Film vs Digital - Color Rendition

- We don't know that for certain, since we don't have access to how those pictures of a map, or maps were taken. Without visual sight of the original, everything you're stating as fact is just assumption and guesswork. And it's just one example that you're refusing to give us a link to the source of.

It's right here in this thread as I believe I mentioned earlier. Unless the guy who posted it for some odd reason fiddled with it to disproved his own point, I see no reason do disbelieve it.

Errors like that are typical of bayer demosaicing. They are often blatantly visible in certain details, but they profoundly affect image quality in the whole picture in subtle ways.

There are thousands more examples online that clearly show the superiority of detail and colour rendering of digital, for a comparable sensor/film area. Such as here, and posted by a film user, but not a totally biased one.

Did you even read the article/test you linked? It was very poorly and loosely executed, admitted by the authors themselves in the comments and they actually didn't conclude anything.

Edit: Sorry for the broken up post. Apparently I wrote something "inappropriate" or spam like that I wasn't able to purge, and I had to play divide and conquer.

- It doesn't matter about grain size. The point is that reduced silver is opaque, and that dye clouds are fairly uniform in size and density.

Opaque but different in size and the dye follows the silver.

What is seen in film is a dithering of opaque or coloured specks that only give the illusion of continuous tone.

You could say something similar of just about any imaging technology.

It's exactly analogous to the half-tone printing method, whereby various sized dots of a single-colour ink give the same illusion. But the colour depth or refinement is limited by the smallest sized dot that can be produced.

Not exactly analogous. There is no raster and the "primitives" are can be approximated in shaped to the feature they are conveying. And of course the "dots" is very much smaller.

I don't see your point with the comparison?

- Another example of unsupported nonsense. Have a look at DXO labs analysis of digital sensor colour depth.

Did you have a particular one in mind? Like this one: Nikon D850 : Measurements | DxOMark ?

There is still "only" 14 bits per colour channel. That is what the hardware can sample. As you can see the actual range is different.

No quantisation steps in film.

Opaque but different in size and the dye follows the silver

- No, it doesn't. The dye couplers are activated by developer oxidation products, caused by reducing AgX crystals in close proximity to a coupler globule. The two aren't physically linked except by proximity.

I don't see your point with the comparison?

- No, and you obviously don't want to.

There is still "only" 14 bits per colour channel

Well how many bits colour depth do you expect from a 14 bit A/D converter? That's 2^42 different shades of colour and greyscale, and would require 4.398 x 10^12 separate pixels to display them all!

"No quantisation steps in film."

- Yeah, keep saying that and it's bound to be true.

I suppose that's why trying to change saturation or tone curves in a film scan rapidly results in visible posterisation?

Still waiting for the link to that dubious map comparison.

- No, it doesn't. The dye couplers are activated by developer oxidation products, caused by reducing AgX crystals in close proximity to a coupler globule. The two aren't physically linked except by proximity.

That's it. Proximity and general volume. That is the factor that regulates the size of the dye clouds. And in turn their step-less size.

- No, and you obviously don't want to.

Look, we're not enemies, I'm not here to irritate you. If you have some real point to make I'd be curious to hear it.

The storage and displaying of the image captured by the analog sensor in a digital camera needs go through quantisation and locking into bit depth and a grid so in a sense if anything it's the digital image that is more akin to raster printing.

Silver halide imagings particular looks nearest cousin, would in my eyes be something like Floyd-Steinberg dithering

Well how many bits colour depth do you expect from a 14 bit A/D converter? That's 2^42 different shades of colour and greyscale, and would require 4.398 x 10^12 separate pixels to display them all!

It's not the number bits themselves I'm pointing to (though they could easily be a source of artefacts too) It's the actual range and the distribution of the range.

IE, where and how are the quantisation steps placed? That's also a factor to consider.

The DxO test shows 22.6 bits at ISO 400. So that's under 8 bits per colour channel. Even at ISO 32 it's barely more than 8 bits of actual discerning ability per channel on average.

"No quantisation steps in film."

 

- Yeah, keep saying that and it's bound to be true.

I suppose that's why trying to change saturation or tone curves in a film scan rapidly results in visible posterisation?

Because you are quantising the image when scanning it?

Real darkroom postersation relies the the upper and lower plateauing of film. In particular lith film and paper. That is still not a quantisation stepping though, since there is still a toe and shoulder, albeit a small one petering out into almost but not quite horizontal curves.

Still waiting for the link to that dubious map comparison.

Really? You couldn't be bothered to scroll though the thread? Then let me do it for you:

Film vs Digital - Color Rendition first post by danielleetaylor

If you have some real point to make I'd be curious to hear it.

- Like I said, you seem to be deaf to any opinion except what you've already made your mind up to.

So have you paid for the 5 minute argument, or the full half-hour?

The sensible thing to do would be to message Daniel Lee Taylor and see if he has any reasonable explanation for the missing lines. Maybe there was a speck on his camera sensor that he 'airbrushed' out? Because I can't believe it's due to your implausible interpretation of Bayer de-mosaicing skipping an obvious change of density. Otherwise there would be numerous reports of such missing detail, which there obviously aren't.

I couldn't find a duplicate of Daniel's map, but I came across a very similar one in a 1983 3rd edition of Philips' Concise Atlas of the World.

I measured the width of the blue lines of latitude and longitude. On average they're 0.2mm wide, with some small tolerance for ink spread.

I set a 16 Megapixel Sony Nex-6 camera up at a magnification (1:42) that would render those lines the same width on the sensor as the photosite spacing - 4.8 microns. A setup, which according to your - incorrect - theory of de-mosaicing should provoke loss of colour and definition in those lines.

As can be seen, no such thing happened.

To complete the methodology list: The lens used was a 50mm f/2 Nikkor at f/4, and the map was lit by an SB-28 speedlight flash off-camera. Camera was set to 100 ISO and flash white-balance.

The colour remains reasonably accurate to the map, as viewed in daylight, and no lines are missing. There is clear discrimination between lines that are black, and those that are shades of blue in the original map. Even slightly faint type is still legibly rendered.

There's no sign of a missing railway, border, river or co-ordinate line. Even at the theoretical limit of the sensor resolution.

Here's a closer view for comparison.

You're obviously not easily convinced. Therefore I suggest you do the experiment for yourself.

No, I'm not about to waste time and money duplicating the test with film. I've been there, done that, and bought the tee-shirt. So please stop simply regurgitating poorly done tests and dubious 'facts' found by Google, and actually get some results of your own to show. In short - put up or shut up!

Well I don't see the sense in mailing someone who hasn't been active here for over two years about something he posted eight years ago. I'm all for using old threads and posts, but within reason.

Also "A" spec of dust?! There's evidence over the whole crop.

There is not "numerous reports of missing detail because it's rare that digital cameras are used for repro work at such low magnifications. Usually in the real world you don't have anything to compare against.

As I said It's not really about the small artefacts. Those are just immediately visible symptoms of how much of a bayer demosaiced images detail is invented by the algorithm.

This contributes to the whole overly sharpened look of digital photos. A look that might lure you in at first but is tiring (not to mention not conductive to PP and printing) in the long run.

I applaud you for doing your own map image, even though I can see several bayer related ailments in both versions of the image. IE squared of circles. coloured noise around lines, coloured moire, chunky coloured aliasing blocks etc.

It's rare that I lug around two cameras, especially a digital and a film one. But I will see if I can remember to do a controlled test with film and digital in the near future.

I' be impressed if you could find what I said via google. This is simply deducted and observed, and then discussed with others.

I hate to be the one to pull the tone card, but I really don't get your testy tone. You haven't said or shown anything out of the ordinary that anyone with modicum of interest, in the technical aspects of photography doesn't already know.

As I said It's not really about the small artefacts. Those are just immediately visible symptoms of how much of a bayer demosaiced images detail is invented by the algorithm.

- Where's your evidence for this assertion?

Immediately visible? I think not.

The ability to 'pixel peep' at this level has only arisen since digital photography became popular and digital sensors achieved resolution levels (area for area) far greater than film.

The 'grain' or dye-clouds of film can easily be seen in modest (A4) enlargements from 100 ISO film. No such visual disturbance is seen with a digital camera of 10 or 12 megapixels. We have to consider reasonable usage - not scrutiny with an eye to criticism of tiny imperfections - which film has in abundance.

even though I can see several bayer related ailments in both versions of the image. IE squared of circles. coloured noise around lines, coloured moire, chunky coloured aliasing blocks etc

- Square circles? No! You're seeing things that simply aren't there.

The moire aliasing arises because of the extreme resolution that the digital camera is capable of. The individual ink dots simply wouldn't be resolved on film of an equal area and ISO rating. Film has an emulsion depth, which prevents the image being totally sharp. If we also throw the digital image slightly out of focus, then most of the moire goes away.

As for sharpening. I left it at the software default. It was much less noticeable in the RAW and TIFF versions I used before cropping. However, conversion to the crappy JPEG format for display here entails the generation of artefacts that are beyond control.

Sharpening is a matter of taste, rather than an inbuilt 'feature' of the digital image. However, many B&W film workers seem to want to provoke the 'edge' or 'adjacency' effect - AKA Mackie lines - by using stand development or so-called acutance developers. Do you object to such practises as well? Since they can look just as artificial.

I' be impressed if you could find what I said via google. This is simply deducted and observed, and then discussed with others.

- Exactly. Deducted and observed from an apparently very film-biased viewpoint.

If film was so good, and digital-capture so bad, why would the general 'bias' be about a thousand to one in favour of digital?

I simply have a passion for the photographic image. It really matters not what medium is used, and when film was all that was available, I used that. Now we have a medium that is demonstrably superior in its ability to capture detail, to give a vibrancy and 'cleanliness' of colour that's unparalleled, and which is less polluting and arguably less costly.

I can now use a digital camera barely bigger than one used for 35mm film, and just as portable, but which returns prints superior to those I had to lug a medium format rollfilm camera around to get in the past.

I don't need to compare those results at the pixel level to simply see that their clarity and vibrancy is better than those that film ever gave.

I wonder why you seem to be doing a lot if not most of your posting in the film section of this forum?

I can see very little, if any film artefacts on A4 or even larger. I wonder what kind of scanner or enlarger you are using? Ever heard of grain aliasing?

With a digital image it doesn't take much enlarging before you begin to see matrix artefacting.

Consumers turned on film because they never really cared about their images. "Free photos" was just too much of a lure. Only later came the claim of superior quality of digital.

Professionals less enamoured with cost, though still a factor of course, thought they where in more than a hurry than before and used that as an incentive to switch. That depended entirely on the profession and type of photography though. Many professionals helped keep film alive in the intervening last fifteen years.

"Detail, vibrancy and 'cleanliness" is all very debatable.

Cleanliness is again in part a result of interpolation and demosaicing. There is no such thing as correct colour. The most important thing in that regard is to get as much spectral information as possible to begin with.

Vibrancy in an electronic sensor is a result of the dyes in the Bayer array and cross contamination from sensor site to another. Neither of which has anywhere near ideal properties in the fundament implementation of CMOS.

Detail, is again as already stated more complex than having it or not. How is it arrived at? How is it represented? How does the visual system of humans like it at different scales af reproduction?

Have a look at a pre demosaiced pre-RAW capture from a sensor. and then from the correct but dull linear interpolation to get a feel for what is going on.

This page for example: IPOL Journal · Malvar-He-Cutler Linear Image Demosaicking

The "I can see it's better with my own eyes" argument doesn't really hold up historically or to closer logical scrutiny.

Looking at the area around Port Harcourt in magnification on both images reveals all the artefacts I mentioned

Moire and colour fringes are not a result of extreme resolution. It's a result of not enough resolution, a matrix of pixels and bayer filter. Any kind of post filter will not fix it.

Films depth is of often used as an argument. For the aperture and mechanical tolerances users in any camera, it will very seldom bee an issue, Film flatness is a bigger problem, but neither are pathological.

"Pixel" peeping was always possible with a strong loupe or a microscope and as such is not new.

Just as demosaicing interpolates the sensor captured image, you can, and to a larger extent in the near future, with new deep learning algorithms, will be able to filter out the luminous noise on top of detail that grain is, and get a "cleaner" image if that is ones desire.

Zooming in on an image is not the correct way to evaluate the quality. The faint halos around black lettering are probably due to sharpening and the JPEG conversion process. The correct method is to sample the original image, RAW or TIFF, before sharpening, then convert to JPEG for publication in PN.

I don't see Moire patterns in the example above, rather patterns created in the printing process of the dots where colors overlap. The best way to mitigate aliasing and Moire is to reduce the resolution of the optical system ahead of the digital sensor. This is tantamount to squinting when reading. In this example, I don't see individual dots, so the regular patterns may be subject to aliasing, which causes repetitive patterns to appear. It is impossible to remove aliasing or Moire after the fact without leaving gross artifacts in the results.

Artifacts due to de-matrixing the Bayer patters are usually seen as smeared colors, and is particularly noticeable in monochromatic patterns close to the colors of the filter elements. Other Bayer artifacts occur in the edges and corners of the image, when a high angle of incidence creates parallax between the filter and sensor. This is improved by increasing the height of the rear node above the sensor when designing the lens.

Moire usually appears as wavy or curved bands of color, and is especially visible where there is no color in the subject (corrugated steel, picket fences and resolution targets). It only appears when the lens has greater resolution than patterns in the subject, and the cell spacing of the sensor is comparable to the spacing of those patterns in the image. In standard resolution sensors (<= 24 MP), anti-aliasing filters (low-pass) are. used to eliminate aliasing. High resolution sensors get by without AA filters because Moire patterns only occur in details too small to see without considerable magnification. Even then, it only occurs if the lens resolution exceeds that of the sensor by a factor of 2 or more.

Resolution aside, the biggest difference between film and digital is the way in which each renders colors. I find it impossible to match the way Ektachrome, for example, renders color in a direct digital capture. On the other hand, copying an Ektachrome slide with a scanner or digital camera produces an image which is nearly identical to the colors of the slide. My conclusion is that digital capture is more accurate than film. One may prefer the colors on film, because colors are distorted in a way that pleases you.

I created this composite image to demonstrate the resolution if a 42 MP sensor, taken with an excellent lens, and tripod. I also used pixel-shifting, a set of four consecutive images, rendered in a way which eliminates any Bayer filter effects. The composite was created in TIFF format before being re-sampled and converted to JPEG for display. You won't see any halos, but you can almost see individual branches in trees over a mile away, with a very wide angle lens. The inset image is at the original resolution, pixel = pixel. Pixel-shifting is often used when making archival digital images of art and historical documents.

Sony A7Riii + Loxia 25/2.4, 1/250 @ f/8, ISO 100

L

I wonder why you seem to be doing a lot if not most of your posting in the film section of this forum?

 

I can see very little, if any film artefacts on A4 or even larger. I wonder what kind of scanner or enlarger you are using? Ever heard of grain aliasing?

- You wonder why I'm getting 'testy' when you're so patronising?

I was looking into the issue of grain aliasing before you probably got out of nappies. That's if you're out of them now.

With a digital image it doesn't take much enlarging before you begin to see matrix artefacting

- Total tosh!

Vibrancy in an electronic sensor is a result of the dyes in the Bayer array and cross contamination from sensor site to another.

- More tosh!

Have a look at a pre demosaiced pre-RAW capture from a sensor. and then from the correct but dull linear interpolation to get a feel for what is going on.

- Do you think I haven't?

It's difficult to get any idea of the final image through the excess of green.

Have you looked at the undeveloped latent image on film? An equally irrelevant and pointless question.

Cleanliness is again in part a result of interpolation and demosaicing

- Keep these howlers coming!

Moire and colour fringes are not a result of extreme resolution. It's a result of not enough resolution

- Well here's the same map thrown slightly out of focus, at the same scale as previously shown. Strangely, the moire has disappeared. The definition has also dropped close to the smudgy level of Daniel's film example.

"Pixel" peeping was always possible with a strong loupe or a microscope and as such is not new.

- Yes, but you can thank digital technology for the ease and clarity with which such magnification can be done.

It's also responsible for debates such as this.

Film, or not film?

Apart from Googling; what have you actually done by way of research?

At this point, I think any discussion of the superiority of a capture medium is academic. My 12mp full frame DSLR looks better(and captures more detail) at ISO 400 than any 35mm ASA 400 speed film I've ever used, while my 24mp DSLR destroys any speed film I've ever used(except perhaps tech pan) and at 36mp I'm well into "6x7 with perfect technique" territory.

I still shoot a lot of film because I like the process. I enjoy playing around in the darkroom, and the results I get with wet prints from B&W film. I like the color rendition I get from specific emulsions-but as Ed says that's not because they give a more accurate color rendition. Rather, it's just that I've come to like how Velvia 50, for example, "sees" the world.

I use and will continue to use film because I enjoy it and like how it looks(and again enjoy the process). I use and will continue to use digital because it gives me technically superior results and allows me to do things that were impossible with film(like hand held ambient light photography in dark indoor spaces without using rediculously grainy films at like TMAX P3200).

I'm sick of looking at that section of Africa. Let's get back to the original question of film colour-depth.

What seems not to be appreciated is exactly how big and clumsy film 'grains' and dye-clouds are, compared to the scale of modern digital camera photosites.

Therefore I dug out my microscope for the first time in years, and stuck some bits of film under it. Together with an accurate resolution graticule that last saw use determining the true resolution of a scanner. (Incidentally nowhere near the maker's claim!)

The piece of Fuji 100 ISO negative film was placed emulsion to emulsion with the glass graticule, and photographed by direct projection using an 8mm Leitz oil-immersion objective.

Here's part of the resulting photomicrograph - warts and all, complete with hotspot.

The area of film was a partially fogged section of leader, but even at this scale the individual dye-clouds can be seen to give nothing like a continuous tone.

It was impossible to get both the graticule and dye clouds in focus simultaneously. So here I've reconstructed the 5 micron bar-space pattern and superimposed it on the dye-clouds.

The top line has been subdivided into 5 micron squares, which is a pretty average size for digital sensor 'pixel' spacing these days.

If we zoom in on the pattern - which we can easily do, thanks to digital technology - then it becomes pretty obvious how few dye-clouds can be contained in the same area as a digital pixel.

I see no evidence of millions, or even thousands of shades of colour being represented in each 5 micron-square area of film. It's also fairly obvious that the dye clouds have a fairly well-defined boundary, and that their size defaults to a fairly large average.

The conclusion must be that; because the overall colour is dithered from small and finite-sized patches of cyan, magenta and yellow dye, that the impression of shades of colour is area-dependant. Ergo, we cannot have a perception of infinite shades of colour in a finite area. As the averaged area gets larger, so does the perceived colour-depth, but we cannot perceive small detail and good colour-depth together. They are mutually exclusive.

I also think that Ed's example above adequately squashes any accusation of 'blockiness' or pixelation aimed at modern digital cameras. I'm pretty sure you would have to resort to 5"x4", or even 10"x8" film to come close to seeing the kind of detail and colour clarity shown in the above example.

And the map 'mystery' is solved!

See this thread.

Where Danielleetaylor replies to a similar query -

"Also, what's with the missing lake in one of the comparisons?

Same map "model", but not the same map. That said...what lake? There are some color and line differences, but I never noticed a lake difference."

It transpires that the map used by Daniel was a different copy from the one used in the film example, and that the film example wasn't even taken by Daniel.

We can all sleep easier in our beds tonight for knowing that.

To quote Sherlock Holmes -

'When you have eliminated the impossible; whatever remains, however improbable, must be the truth.'