

But I urge you to do the experiments I suggested. You can then "see" for yourself.

<p>

Or why don't you just google for "optical illusion" - there are plenty of images; parallel lines that don't look

parallel, static images of disks that appear to rotate - endless fun.

<p>

The point is that the human eye/human brain is a fallible instrument, and it could be causing some erroneous data

interpretation; to base a theory on that would make the theory fallible as well.

<p>

And you are quite right, dynamic range will be limited on the print as well. But here is the trick question, pay

close attention: <b>If you were to enlarge by 400x a section of a negative that contained a middle gray tone onto

paper, would you get black grains on a white background, or would you get black grains on a gray background?</b>

<p>

Think carefully, and use your imagination a bit before answering.

[vijay_nebhrajani]

And oh yes, absolutely, it is common to observe gray tones, colors, and all sorts of optical deliciousness under a microscope. No question, and I agree.

[danielleetaylor]

<P>Vijay,</P>

<P></P>

<P><i>To talk about film - to get a density of those tiny silver specks that you could observe through a 400x

microscope, you'd have to expose the film enough - but this "enough" would also create larger silver specks,

which would make the smaller ones harder to see. That does not prove that the smaller specks don't exist, or that

they wouldn't attenuate light also, playing their part in the generation of tone.</i></P>

<P></P>

<P>If a small speck is directly over or under a large speck then it means nothing. Silver is opaque, it cannot be

made "more opaque" in this context.</P>

<P></P>

<P>If a small speck exists in the space between larger specks, and is large enough to be observable, then it is a

part of the halftone pattern. It itself is not "gray", though its existence helps produce the perception of gray

at magnifications too low to resolve individual grains.</P>

<P></P>

<P>If a small speck exists in the space between larger specks, and is too small to be observable at 400x, then it

cannot be observable at 10x. It cannot influence perception. It's meaningless. If it were otherwise, then B&W

film would be useless due to all the spots and obstructions caused by sub-microscope scale crude both on the

film, on the paper, and in the air in between.</P>

<P></P>

<P><i>Try a couple of simple experiments.</i></P>

<P></P>

<P>Your experiments have nothing to do with what we're discussing. No offense, but I am so freaking sick of

hypotheticals and "experiments"! Where are these gray grains? If they are not observed, then film tone is not

created by them. Science doesn't get much easier than that!</P>

<P></P>

<P>Here's an experiment I want you to perform: observe real B&W film under a microscope. Tell me when you find a

gray grain.</P>

<P></P>

<P><i>The point is that the human eye/human brain is a fallible instrument, and it could be causing some

erroneous data interpretation; to base a theory on that would make the theory fallible as well.</i></P>

<P></P>

<P>So now you're suggesting that human vision sees only opaque grains when necessary to fool us into a false

theory of B&W film, but sees gray grains when we're not thinking about B&W film, but only enjoying prints? Do I

have this right? Is Descartes Demon preventing us from understanding the true nature of B&W film? <i>I post to

photo.net, therefore I am.</i></P>

<P></P>

<P>I could have created a unified field theory in less time than it is taking to establish that gray grains

aren't observed at all, much less at a frequency necessary to contribute to gray tones in film. Adams was right,

B&W film is a halftone process. And Reichmann was right, grains are not pixels.</P>

<P></P>

<P>And the economy blows, but that's a completely different topic....</P>

<P></P>

[danielleetaylor]

Vijay, can we apply Occam's razor to all of this? Right now you are jumping through hoops trying to explain gray grains that influence tonality at 10x but are not observable in any other manner. It's just not the explanation that naturally flows from the evidence. If we took someone who had no interest in photography and showed them a magnified view of a halftone pattern and then a magnified view of B&W film and asked them what the film looked like, they would say "Hey, it looks kind of like that halftone pattern thing you just showed me!" Pretty much...

 

Maybe we should give up looking for elusive (yet some how significantly contributing) gray grains? :-)

[bernardwest]

<I>Vijay, can we apply Occam's razor to all of this?</i><p>

 

I agree wholeheartedly. <P>

This is becoming more farcical with each post by Vijay. Back to real life for me for a while. See you all next lunch

break.

[vijay_nebhrajani]

<i>Daniel: "If a small speck exists in the space between larger specks, and is large enough to be observable,

then it is a part of the halftone pattern. It itself is not "gray", though its existence helps produce the

perception of gray at magnifications too low to resolve individual grains."</i>

<p>

OK, we're getting closer.

<p>

Lets talk about that small speck between the larger specks shall we?

<p>

Lets see what happens if that speck is out of focus through that microscope, shall we?. Its edges would be fuzzy,

light would be diffracting around it, and if sufficiently out of focus you'd get a uniform gray tone in its

place. Correct? Easily observable even in Kodak's photos that you posted.

<p>

That is not how a halftone process works, right? In halftone processes, there are no gray areas between black

areas, are there? Its all black or all white.

<p>

Now, you can have a gray area between opaque grains because of diffraction from out of focus grains, correct?

What is the mechanism of the diffraction? To spread light out, to attenuate its local intensity, correct?

<p>

What would happen if you had a silver speck that was really tiny, too tiny to see through your optical

microscope, because the light from the bright field diffracted around it? It would still be there, attenuating

the intensity of the light falling on it by a tiny amount, right, both by diffraction and by absorption? Helping

form the "gray" between the black, correct?

<p>

If that is not the formation of tone, what is?

<p>

Now what if that speck were thin enough and small enough? Light would diffract around it; perhaps some light

would pass right through the silver itself.

<p>

What would that do? Help formation of tone. Tone is nothing but attenuation of light. A big speck of silver

attenuates light in big chunks, smaller specks in smaller chunks. The total amount of light absorbed by the

silver determines tone on the final print, right?

<p>

You can't be saying that all silver specks that are too small to be seen under an optical microscope can't be

engaging in attenuating light can you?

<p>

Even if they attenuate by a tiny amount, they do so. Perhaps you are thinking that when I say "gray specks of

silver" I mean specks that are as large as black ones, but partially transmissive - like 5 microns x 5 microns,

but 50% transmissive. Perhaps something I said seemed to imply that. That is not what I meant - I've clearly said

that there would be specks at different stages of growth (see Rich's SEM micrographs, on the left there is one

tiny speck about 100-150 nm in the 102kx magnification image). Are you implying that that speck, sitting there,

is contributing nothing at all to tone in the final print?

<p>

<i>No offense, but I am so freaking sick of hypotheticals and "experiments"!</i>

<p>

I'm sorry, its not like I'm thrilled either. But for the sake of all the time and energy already spent, we owe it

to ourselves to come to some common ground.

<p>

And its not like the Adams/Reichmann theory can explain everything either. For instance, it can't explain how

film resolves 80 cpmm when the target has a contrast ratio of 1.6:1; something that would be impossible because

it would take too large a film area to form tones that close - meaning that you should only be seeing a random

jumble of silver specks rather than finely differentiated tones at such high cpmm.

<p>

<b>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. </b>

<p>

Reichmann's theory predicts that increasing number of tones requires an increase in the area of film; for

instance if you represent tones with a 16-bit number, then creating a tone was really dark gray, like a ratio of

65,535/65,536 would inordinately large area.

<p>

Reichmann's theory predicts that the resolution of silver halide based film will be much less than chromogenic

film - which is dye based, and whose dye clouds "pixels" can represent many tones.

<p>

Daniel, please come up with an explanation for these two cases. Ignore all my talk of gray grains and whatnot.

Ignore everything, and just tell me why these predictions of Reichmann's theory don't lead to valid observations

in the real world.

<p>

Just tell me why silver halide film resolves 80 cpmm (use this number, you have to show that this is possible,

because this is observed) at contrast 1.6:1 and why it outresolves chromogenic films, including color films,

again at 1.6:1. Reichmann explains why it would resolve a high resolution (1000:1) target by saying it takes just

one pixel to record black. Fine, why is the resolution so high at 1.6:1? And why, when we make the "pixels" such

that they can take on any tone as in chromogenic or color film do they still not outresolve silver halide film?

<p>

You have, so far only negated my point of view; you have not enhanced Reichmann's or Adams' one bit by providing

appropriate explanations. In other words, if I go through all your posts, I don't come away with an iota more of

understanding how film works, how tones are formed or how film resolution works than if I read a hundred lines of

a pithy Reichmann article.

<p>

I have done my bit of this by providing a theory, some numbers etc. Now it is your turn. Please explain the

difficulties I have pointed out with the Reichmann/Adams theory as a first step.