-----Original Message-----
[tirade deleted]
You regularly make snide and derisive comments about anything having to do with digital.
Or maybe it's just you that take them like that.
Before I post something, I make sure I have some arguments to base myself on. In this case, it is the theory of image sampling and reconstruction. Here is what stands behind my affirmation:
Both film and digital sensors are sampling device. In the case of digital, for each sample, you have a measuring of the weighted average of light intensity for a small area around each sensor pixel's center. This measuring is represented by a number - the "value" of that pixel. In the case of film, you have the same thing, with the following differences:
- the sampling grid is irregular
- the light intensity is represented in analog form, by a dye cloud density and area (shape)
None of them are ideal samplings (which would be the convolution of the continous image field with an array of Dirac delta function). However, they are close enough and could be used within certain limits instead of them.
Then comes the image reconstruction stage. The theory says that this can be optimally done with *interpolation* functions. And this is what happens with digital images - for your display purpose, the values of the points between the samples are calculated. In the simplest case, with a square pulse function, which gives you what's commonly known as a "nearest neighbour" interpolation. An optimal interpolator is for example a "sinc" function (a sum of sin(x)/x functions) - for an ideally sampled image, it would give you a perfect reconstruction.
The interesting thing happens with film. As long as the grain is smaller in size than the circle of confusion of the enlarging lens, you have an interpolation function applied too. Not an optimal one, but you get close to "grainless" images. When you get to magnification orders that make the grain comparable or bigger than the COC, there's no more sensible interpolation happening. You render grain-blank-grain-blank. The reconstruction function is the worst possible. There's nothing interpolated for the space between grains. You can come back to the interpolation situation by adjusting the COC and making it larger - thus the suggestion of out-of-focus enlarging lens.
Personally, I find your attitude annoying, and your opinions ignorant.
You're entitled to have any opinion you want. It's a free world.
cheers, caveman
