The first serious expectations of near death of film technology appeared after the emergence of digital photosensitive matrix capacity of 6 megapixels. Optimists immediately said that in most cases, such number of elementary carriers of the image is enough.
“Yes, yes!” – Harmony become the second of marketing professionals.
New blow to the traditional chemical photography
suffered on Feb. 11, 2002. On this day it was announced that appeared on the market Foveon X3, a new type of matrix. Prior to that pixel digital matrices were not quite full-fledged: each of them could perceive only the signal of one of the three primary colors.
The other two missing color had
determine by interpolation, based on information from neighboring pixels. New technology has allowed to obtain information about all three colors at each pixel of the matrix.
The founder of Foveon Carver Mead (Carver Mead) hastened to declare in an interview with New York Times (number from 11 February 2002):
«There is no longer any need to use film»
“For there is no need to use the film”
Official web site of Foveon contains more cautious statements. There’s a FAQ section on the impact of new technology on the film image is given the answer:
«While the many advantages of digital photography have led to explosive sales growth, image capture systems based on silver halide will be around for a long time to come.»
“Despite the fact that many of the advantages of digital photography have led to rapid growth in sales volumes, the system of registration of images using silver halide to be used for a long time”
At the moment, the hype around the Foveon X3 little abated. But fotomir continues to discuss digital technology and to build the prediction of the date of imminent extinction of film photography.
Here we will not build empty predictions. Let us better than the mundane things. Let’s try to find out how much the equivalent of pixels contained in one frame narrow screen the film.
No, no, dear reader, in search of an answer, we are not going to go into the theoretical thicket.
We get the answer to this question in two ways, relying on common sense and
Universally evaluation.
1. The simplest and most obvious way
According to the experts (so they say always, when they want to avoid detailed explanations and to give additional weight to his words 
) to scan 35 mm frame with high-resolution 4000 dpi has no meaning. Most professional film scanners are working precisely on this resolution. There are, however, and more accurate sampling device, but as I said, many believe that a great need for them there.
But around 4000 dpi equivalent of 150 dots per mm. If we had placed on the narrow screen frame size 24 x 36 mm pixels with a density, we find that their total number will reach about 20 million This is the answer to this question.
Notes:
Yes. For fans of precision: 24 x 36 x 150 x 150 = 19440000.
b.. For more great fans of precision: 4000 dpi are in fact 157.48 pixels per mm. Therefore, the number of pixels per frame will be:
[24 x 157.48] x [36 x 157, 48] = 21423151.
(Square brackets denote the operation of taking the integer part.)
in. From the standpoint of common sense and notes. and b. do not have much sense.
2. A little more rigorous approach
Let’s try to estimate the number of pixels in the narrow screen frame, based on data on the resolving power of the system “film + lens”. In this approach, we absolutely do not care what factors determined by this resolution. We easily forget (if you know, of course) about the theorem of Kotelnikov-Nyquist, about the effect of layering multiple digital processes (aliasing), about the size of grain and about many other things. Let your eyes only on the final result.
When in Internet publications indicate the resolving power, they often refer to
Norman Koren website. In general, placed on the site Koren materials is very instructive and interesting. Not all his conclusions can agree, but the author clearly knows and understands what
writes about. And this is a lot. In his subsequent arguments will be based on data Koren.
At the end of the first part of his comprehensive article «Understanding Image Sharpness» Coren gets an assessment of the resolution of the system “film + lens”. For a combination of “film Velvia + Lens Canon 28-70/2.8» resolution amounted to 36.8 pairs of lines per mm. Drop the contrast to 50% and 68.6 pairs of lines per mm at the fall of the contrast to 10%.
Measured by the reduction of the contrast resolution of up to 50% can only extreme pessimists. Aim at a level of 10%, perhaps, not risk, and many optimists. Truth is somewhere in between. It is reasonable to assume that the resolution of the “lens + film” is approximately 50 pairs of lines per mm. Most likely, Koren would agree with this conclusion. Note that this is not the ultimate score. Properly treat it as a standard size.
Let us ask ourselves: how many pixels need to resolve a pair of lines? Coren discusses this issue at the beginning of the second part of the same article. Two – a little, but four – a bit too much. Coren more like figure 3. I must say that he was not alone. Thus, the firm Sony has long been proposed to evaluate the resolution of cameras based on just 3 pixels on a couple of lines.
Feel free to side with views of Koren and firm Sony. Multiply 50 pairs per mm for our top three. Voila! We get the same 150 pixels per mm. This means that these pixels in a frame size 24 x 36 mm fit almost 20 million
Findings
So, we can safely assume that, when there will be a digital matrix size 24 x 36 mm, consisting of at least 20 million pixels, the demise of film photography is just around the corner.
Sometimes you can hear the objection: but it is not always such as high resolution film okazvaetsya demand. In many cases, is enough and more modest resolution. No doubt, this is so. But “many” does not mean “all”.
These devices are quite common and far from worthless, despite the fact that they do not allow to consider all the information from the film.
To print enough to have a resolution of 300 dpi, therefore, the scanned image is thus designed to increase the pocketbook shot 9 times (to about 20 x 30 cm).
It is also useful to remember that the typical diameter of the permissible range of blur is 0.03 mm. When scanning with a resolution of 2700 dpi at this diameter will fit three pixels. In other words, the resolution of 2700 dpi is really quite enough in many cases.
Resolution 4000 dpi is necessary either at high magnifications, or with higher quality requirements. And that, in both cases when shooting should take into account that the diameter of the permissible range of blur should be less than the typical value of 0.03 mm.
Scanned with a resolution of 2700 dpi frame will contain 9.7 million pixels. Apparently, this value should be acknowledged as the most minimal estimate of the number of pixels in the narrow screen frame.