The K40 Saviour...(s)....

[carllooper]

video? no big deal any remedy which works will do..

But what's wrong with the proposal I put forward?

This is part of what theorising is all about - not just proposing theories - but knocking them down.

I've presented a theory for you to knock down. Have a think about what is wrong with it. I'll give you a clue: there is something wrong with it.

Carl

[S8 Booster]

nothing wrong . possibly unclear from me but i supported your suggestion,

anything goes to solve the matter...

shoot..

[carllooper]

Hey - I'm on your side. Don't worry. You're thinking.

But the proposal I put forward doesn't work.

My question about what was wrong with it was a serious question. There is something wrong with it.

But I want you to work that out. When you've worked it out, or at least made an attempt to work it out - we can continue.

cheers
Carl

[S8 Booster]

if you are aiming at the 3 layers - my thinking was...

once upon a time on this forum i made a post - how to reduce the contrast of the K40 in transfer - to highlight the dark shadows i tried to use the own film´s duplicate overlay neg frame as a filter - adjusting the transparency and opacity to balance the high and low light - and it worked....

mattias on this forum then wrote a script or plug to equal this option - it worked well.... ill try to search the forum for the post.

anyway/s i´m not sure if it was this you had in mind but it is something i thought of as a possible problem.

additionally i have been thinking about extracting the 3 layers - if a precision optics can "almost" extract each layer but not clean - in theory - if the layers are not possible to extract clean out of the film - ie. that for layer 1 some of 2 and 3 will be "ghosting" on 1 it could be possible to take layer 1 and subtract 2 and 3 from 1 to make it clean/er. same goes for layer #2 and layer #3.

one way or another - possibly....

half way down this post you will find mattias´script:
viewtopic.php?f=1&t=3941&hilit=script&start=15


shoot....

[S8 Booster]

just adding it here:

Kodakchrome processing procedures:

http://www.kodak.com/global/plugins/acr ... z50_03.pdf

pps:
http://www.randrews4.com/ProcessK-14.pps


shoot....

[carllooper]

Hi S8 Booster,

Regarding the theory I was proposing. This is what is wrong with it:

If you encode red as a greylevel between 0 to 33%. And green as greylevel beyween 33% and 66%, and blue as a grey level between 66% and 100%,

then the problem is this:

You can only have one of these primarys encoded at any one place.

For example, how do you encode the colour Yellow? Yellow is a mixture of Red and Green. So you would need both a greylevel between 0 and 33 (red), AND a greylevel between 33 and 66 (Green) encoded in the same place.

One solution is to separate them, for example, into separate layers. Starts to sound like what K40 is all about ...

But then you are back to the problem of how you separate the information in each layer. Until they can be separated (eg. using dyes) you don't know how much there is of each, in each layer.

Ok. One way to separate them is to inject dyes into each layer.

In other words, one way to process K40, so that it encodes colour, is to process it as colour instead of BW.

Carl

Addendum.

I'm a fan of science fiction. Especially that sci-fi from the 60's that border on scientific incomprehensibility (eg. most of Philip K Dicks work). The meanings of these works don't lie in the science per se, but the social and political contexts which frame and are framed by the work. A contemporary equivalent of P K Dick might be a writer like Kaufman (Being John Malcovitch, Eternal Sunshine of the Spotless Mind).

Scenario.

In the distant future, there are supercomputers that can analyse a roll of unprocessed film and computationally reconstruct the image exposed on such. Indeed they are so powerful they can even compute the image recorded on an unexposed roll of film - eg. what the photographer was going to shoot but didn't.

So what we should do now is go into a museum where a roll of unexposed K40 is on exhibition and simply meditate on what what we might shoot on it. That's all we have to do.

[S8 Booster]

good great fun maybe some day.... only need a Chinese Cray killer..

ive allways been fancified by the theoretical possibility of extracting images from exposed but unprocessed film

i see your problem as described ... no easy solution to that.

anyway, sometimes there is a workaround going in the opposite direction if it will solve anything here is impossible to know.

one example of the opposite way to attack a certain problem was the mathematcian Aabel who proved that equations of 5th power can not be resolved

secondly ive worked out a method of visualise and shoot down any stealth plane around very simple in fact ;)

here is a supposed sample of hand processed K40 as BW on the tube. no idea if it tells us anything at all .. at this time.

Hand developed K40 experiment

possibly one could part K14 process (some of the initial not too complex processing stages) the K40 to establish some useful parameters for digital colour re-processing ....

lots trial and error in the coming if so....

shoot.....

[carllooper]

The posted film looks quite interesting but it's hard to fully appreciate in the digital transfer. I imagine it looks fantastic when projected. The sound track is excellent. The context makes it powerful - hand processed K40.

possibly one could part K14 process (some of the initial not too complex processing stages) the K40 to establish some useful parameters for digital colour re-processing

The processing is irrelevant if you can't, in some way, separate the layers. That's the main problem. The rest is easy. You can process it as crazy as you want and then map the results back to the original signal that entered the lens.

BUT that's the easy part.

The difficult (but not impossible) part is physically distinguishing the information in each layer. That's what you would have to solve. One can solve it theoretically/mathematically (fourier transforms etc)- but the implemetation of the theory is not image processing software.

The implementation involves assembly of hardware (lasers etc) for interrogating the medium. A simple optical image using filters will not do it. The theory will tell you that.

Only when the information has been physically separated is the remainder simple arithmetic (ie. software).

By way of analogy, I'm developing software to increase the definition of film signals (encoded on Super8 or any other gauge) but the original signal (where the definition is limited by noise) has to be physically transferred at the target high definition in the first place before the noise can be digitally removed and the resulting target definition signal reconstructed.

Carl

[S8 Booster]

By way of analogy, I'm developing software to increase the definition of film signals (encoded on Super8 or any other gauge) but the original signal (where the definition is limited by noise) has to be physically transferred at the target high definition in the first place before the noise can be digitally removed and the resulting target definition signal reconstructed.

Carl

i have followed your postings about this which is very interesting and one of the main reasons to trig this post. i once joked with the transfer guys here about the need for the highest resolution - down to the very last silver grain is the way to go if one wants full control over the imagery. seems like you are about to get there

now, back to quasibusiness... what about a hi-res MRI style cross section scan of the film using and electron microscope? could that be used to separate the layers? ... using a Cray or something to process the data? combination scan? Electron microscope to detect cross section layers while the image scan is done by a different scanning method or going for it full... using all out electron microscope cross sections adding up to the complete frame.
some will call it an overkill for sure but it woulda been good great fun to see the result even for only 1 frame.

could be a suitable hi-res suitable cross section scanner system out there..

keep shooting...

just a sample - non photographic film just to show possibility/ies

[carllooper]

Yep - something like that should work
Slicing the film (like a bread slicing machine) into very fine slivers, and scanning each sliver ...


Carl

[S8 Booster]

Yeah, but my mind was more set on super hi res digital slicing creating an 3D hologram frame

Shoot......

[carllooper]

Ah yes. Something more practical.

If you consider a number of laser beams passing through the same point in the emulsion, but from different angles, the densitys sampled along each beam should present a composite signal where the density signal at the focus point (where the beams intersect) can be made to reinforce while density signals elsewhere, can be made to cancel out. The more beams there are the better the result.

Carl

[S8 Booster]

Yeah, results becoming more like this:
http://www.youtube.com/watch?v=Eugu1qGu ... ata_player
Giving the tool needed to do just about anything creating the "best" film look/imagery ever seen....

X-Y axis forming most of the imagery while vectors in the Z axis deciding most of the layer information and hologram of the image surface,

Which makes me think; image & image data all determined by vector calculations, isnt that a very efficient way to store and control the image and image data?

Now, that would be back to you i guess, not even needing a Cray for image processing this time.

Shoot
Shoot.....

[S8 Booster]

Oooppppsss.... i mistakingly came to post an MRi image of my brain and can not get it deleted....


Shoot...

[carllooper]

:lol: