Luminous Landscape Forum
The Art of Photography => The Coffee Corner => Topic started by: Diego Pigozzo on October 08, 2015, 09:12:50 am
-
Given what (very very little) I know about the development chemical process, the answer should be "yes".
But I'm probably missing a lot, so I'm asking.
-
If I understand the question properly, the answer would be yes. You simply lay the (same) image over a larger area and benefit from less visual interference from the structure of the medium and the development processes themselves.
Whether this is always a good thing isn't as clear: many times the use of a smaller film gives you effects that the use of larger formats can't. And, obviously, vice versa.
Rob C
-
If I understand the question properly, the answer would be yes. You simply lay the (same) image over a larger area and benefit from less visual interference from the structure of the medium and the development processes themselves.
Whether this is always a good thing isn't as clear: many times the use of a smaller film gives you effects that the use of larger formats can't. And, obviously, vice versa.
Rob C
I was thinking more about the fact that a large format film has "more samples" of the same area (which has a relation with the visual interference you're talking about, but I'm not sure they are the same thing).
-
I think you are both saying the same thing, non?
-
I think you are both saying the same thing, non?
Maybe, but I'm not sure.
As I understand it, what RobC said is something akin to "for a given ppi, a bigger monitor will show more image details than a smaller one (assuming the image has enough resolution)", which is something more about spatial resolution than "tone resolution".
My doubt was more about "tone resolution", that is the ability of large format film to resolve more tones (not just more details).
As I understand it, what the development process does is to "propagate" the chemical changes from the exposed molecules to the unexposed ones.
And as I understand it, this process is necessary because very few light-sensitive molecules get struck by photons.
If all the above is correct, a film can be seen as a set of binary samples (where "binary" means "just two states") of an image: each light-sensitive molecules is either exposed or not esposed.
So I'm supposing that a large format film has more samples of an image, and therefore can resolve both more details and more tones.
I tried the following experiment:
1) I took and 2048x3029 JPG (original_image.jpg)
2) I converted it to a 1 bit image (that is, pixel can only be black or white, no grays) (just_1bit_convertion.gif)
3) I reloaded the original image, resized at 25% and converted to a 1 bit image (resize25_and_1bit_conversion.gif)
As you can see, the just_1bit_convertion.gif has (or looks to have) both more details and more tones than the resize25_and_1bit_conversion.gif.
I don't know if this test is meaningful in any ways, but that's the best examples I came up with.
By the way, both what I'm saying and what RobC is saying are strongly correlated.
-
But you can't think of film really as digital capture.
A closer analogy of what I think you mean now, might be taking the concept of enlarging.
From whatever format of negative you begin, the greater you enlarge, the film's 'pixels' are being spaced over a larger and larger print area with no interpolation so the image falls apart, but in an honest way. ;-)
See the movie Blow Up!
Rob C
-
But you can't think of film really as digital capture.
That's why I used "binary" instead of "digital": because binary implies only two values (which are, in the case of silver halide, "exposed" or "not exposed").
A closer analogy of what I think you mean now, might be taking the concept of enlarging.
From whatever format of negative you begin, the greater you enlarge, the film's 'pixels' are being spaced over a larger and larger print area with no interpolation so the image falls apart, but in an honest way. ;-)
See the movie Blow Up!
Rob C
I agree, but what I mean was (more or less) this: "If you print a large format film to a certain size, do this print show more tones than a 35mm frame printed to the same size, when viewed at the same distance?"
-
Large format is basically Dolby noise reduction for pictures.
By making the signal bigger and keeping the noise levels the same you make the whole thing look better.
Ideas from sampling theory map to analog quite poorly. The correct model is signal to noise ratio, usually.
-
"As I understand it, what the development process does is to "propagate" the chemical changes from the exposed molecules to the unexposed ones.
And as I understand it, this process is necessary because very few light-sensitive molecules get struck by photons.
If all the above is correct, a film can be seen as a set of binary samples (where "binary" means "just two states") of an image: each light-sensitive molecules is either exposed or not esposed."
No, it's not like that. It is not a matter of exposed or not exposed, yes or no; it's a matter of degree (as in strength) of light hitting the individual molecule.
The characteristic curves show you that highlights are better handled by film than digital because they 'roll off' better and allow a smoother transition from max. white into the greys.
And because film is an organic concept, even films of the same nominal ASA/ISO rating have different physical properties: Kodak's Tri-X was a film that tended to be more 'crisp' in the sense of having tighter, more distinct-looking grain, whereas Ilford's HP3/4 showed a characteristic rolling, slightly more mushy, look in its grain structure if you made similar sized blow-ups from the same neg formats.
For what it's worth, comparisons between the two mediums are best forgotten because the differences are too great to make sense. Or so it has come to seem to me.
Rob C
-
You'll also need to consider resolution and contrast characteristics for both film and lens along with film grain size and how many tonal distinctions the lens can expose to silver halide crystals compared to how much is just grain & noise. It's the same issue with digital sensors.
When I was a staff photographer in high school for both the annual and newspaper, I noticed my instructor's 120 B&W film captures off an old Rolleiflex camera was so sharp and detailed compared to the 35mm B&W captures off the school issued SLR with 50mm prime lens. At the time I couldn't tell if it was the lens or the film stock.
About 15 years later I got my own Yashica 35mm SLR/50mm prime lens and shot on Agfa color negative print film where I was surprised by its sharpness and detail on the 4x6 print, but the lab cranked up the contrast and saturation to where I couldn't be sure of causality until I finally scanned the 35mm negatives myself 20 years later when I bought my Epson scanner where I was getting tonality that rivaled the 120 B&W but it was of close-up head/shoulder portraits.
So I guess it depends on a lot of things that would require controlled tests using a fixed target at a fixed distance and all the other stuff I mentioned above.
-
No, it's not like that. It is not a matter of exposed or not exposed, yes or no; it's a matter of degree (as in strength) of light hitting the individual molecule.
I'm not sure I get this: I thought that a silver halide molecule could only be on the "exposed" or "unexposed" state.
Am I wrong?
-
...
I noticed my instructor's 120 B&W film captures off an old Rolleiflex camera was so sharp and detailed ...
So I guess it depends on a lot of things that would require controlled tests using a fixed target at a fixed distance and all the other stuff I mentioned above.
The same thing happened to me with my Flexaret VII: the shots looks much more sharp than both the 35mm and the 16mpx DSLR I own.
So, probably, mine was just a masturbaccademic question ;D
-
I'm not sure I get this: I thought that a silver halide molecule could only be on the "exposed" or "unexposed" state.
Am I wrong?
I'm sure that you are under a delusion based upon, I suspect, the fact that you are making comparisons of 1 and 0. It is flexible, or everything would be either totally black or totally white, just as you'd get using a line film. You'd get no shadow detail - just blackness. The shadow exists (in the image) because of the lesser degree of light reaching the silver halides, but still being just strong enough to have some slight photochemical effect on the film.
But it gets worse: different emulsions are also sensitive to colours in different ways, and that makes huge differences too. Panchromatic films manage to record all colours of the spectrum that are visible to us, and orthochromatic emulsions do much the same except for red and orange. Basically, that's why coloured filters are used in black/white photography and can make huge differences - as St Ansel, famous fan of red, would tell you!
I hope this helps, but it really demands that you forget totally about digital capture and clear your mind before you get into the mental process of figuring it out.
I could be totally wrong: I only remember what I was told in the night school classes I was obliged to attend for a while. I abandoned them as soon as I realised that the place - well, the tutors - had nothing to offer me for where I was wanting to go. Turned out I was right about the latter bit anyway! Regarding the rest, photographic materials theory, it interested me very little: I knew what I had to know in order to be able to shoot and process and print well, but apart from that, it held no interest. I felt I was a photographer not a scientist. Which, really, is why I find digital hand-wringing and pixel peeping etc. coldly uninteresting. For me, it's not about the technicalities, it's about the art.
;-)
Rob C
-
I'm sure that you are under a delusion based upon, I suspect, the fact that you are making comparisons of 1 and 0. It is flexible, or everything would be either totally black or totally white, just as you'd get using a line film. You'd get no shadow detail - just blackness. The shadow exists (in the image) because of the lesser degree of light reaching the silver halides, but still being just strong enough to have some slight photochemical effect on the film.
I'm still don't getting it, sorry.
I mean, in the hypotesis that a silver halide molecule could only be "exposed" or "non exposed" doesn't implies that the resulting image will have only two tones: given that the molecules are so small and so much, our eyes would perceive them as shadows.
(more or less like in the images I posted).
I looked on Wikipedia (https://en.wikipedia.org/wiki/Silver_halide) and they say:
When a silver halide crystal is exposed to light, a sensitivity speck on the surface of the crystal is turned into a small speck of metallic silver (these comprise the invisible or latent image). If the speck of silver contains approximately four or more atoms, it is rendered developable - meaning that it can undergo development which turns the entire crystal into metallic silver. Areas of the emulsion receiving larger amounts of light (reflected from a subject being photographed, for example) undergo the greatest development and therefore results in the highest optical density.
I understand this to means that there are no "degrees" of exposure: just "exposed" or "non exposed".
(truth is, they are talking about "render developable", but the result is the same).
Let me use your examples of characteristic curves: they looks like continous curves because the numbers of silver halide molecules is huge,
But let's imaging to have a tiny tiny piece of film with just 10 molecules: the characteristic curve will still look continuous or it will looks to have discrete steps?
-
i agree with Diego: under a microscope a silver halide negative is all black and white, with no intermediate shades of gray: at any locations, either there is developed silver (black) or there is not (white). The same with silver halide prints as far as I know. What we see is the amount of light reflected from a region with hundreds of solver specks, because our eye cannot resolve down to the level of individual specks, and the shade of gray that we see depends on the fraction of each tiny part of the negative/print that is black. The same with half-tone printing: look at a billboard close up! The term is "dithering", I believe.
Aside: this fact that a silver-halide negative is "binary", so with a very low "pixel level dynamic range" if viewed close enough, is why people should stop obsessing about per pixel dynamic range specs, and instead look at the DR and tonal gradations seen when displayed and viewed at a particular size.
-
Those of you pushing the "binary" nature of silver halide crystals seem to be forgetting or ignoring the fact that exposure isn't the only thing that matters. If, indeed, every molecule were "on" or "off" depending on whether it was struck by light or not, then development would make no difference at all.
In general, longer development creates higher contrast. This is the basis of Ansel Adams' famous "Zone System."
And, in my fifty years experience of darkroom work with films from 35mm to 8x10" I saw time and time again that a larger negative was able to display more visibly different shades of grey than a smaller negative, and would have a significantly greater dynamic range as well.
Eric
-
Crystals and molecules are different things and the ugly truth is that the photochemistry of silver halides remains somewhat of a mystery.
We have some very useful models and a lot of understanding but there remain some corners where we just don't know what's going on, I am given to understand.
-
To which, I suppose, one could add the phenomenon whereby even very well-known PJs subscribe(d) to the theory that increasing development increased film speed. It didn't. What it did do was increase contrast within the faint shadow areas and give the impression of higher film speed because those shadow areas were made slightly more distinct. But that was just an increase in local contrst, not in detail captured. If anything, the greater achievement was really a loss of highlight detail due to clogging up through over-development. But the myth lived on, and even became a style.
Such is photography.
Rob C
-
Those of you pushing the "binary" nature of silver halide crystals seem to be forgetting or ignoring the fact that exposure isn't the only thing that matters. If, indeed, every molecule were "on" or "off" depending on whether it was struck by light or not, then development would make no difference at all.
In general, longer development creates higher contrast. This is the basis of Ansel Adams' famous "Zone System."
Agreed, but that does not change the fact that the information on a developed "silver" negative, is binary at the microscopic scale: those tiny bits of silver are black, not shades of gray. Changes in development only change the density of the silver particles, so that there are more regions of very high and very low density and less region of intermediate density. The bottom line is still that measuring dynamic range on the original "raw recorded signal" is meaningless until you account for how much local averaging (dithering) goes into determining the gradations of luminosity (and color) that are detected by a viewer's eyes.
And to the original question: printing from a larger negative, and so at lower enlargement for the same print size, allows more of this local averaging, and thus can give finer tonal gradations. Diffraction from the enlarging lens alone does some of this averaging.
-
Those of you pushing the "binary" nature of silver halide crystals seem to be forgetting or ignoring the fact that exposure isn't the only thing that matters. If, indeed, every molecule were "on" or "off" depending on whether it was struck by light or not, then development would make no difference at all.
In general, longer development creates higher contrast. This is the basis of Ansel Adams' famous "Zone System."
And, in my fifty years experience of darkroom work with films from 35mm to 8x10" I saw time and time again that a larger negative was able to display more visibly different shades of grey than a smaller negative, and would have a significantly greater dynamic range as well.
Eric
I can't talk for the others, but I'm not "forgetting or ignoring the fact that exposure isn't the only thing that matters": I'm just simply toying with the idea that large format film may have more dynamic range than smaller ones (which is something you're confirming to be true).
About the uselessness of development if a silver halide molecule/crystal were "binary", I'm not sure it really matters.
For sure silver halide molecule/crystal have a finite number of stable states and I think this number is "two".
But even if this number was not "two", it must be very low: I don't know of any molecule/crystal that has hundreds of stable states.
About the Zone System, thinking about it seems to confirm my idea that development "propagates the exposed state" from the exposed molecule/crystal to the unexposed ones.
If I'm not mistaken, the Zone System says to expose for the shadows and to develop for the highlight because the development process has a stronger effects on highlights.
Now, if the development process just "do something to the already exposed molecules/crystals", you could get an increase of constrast only if the silver halide molecule/crystal has a huge number of stable states (so that you're making the black "blacker" and the whites "whiter".
But if the development process "propagates the exposed state", it's easy to see why you get more constrast: if the process doubles the exposed molecules/crystals (by propagation) then the part of the film that starts with 1N exposed molecules/crystals will end up with 2N, while the part of the film that starts with 2N ends up with 4N.
This means that if the starting difference in contrast was 1N (2N - 1N), the end contrast will be 2N (4N - 2N).
-
Film format does not have anything to do with dynamic range. You can cut the same emulsion up large, small, or in to paper dolls, the properties remain the same.
Does a digital camera lose dynamic range when you set it to crop mode, only using the middle of the sensor?
-
Does a digital camera lose dynamic range when you set it to crop mode, only using the middle of the sensor?
The pixel doesn't lose dynamic range but the over all image does, comparing shots of the same scene which have equal FoV, and using the same ISO and same ETTR process.
When the full sensor is used, each part of the scene examined will consist of more pixels than the same parts of the same scene shot in crop mode. More pixels means more light has been gathered to portray each part of the scene, whether shadows or mid-tones, and the over all signal-to-noise ratio is therefore better. This is one of the main reasons why some people prefer the larger format.
-
Film format does not have anything to do with dynamic range. You can cut the same emulsion up large, small, or in to paper dolls, the properties remain the same.
Does a digital camera lose dynamic range when you set it to crop mode, only using the middle of the sensor?
The tonal resolution of a print does depend on the format indirectly, through the different degrees of enlargement needed to get a given sized print: please consider what has been said about dithering, and the way that information gets "averaged" to form the signal that the eye eventually receives. Averaging multiple signals increases the SNR of the resulting signal.
And to your second question: Yes, in the important practical sense of signal to noise ratios of the final displayed image (print or on-screen), working with the whole file rather than a crop does allow you to produce a final image with "higher DR" or "finer tonal gradations" because for example the data from more pixels can be averaged to produce output pixels of higher SNR (higher DR).
This should be easy to see: take a high pixel count file (preferably taken at high ISO, so that the raw per-photosite DR is lowish) and first display the whole thing at suitably high PPI, and then view a small crop at the same size, so at far lower PPI. (Two prints of the same size, or just zooming in on screen will do it.) You should be able to see differences in "tonal gradations", or "DR" or whatever one calls it, in part through the more visible noise in the second image.
I plea once again for people to stop taking the engineering measures of per photosite DR and SNR alone as being direct measures of visual quality without taking account of how pixel count, PPI used when printing/displaying, degree of enlargement, and so on effect on the final visual result.
-
This situation is demonstrated by the DXOMark graphs, comparing full-format cameras which have the same pixel density and pixel quality as a preceding cropped format camera.
For example, the 36mp D800 is basically a full-format version of the 16mp D7000, and the Canon 5DS is a full-format version of the 18mp 7D mark II.
The DXO graphs show equal pixel performance for both brands when comparing full-frame cameras with the cropped format of the same brand. However, at same print size both the Canon and Nikon full-frame cameras show higher DR, higher SNR and better Tonal Range than the cropped-format equivalents. Check out the following DXOMark comparison at:
http://www.dxomark.com/Cameras/Compare/Side-by-side/Canon-EOS-7D-Mark-II-versus-Canon-EOS-5DS___977_1008
-
Fair enough. In pictures we tend to think of dynamic range as the distance from black to white, but yes, the noise floor does drop, which is an extension of DR.
I think dynamic range is the wrong thing to be looking at and thinking about.
The noise floor is the point here, and with large format film it's most obvious not at the ends of the range but in the middle. It is the relatively reduced noise that yields the sleek midtones, and if there's anything at all to a 'large format look' it is surely that.
-
I think dynamic range is the wrong thing to be looking at and thinking about.
I heartily agree! For most photographers, it would be good to return discussions to concepts like subject brightness range, tonal gradations, ISO film-speed like measures of how dark a region of the scene can be before the SNR is above a threshold of usability (not 1:1!) etc., rather than adopting new and often mis-understood measured from an electronics text-book. Used properly, various measure of SNR and DR can help, but only if we connect them properly to photographic objectives and effects on the final viewed image.
The noise floor is the point here, and with large format film it's most obvious not at the ends of the range but in the middle. It is the relatively reduced noise that yields the sleek midtones, and if there's anything at all to a 'large format look' it is surely that.
That makes sense.
-
For most photographers, it would be good to return discussions to concepts like subject brightness range, tonal gradations...
That's the subject of the thread: do large format films resolve more tonal gradations?
-
That's the subject of the thread: do large format films resolve more tonal gradations?
Subject to some fairly careful caveats about what, exactly, we mean, the answer is "generally yes".
-
"More" gradations?
Surely we are talking about Nyquist sampling: higher spatial sampling frequency relative to image size means the ability to record higher spatial frequency information.
However I have a feeling we are talking about "smoothness" of gradation, which is about re-construction... that means dithering essentially, and an appropriate bit of microscopic blurring added by the printing process... which may be the real reason some enlarging lenses have a magic reputation, they are sharp without being too sharp.
-
"More" gradations?
Surely we are talking about Nyquist sampling: higher spatial sampling frequency relative to image size means the ability to record higher spatial frequency information.
Yes, I think it is this (and maybe I just didn't get that "higher spatial frequency information" also implies "higher tone resolution").
However I have a feeling we are talking about "smoothness" of gradation, which is about re-construction... that means dithering essentially, and an appropriate bit of microscopic blurring added by the printing process... which may be the real reason some enlarging lenses have a magic reputation, they are sharp without being too sharp.
Yes, I think you're right.
I'm not an expert in signal processing, but I start to think that a given recostruction result may be achieved by both low-frequency/high-bit-sample and high-frequency/low-big-sample procedures.
Something like a 1-bit DAC (https://en.wikipedia.org/wiki/1-bit_DAC).
ADDED:
This looks spot-on (https://en.wikipedia.org/wiki/Oversampling#Resolution)
In practice, oversampling is implemented in order to achieve cheaper higher-resolution A/D and D/A conversion.[1] For instance, to implement a 24-bit converter, it is sufficient to use a 20-bit converter that can run at 256 times the target sampling rate. Combining 256 consecutive 20-bit samples can increase the signal-to-noise ratio at the voltage level by a factor of 16 (the square root of the number of samples averaged), adding 4 bits to the resolution and producing a single sample with 24-bit resolution.[3]
When oversampling by a factor of N, the dynamic range increases by log2(N) bits, because there are N times as many possible values for the sum.