The following article is about how to achive maximum quality from gigabitfilm. The Altavista traslator does not provide readable result. http://www.gigabitfilm.de/html/deutsch/vortraege/linear.htm

-- martin tai (martin.tai@capcanada.com), October 17, 2000

Answers

Response to Can some one translate this article (German) ?

Here's your translation Martin. May it give you much pleasure.

A thousandfold linear, or Gigabit film with its back to the wall.

To obtain the maximum enlargement quality with Gigabit film, first an introduction.

Please imagine the emulsion of a developed film in 3 dimensions. In this thin layer the developed Silver Halogen crystals are not only sitting next to each other, but also one on top of the other. This can involve a stack of 8 to 10 crystals. The light which comes from the taking lens hitting this layer has to pass through the depth of this layer. If this exposing light falls on the edge of the film format, then more area is needed to reproduce the same level of detail. In other words, the resolution goes down. One could also say: The more acute the angle [of the light] incident on the layer, the worse the resolution becomes. There are well-known optical laws governing [or concerning] this.

What does the developed negative look like, particularly if it's been exposed at an acute angle and if you are trying to image high levels of detail? What follows next is theoretical, the practical details will follow later.

If I look through a powerful loupe or a microscope at the developed film, then I will see so-called 'grain'. But this isn't really the silver grain, but only gaps between the individual developed crystals. These gaps form a 3 dimensional layer (as previously explained), and are also able to image with a certain depth the more-or-less acute light ray, with its level of detail. If I look perpendicularly onto a layer which has been imaged at an angle, then I will see a certain overlap [of the crystals], and the resolution will look worse than if I looked from the exact same angle of the original light ray.

Until now, this theoretical view has never played any practical part in gaining the maximum resolution.

The following happened to me when I attempted a 1000 x enlargement of my Gigabit film. (State of the art, as of 1989.)

First I made a special lens from ready made parts, to fit to a standard enlarger. The purpose being to enlarge an area of 16 x 21 cm by 100 fold, using a normal separation of 70 to 80cm between the head of the enlarger and the baseboard with normal separation from the film to the rear node of the lens. A modification of the light source with a weaker diffuser of my Agfa Varioskop enlarger was advantageous, for the achievement of normal exposure times.
Without the diffuser, there were disturbing diffraction overlays due to imperfections, or fluctuations in the refractive index [of the emulson], as with a point source [of light].

Test enlargements with Polaroid instant colour film with the 3 colour stripes showed freedom from distortion, image field smoothing, and even resolution, without disturbing chromatic effects. A lens designer [optical technical expert] who had just computed a new range of optical systems for aerial photography said after seeing these images: "If you had asked us to produce these quality parameters, it would have cost you plenty of money [a fortune]"

These enlargements produced with this 100 fold lens looked substantially better than with a conventional Apo. enlarging lens. With conventional Apo lenses, a direct 100 x could only be produced with difficulty in the laboratory, - this isn't really acceptable for the 'normal user'; because who has access to a projection area of 5 metres to make such enlargements?

I was still not happy with the [resolution of] edge detail, even with the 100 fold enlargements using my special lens. I knew that by using 'optical tricks' under the microscope, that I could get more detail from the negative. Thus I decided- after a vague clue from a publication of the 1930s (text attached)- to do as complete as possible a reconstruction of the conditions surrounding the taking of the image, in order to obtain the maximum image quality in ray path retracing during enlargement. [??]

The same lens as used in taking the image was attached to the enlarger, the same aperture (1:2.5) was used, as well as the same register [focal distance?] as on the camera. Of course the negative was mounted [absolutely] centrally, therefore everything was in exactly the same position as for taking the image. Since the taking lens used was one that can be corrected for every image ratio by using floating elements (Vivitar series 1 macro 90mm f/2.5, selected by the maker from only lenses which meet the design criteria), only instead of a distance setting of 50m, only 2m was used for the enlargement. I considered this acceptable. A KB camera, minus lens, was installed at the image plane on a micro slide. The camera was also loaded with Gigabit film. The focussing region was stepped [driven] through in about 40 increments of a few hundredths of a millimetre each. The best looking positive, with a low gamma, was again copied at 20 x linear onto Gigabit film, again using anumber of fine focussing steps. The finished negative served as the optimised negative for -all [scale] factors taken into consideration - the 1000 fold linear enlargemnet on normal photographic paper. Here was the big surprise: This image showed more detail than I seemed to see under the microscope!

About this image quality (State of the art of Gigabit film progress in 1989 - today it's even better.) it should be remarked that losses during multiple copying added up, and to further optimise the process - any condenser which is in use today presents some fundamental optical problems.
This article will give notes [hints] for further improvement to the optical specialist.

The note [clue] from the 1930s.

Photographic Industry 1933, page 880, Dr. K.Fischer, "Errors in enlarging apparatus." Quote:
.................. you should be aware of how the greyscale of the negative is altered during enlargement, for example, after notes from the Journal of Scientific Photography 1933 ,31, p 306, and Journal of Scientific Photography 1933, 32, p 410 (Photometry of areas of similar density but differing grain size.) Because it can be seen that the differences in tone from the negative are not translated [transferred] proportionally into the positive, but are substantially modified, depending on the aperture and on the surface area of that tone [density]. Starting from a certain size, say 1cm square, they are translated accurately into the positive, however if there is a 1mm square dark area within a bright field, that will be much lighter. As can be seen from an example on page 309, fig. 5 for instance from an absorbtion of 96.3% using parallel light, to 60% with an aperture of 1:0.5. From this result, in passing, the requirement is to enlarge using the same relative aperture that was used for taking the image, as well as to position the enlarging lens over the same point of the negative as was used for taking the image. Only then will the enlarging optics not introduce any errors into the translation of the tonal range. If this [deleterious effect] isn't noticed in daily use, it's because the errors in the photographic emulsion (non-linear density curve) cause much larger deviations in the tonal range translation.
......................................

This was the state of the art in 1933, now we are in the year 2000, and at last we have perfect films, and we would gladly have a tone value translation without errors.

-- Pete Andrews (p.l.andrews@bham.ac.uk), October 19, 2000.

Response to Can some one translate this article (German) ?

Pete, thank your for providing this translation. Note: the KB camera referred in the article means 35mm format camera. KB is an abbreviation for kleinbild, ie miniature. For subminature camera, they call it kleinstbild-- smallest format.

-- martin tai (martin.tai@capcanada.com), October 31, 2000.