A Super film from Germany Gibabit film

greenspun.com : LUSENET : B&W Photo - Film & Processing : One Thread

Agfapan APX 25 is on the way out. Agfa has discontinued this film. However there is a new super fine grain high resolution B&W film from a small comapny in Germany

http://www.gigabitfilm.de

The English site is not up yet, but there are great example of the increidble resolution of this film. It has a resoultion of 720 lpmm at 1:1000 and ASA 40, far exceeds Kodak technical Pan ASA 25/ 320 LPMM Look at the 1000x enlargement of watch on the lady's hand at upper left corner, which in turn is a 100x enlargement of the 35mm negative on upper right A 35mm negative can easily enlarge to 100" x 150"

-- martin tai (martin.tai@capcanada.com), October 11, 2000

Answers

To heck with the film!
I want to know where I can get the lens that took the shot. By my reckoning, that sucker has to be capable of 600 lppm, and must have been used at f/2.
I also want the camera that's capable of holding the film in the focal plane within a wavelength of blue light.
I'm not doubting the authenticity of the picture (much), but how come the girls wrist looks like it's been pebble-dashed, while the watch is almost perfectly defined?

And here I was thinking that those B movie cliches were all make believe.- "Hey, Joey, that guy across the road, just zoom in on his fingerprints will ya? I think we gotta match for our suspect."

-- Pete Andrews (p.l.andrews@bham.ac.uk), October 11, 2000.


So far it's vaporware, and imho probably will remain so.

Several people who know much more about optics and physics that I have pointed out that Gigabit's claims border on and may surpass the physically impossible. In order to get that sort of resolution on film you'd need a perfect lens working at f1 or so, perfect film flatness, perfect focusing etc.

-- John Hicks (jbh@magicnet.net), October 11, 2000.


Does anybody happen to know what the lpmm is for holographic film? Somebody on a forum boasted that they worked with holographic film, and it had 1000+lpmm. If that is so, then this film is quite plausible. (Will this film available in 8x10 sheets like TechPan? >:- })

Can someone clue me in?

Thanks.

-- Brian C. Miller (brian.c.miller@gte.net), October 11, 2000.


Hmmmm, I smell a rat. Which reminds me that I have to vacuum the floor!

If this "Gigabit" film comes in 120 I will try it. For the time being I have stored twenty rolls of 120 APX 25 for future use, and will order some more. Efke will sell more films after this desicion by Agfa to discontinue APX 25.

-- Patric (jenspatric@mail.bip.net), October 11, 2000.


They announce a 35mm roll film. I am curious. www.monochrom.com sells already 4x5 sheets of this film:

http://www.monochrom.com/MonoC/asp/rmiArt001.asp?wareng=513

Meanwhile I ordered 5 bulk rolls of APX25.

-- Marc Leest (mmm@n2photography.com), October 11, 2000.



Microfilm with resolution of 800 lpmm is used quite often in subminiature camera circle ( Minox, Minolta 16 etc) For example Fuji Super HR film is capable of resolution of 800 lpmm. Holographic film has a resolution of 7000 lpmm. High resolution film is nothing new. The only thing worth noting is the speed of this film at ASA 40 it is considerabley faster then Fuji Super HR at only ASA 6.

-- martin tai (martin.tai@capcanada.com), October 11, 2000.

Martin, this reminds me very much of H&W Control film of the '70s.

There were two H&W films; one was an Agfa microfilm of EI 16 and the other was an Agfa copy film of EI 80. Both were developed in H&W's proprietary developers, which I suspect was a POTA concentrate; it was sold in small air-excluding ampoules to be mixed immediately before use.

The EI 80 film, H&W VTE, for which Kodak High Contrast Copy Film could be substituted, was very much like today's Tech Pan except that it had rather reduced red sensitivity.

The slow film, VTE Ultra, was another matter. With careful work I could get prints that showed much higher sharpness, an overall combination of resolution and local contrast, than I can get with Tech Pan developed in anything. VTE Ultra was also panchromatic.

Now...I can get EI 50 with TP with reasonable contrast (Ethol TEC 1:15)...so the speed claimed for Gigabitfilm, assuming it's some sort of microfilm, to achieve a higher speed for the same quality 25 years later isn't unreasonable.

H&W vanished due to lack of demand. That Kodak has essentially shut down R&D in b&w materials and Agfa has dropped APX 25 reflects a market that doesn't bode well for Gigabitfilm.

It'd be interesting to learn what Gigabit's developer is.

-- John Hicks (jbh@magicnet.net), October 11, 2000.


I wasn't contesting Gigabit's claims for its film, just questioning its sensationalist publicity material.
Let's put some facts on the table:
1) The resolution of a film cannot exceed the limitations set by laws of physics. Primarily, light cannot be focused to a spot size smaller than its wavelength. For blue light this is 450nm, a resolution limit of ~1100 lppm, for red light its 700nm, or 700 lppm. To actually test a film to these limits of resolution isn't a trivial task.
2) The resolution of a lens is diffraction limited to a figure of 1392/N using light of wavelength 698nm (N = f stop number). Most lenses approaching this level of performance work at an aperture of f/5.6 or so, and cover a very small field. This means that the practical limit of resolution is 248 lppm.
3)The OTF of the combined film and lens reduces this figure even further. Optimistically, a figure of 200 lppm might just be obtainable.
4)All the above assumes perfect focal plane alignment and flatness. An error of one wavelength of light results in the limit of resolution of the lens being reduced to 20% of its maximum. i.e. 248 lppm drops to 50 lppm. The surface roughness of a film based gelatine emulsion is probably on the order of + or - 100nm.

I used to work in micro-photolithography, producing masters for semiconductor devices, so I've seen these theoretical limits in action.
Achieving 5 micron features using visible light is not an easy task, and that's using a Lippmann emulsion on glass plate, exposed with monochromatic light, through a very expensive specialist lens, set on an accurately aligned optical bed, in a temperature and vibration controlled environment, and developed to maximum gamma.
Perhaps now you can see why a continuous tone snapshot of two girls on a museum roof, showing the same kind of resolution stretches my credulity a little.

-- Pete Andrews (p.l.andrews@bham.ac.uk), October 12, 2000.


Perhaps Gigabit's line pair testing was not done with lenses, but with lasers. If holographic film is in the 7000lpmm ballpark, maybe the Gigabit film was tested with the same methodology.

Anybody know whose camera lenses have the highest lpmm resolution?

-- Brian C. Miller (brian.c.miller@gte.net), October 12, 2000.


John, the Gigabit film might well be a reincarnation of one of Agfa's Copex Pan copy film developed in special developer. The secret may be in their developer; I doubt a small company has the resource to develop film. 720 lpmm is about doublt that of TP, and on par with Fuji super HR. However this gibabit film may has much finer grain than Fuji Super HR.

The watch surface on the 1000x enlargement is about 1 cm x 1 cm, meaning the image on negative is about 1/1000 cm or 1/100 mm There are only 2 line pair of details in the watch surface, which translates into only 200 lpmm, that is easily achievable with a good 35mm lens

-- martin tai (martin.tai@capcanada.com), October 12, 2000.



Well, I measured the '1000x' picture as only a 600 times enlargement on my monitor, and the watch hand and hour markers measured less than 2mm across. Now 2mm / 600 = 3.333microns = at least 300lppm.
200 lppm from a 35mm lens? Yeah, right! And to think we paid thousands of pounds 20 years ago for a lens capable of imaging 5 microns at a fixed 10:1 ratio, when we could simply have walked into the nearest camera shop and got a 50mm standard lens off the shelf, that would do the same at any old reproduction ratio, and for less than a hundred quid at that time.
BTW, I made a boo-boo with the depth-of-focus statement. It should have factored in the numerical aperture of the lens, giving a depth of focus of several wavelengths of light. Even so, if the emulsion is a couple of microns deep.....

When you can duplicate the results shown on Gigabit's website Martin, tell me exactly how you did it, and how much it cost you.

-- Pete Andrews (p.l.andrews@bham.ac.uk), October 13, 2000.


In the days of Kodak High Contrast Copy film, I determined the resolution of my Leicaflex 50mm Summicron and 90mm Elmarit R lenses with USBS targets placed at 76 focal lengths from the nodal point of the lenses. At this distance the maximum resolution that can be measured is 240 l/mm. The Summicron achieved this resoution at f4 in the very center of the field! And the resolution was distinct, meaning I did not have to strain my eyes to see the spaces between the lines. I tried to see at what magnification this level of resolution could be manifested/used in printing. So I placed the negative in my enlarger and used my Omega grain magnifier to examine the resolution target at 24X with the enlarger lens aperture set at f4.5. I could just make out the target lines, and concluded that this level of resolution would never be seen in a print unless is was feet across and I could find an enlarging lens that could resolve over 240 l/mm at that magnification.

This test, and others I conducted over the years, was a great education. With high resolution lenses, the effects of diffraction can be readily determined. Already at f8 resolution is degraded, and this is especially apparent at f11 and smaller. But at f11, the resolution (i.e., 68-80 l/mm) across the field with high grade lenses is very uniform, and the contrast and illumination is also uniform. With most general use films, practical resolution is 68-80 l/mm, so this aperture is fully capable of using the maximum potential of most fine grain films.

So, resolution and fine grain is fine, but what kind of gradation will the new film provide? If it is like most high resolution copy films, it won't be that hot. But I hope that it is!

Eilert Anders

-- Eilert Anders (eilert@dav.com), October 13, 2000.


Pete, I can get about 170 lpmm with my Minox camera on Kodak technical Pan film. With Gigabit film there shall be no problem at all to get well over 200 lpmm on film

You had experience with process lens, but it seems to me you have seen any really sharp camera lens.

The following is the original picture enlarged to 3.5x 5"

In the above picture, 2 cm from the lower right corner is a street sign post. The street sign is only about 0.1mm on the original negative

This 0.1mm segment enlarged 200 x

You can see the characters "ONE". Each character, for example the "o" is only 0.02mm on the orginal negatives.

That is only techpan, with gigabit film in Minox, I have no doubt 0.01mm is possible. There you go.

-- martin tai (martin.tai@capcanada.com), October 14, 2000.


Pete, as for why process lens cost a lot more. The asnwer: 1. It is absolutely photometric, with no distortion 2. it must be flat field. 3. It works at close range

All these is hard to get in photo camera lens

-- martin tai (martin.tai@capcanada.com), October 14, 2000.


One note: the Minox B picture was taken hand held. If I put the camera on a heavy duty tripod, the image will even be sharper.

I am very excited about this gigabit film. A Minox size 8x11mm negative can be enlarge to 30"x 40" with no visible grains !

-- martin tai (martin.tai@capcanada.com), October 14, 2000.



Martin. I have 35mm negatives showing the same kind of definition as your example, but this is still an order of magnitude or more away from the kind of thing that Gigabit are showing.
If you look at the structure of the image in your example, it has the typical appearance of diffraction limiting; the outlines are beginning to show fringing effects. Now compare that with the slightly blurred outlines of the watch hands shown on Gigabit's site. Gigabit's image just doesn't show the characteristics of a diffraction limited image.
Come on Martin, let's have a bit of a reality check here. Your picture shows a road sign at about the same apparent distance as Gigabit's watch hand. The letters on the road sign have to be, what? 1cm thick at least: A hand on a ladies watch is 1mm thick at most (more likely less than 0.5mm).
Do you really think that Gigabit film will improve the resolution of your lens by 20 fold?

BTW, the lens in question was not an ordinary process lens, but a high resolution f/4, 'wide field' (~15 degrees coverage) 10:1 reduction lens.
Our really high definition (>1000lppm with blue light) f/1.4 lens was hidden away in the step-and-repeat camera, but only covered a few degrees. Nobody attemps 100:1 reductions in one bite and expects them to be 'sharp'.
If you really think that a normal camera lens working at an unspecified ratio, with broad-band light can beat it, then go ahead and try this Gigabit 'miracle film'. Then see if Gigabit can supply the alternative universe, where the fundamental laws of optics don't apply, and less money buys you a better product.

-- Pete Andrews (p.l.andrews@bham.ac.uk), October 16, 2000.


Peter, I believe gigabitfilm will improve the resolution of Minox camera lens by about 2 fold from 170 lpmm to about 300 lpmm

You probably don't know about microdot photography.

In 1925, Emmanuel Goldberg already developed a microdot camera capable of 500:1 reduction in one step. Later, in WWII, Zeiss microdot camerras were widely used. 500:1 in one step.

-- martin tai (martin.tai@capcanada.com), October 16, 2000.


The key to gigabitfilm'ss high resolution seems in its special grain structure: in a developed grain, there may be 8 or 10 silver crystals arranged in three dimensional manner with gaps in between, unlike other films which silver crystals lump together. To take advantage of this special characteristicws of gigabitfilm's grain structure, special enlarger lens may be necessary.

Point source enlarger and special enlarger of 100x with colum height in 80 cm range. Ordinary enlarger lense light incident angle is differert from the camera lens, therefore the individual crystals in a developed grain may block the path of light, resulting in the crystals apparently lumps together and reduced the effective resolution of the film.

The 1000x picture cannot be obtain with ordinarly enlarger lens( with which improvement in 2 to 3 times is still possible ) It was enlarged with a 100x lens(specially made) on Agfa enlarger with a 35mm camera withouut lens, but with gigabit film loaded, small 1/100mm area of the negative is scanned with this special set up to expose the film. Then enlarge this film ten x on photopaper to yield 1000x picture

Seehttp://www.gigabitfilm.de/html/deutsch/vortraege/linear.htm.

-- martin tai (martin.tai@capcanada.com), October 18, 2000.


When 35mm format gigabit film is available, I shall buy a few rolls and cut them up into Minox format film and loaded into my Minox B camera. I may not need any special enlarger lens. Because my Minox enlarger model II uses the same 15mm f3.5 lens as the Minox B camera (COMPLAN lens ) and 100x enlargement is not difficult to get-- all I have to do is to lay the Minox enlarger flat and project the image on the wall, at a distance of 1.5M. Then I can put a piece of gigabitfilm to make an interneg for 1000x enlargment or simply a piece of Ilford paper for 100x print.

There is advantage is using same lens for camera and enlarger, because the viewing angle from the lens to the negative will be very close-- that will keep the perspective of silver crystals (in grains ) right.

-- martin tai (martin.tai@capcanada.com), October 18, 2000.


Your translation is a bit off Martin.
That's not what the article says at all. I've got a German speaking colleague working on a full translation, which I'll post to your other thread later.
Basically, what the guy from Gigabit is saying, is that he tried a specially constructed lens, made from cannibalised parts, but still wasn't satisfied with the results.
He then went on to use the lens that he actually took the negative with (an el-cheapo Vivitar 90mm macro lens), as an enlarger lens (good move or what!?); projecting the image onto another piece of Gigabit film, held in a KB monobar camera. Then from this duped negative he produced the final 1000 times enlargement. (Yeah, yeah, yeah!)

I'm sorry for the sceptical comments Martin, but it's all I can do to keep a straight face at all this moonshine.

-- Pete Andrews (p.l.andrews@bham.ac.uk), October 18, 2000.


> He then went on to use the lens that he actually took the negative with (an el-cheapo Vivitar 90mm macro lens), as an enlarger lens (good move or what!?); projecting the image onto another piece of Gigabit film, held in a KB monobar camera.

Say what?!! That's outrageous.

-- John Hicks (jbh@magicnet.net), October 18, 2000.


There are people here confused the resolution of film WITH resolution of lens. These are two entirely different things. There are a number of Kodak and Agfa microfilm out there which are capapble of several thousand lpmm resolution. Some people then argue it is not possibel to focus 1/7000 mm on film. When one talk about 7000 lpmm resolution film You don't understand what resolution of film really means. Resolution of "Lens on film" is the convolution of resolution of film with the resolution of lens. Since the convulved resolution of Film+ lens is can be measured, and the intrinsic resolution of lens (ie aerial resolution) can also be measured by observing with microscope on aerial image, the resolution of film can be calculated. For example of a lens of intrinsic 400lpmm forms image on a film X with resulting resolution of 250 lpmm, the resolution of film X is 700 lpmm, if the on film resolution is 350 lpmm, the resolution of film is 3000 lpmm; if the combo res is 370 lpmm, the res of film is 5000 lpmm. In other word, the higher the resolution of film the closer the combo lpmm approaches that of the lens's intrinsic resolution. The rule is 1/combo-resolution = 1/resolution-of-film +1/resolution-lens

-- martin tai (martin.tai@capcanada.com), October 18, 2000.

Peter, your method of "measuring : the thickness of one line and from which conclude that the image has 3000 lpmm resolution is dubious at best. The definition of "resolution" involves two points or line, (Raleigh criteria ), ie whether this two points or lines can be viewed as two separate lines and not merged into on. If ther is one line, there is nothing to resolve.

Further, if you get only the image of a line, and you want to judge the MTF , the criteria is the energy distribution inside the Airy disk, not the diameter of the disk itself. A thin line may indicate, contrary to what you 'calculated" a rather low resolution, because it has low energy concetnration. A good example is you cannot see small stars in night sky, the energy distribution of their image is so low, it is no visible-- the diameter =0, which does not means the resolution is 1/0=infinity. In many 35mm camera pictures, power lines could becomes so thin and nearly disappear, it indicate low resolution , not high. In conclusion, your '300 lpmm" conclusion was incorrect, based on your faulty method

-- martin tai (martin.tai@capcanada.com), October 18, 2000.


The best way to get large, sharp, grain free images from a Minox is to trade it in and buy a good 35mm (or medium format) system, not to use "miracle film" which may promise much but deliver little. In this case, that seems strangely appropriate.

High resolution film is nothing new and high resolution optics can take advantage of it, but I'm afraid Mr Tai's knowledge of optics is tenuous at best.

-- Jack Russel (techphoto2000@mail.com), October 18, 2000.


Martin, the watch hand actually does constitute a line pair. It is a dark line surrounded by light. The light on either side of the hand should have been diffracted into the dark area, and in fact the thickness of the hand should appear thinner in the final print than the original. Therefore, I would argue that to image that watch hand as accurately as is shown in Gigabit's sample photo, at least twice the resolution would be necessary than a first order calculation would indicate
The ludicrous figure of 3000 lppm has never passed my fingers until now.

Anyway. Now that we have Herr Gigabit's methodology, we can do some real calculation.
A 90mm lens at 50 metres is a reduction of 555 times. Divide a 0.5mm watch hand by 555 = 0.9microns. That's 1.8 microns per line pair, or 555 lppm! Multiply by 2.5 (the f number used apparently), and we arrive at almost the exact theoretical limit of resolution. What a coincidence! And that's several millimetres off-axis!
I'm surprised Vivitar haven't been flooded with orders for this miraculous lens, and by your formula for calculating film resolution, we find that the film can image an almost infinite number of line pairs per millimetre.

Now, I for one, am more inclined to believe that a commercial company is telling lies, than I am to believe that a relatively cheap lens can perform at the very limits of theoretical resolution, off axis, at its maximum apeture. In fact I'm more inclined to believe in fairies. (Or at least their stories.)

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


Folks, I need to jump in here as the light-handed (or is that light- headed?) quasi-moderator here.

I do not pretend to understand the fine points being discussed here, nor do I find them terribly interesting as they seem to be of the "how many angels can dance on the head of a pin" variety compared to my approach to photography.

However I have noticed several personal flames against one poster in particular, and have exercized my powers to remove those parts of posted messages.

It is one thing to say someone's methodology, theory or technical knowledge is full of holes. Its another entirely to say "oh and he's been kicked out of such-and-such a newsgroup" or that a certain poster is a first-class jerk. I don't give a rat's tushie about someone's past on-line history, as long as their posts are not insulting and abide by the "rules of the house" in this forum (go to http:// www.photogs.com/bwworld/bwforum.html to refresh your memory).

Bottom line? Feel free to keep this discussion going (boring though it may be to me), but please refrain from personal attacks or invective. Challenge a theory or thought? Certainly. Getting personal? Take it to email.

Many thanks for your cooperation.

-- Mason Resnick (bwworld@mindspring.com), October 19, 2000.


I hear this film is now available in Germany. It comes with its own developer. Has anyone tried it yet? The point, I would think, is to do more than Tech Pan or easily available microfilms.

-- Michael Kennedy (m005kennedy@netzero.net), December 20, 2000.

Apparently this film is already sold in some shops in Germany, for about 17DM ($7.5) a 36 exp roll plus chemicals. Some mention Monochrome as a source. Another source quote Karl-Heiz Klang of 520010093110-0001@t-online.de

You may also search "gigabitfilm" in http://www.deja.com/ there are some discussions about this film, some one was testing it.

-- martin tai (martin.tai@capcanada.com), December 21, 2000.


Another dealer in Germany selling gigabitfilm: Foto-Gregor: info@foto-gregor.de

-- martin tai (martin.tai@capcanada.com), December 21, 2000.

I bought some test rolls of 135 gigabitfilm before the actual product release and made test shots from the "highest church tower on earth". You find the results for the next 2 month on http://home.t-online.de/home/stefan.hahn/GBF0102/Muenster.html (I will translate this to English asap).

-----

Concerning the previous discussion, I would assume that the hand on the watch is a special wide one (or are in fact two hands). I did some calculations from the pixel sizes of the image on the gigabit website and concluded, that the "1000x" enlargement must have close to 100 microns across. The aparent width of the hand is 4 pixels which comes to 4/147*100 = a bit less than 3 microns.

------

If anybody is going to prepare minox-gigabitfilms - I would very much like one for test (and compensate for any cost of course).

Stefan

-- Stefan Hahn (stefan.hahn@t-online.de), February 25, 2001.


Moderation questions? read the FAQ