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

[S8 Booster]

Well....just an idea of how to solve the K40 no-mo processing situation....

The plot is this... with so many die hard transfer / software geeks around here ..... consider this..

lets assume this:

1) K40 getting processed as an "original" BW (encoded)
2) The geeks above writes some codecs/matrixs which sort of "duplicates" the original complex Kodachrome processing - ... the encoding lying in the greyscales....

well it may sound weird but old BW movies seemed to get an automated post colouring so isnt it worth looking into?

challenging the geeks...

just throwing out this idea since most 8mm films are transferred to digital these days anyways.....

shortage of suitable K40 supply may make the effort not worthwhile anyway.

good shoot 8 fun....

[carllooper]

Why is it next to impossible?

The main reason is that the three layers, processed as BW, while physically separate, are married together in the emulsion. Without dyes in each layer you can't use light to distinguish (separate out) the information in each layer.

Why is it not impossible?

Well you could separate the layers (don't ask me how) and then digitise each layer separately.

Carl

[carllooper]

Here is a more mathematical analogy.

What are the three numbers, when added together, result in the number 42?

a. 5 + 22 + 15
b. 31 + 10 + 1
c. 14 + 14 + 14

Carl

[carllooper]

In standard mathematics, does two plus two always equal four?
You would think so, if only by definition. But if that's the case, then when does one plus three get an opportunity to equal four?
8O

[S8 Booster]

well, yes, but no, as they say in Japan when things are next to impossible and ieeeeehhh on indraft when things are ultimately impossible,.... nevertheless....

imagine a K40 grey scale coded RAW file ultimately beating the original K40 colour imagery by a huge margin.....

shoot.... more...

[carllooper]

Here's another way.

You use precision optics, with a very shallow depth of field, to bring into focus just one layer at a time. The other layers are out of focus.

Or more realistically: a precision hologram is made, and digitally dismantled.

Definitely not impossible.



Carl

[mr8mm]

I think the colorized B/W films shown on TV were colorized by guessing. You know what skin tones are so color skin and faces pink or whatever. Suits could be brown or blue or black and so on. Trees are green, roses are red , violets are blue and I love Kodachrome.

I have been told Kodachrome can be processed as a B/W reversal. Maybe someone can explain how that works. And what ever happened to people in the Czech Rep. An April Fool Joke?

J.S.

[S8 Booster]

to me the red colours has always tended to show up a certain easy to recognise grey tone/tint viewing BW films...

now, if one shoots a short clip of a colour chart at the beginning of each roll to "encode" the greys...?

would be very interesting if the 3 layers could be extracted separately

if i remember correctly the emulsion side of the K40 when processed the original way looks like a 3D landscape as if etched by acid. now, could this "landscape" formation be "detected" for the layer separation as Carlloper mentions above?

i think it may be possible to get it work one way or another but will it be worth it? to be or not to be....?

if properly resolved it could be the best 8mm Kodachrome results ever...


shoot.. more film ....

[carllooper]

There's no easy way to do it.

As already mentioned, but perhaps not well enough, the way being imagined, of matching grey tones to colours won't work.

Suppose you read a particular grey level in the digital scan, eg. a level of 42. The question your proposed system would be asking is: to which colour does this grey tone correspond.

Okay, the computer looks up it's stored colour chart and finds that the following colour matches the grey level of 42:

R=12
G=13
B=17

But hang on, so does the following:

R=1
G=2
B=39

and indeed a vast cross-section of colours will match the grey tone.

Carl

[S8 Booster]

see the problem...

never thought of it as an easy task... or even possible but it would be fun if it worked out...

now, another highly theoretical idea on how to "identify" the layers... shoot 3 short takes of a grey scale chart or other suitable reference with each R-G-B filters in turn or 1 3x33% cake slice RGB filter for the "encoding".

no idea but still 1 idea....

shoot....

[carllooper]

The only way to decode the film is to use some physical signature regarding the layers, that is different for each layer. The conventional signature (dyes) are not there but other signatures are.

The most obvious signature is that the layers are distinguishable in terms of space, by however many nano-meters. So a precision laser scan using holographic priniciples could separate them out.

But of course that's not easy. I only mention it in the spirit of free speculation where there is no particular agenda (such as resusitating K40).

In the area of film restoration free ideas could very well happen. In a hundred years time some archival project might want to restore colour from a BW processed K40, and this is one way they could do it.

Carl

[carllooper]

see the problem...

never thought of it as an easy task... or even possible but it would be fun if it worked out...

now, another highly theoretical idea on how to "identify" the layers... shoot 3 short takes of a grey scale chart or other suitable reference with each R-G-B filters in turn or 1 3x33% cake slice RGB filter for the "encoding".

no idea but still 1 idea....

shoot....

Ok. In the spirit of theory and experiment I'm with you. Lets forget about why we might want to do this (eg. resusitating K40) and concetrate on the challenge itself - which is - I think - much more interesting than saving the world as we knew it.

And could be a lot of fun. And do-able as an experiment in it's own right.

Carl

[carllooper]

However one thing I'm not sure of is what happens when K40 is processed as BW. I'm assuming, but I could be wrong, that the the three layers just get processed as BW without any dyes. From memory I recall there are specific filters between the layers that determine how each layer represents some area of the spectrum, so in principle, the information is still there after a BW process, but somewhat inaccessible.

Now one way of encoding (which doesn't require K40 - BW will do) is to use a filter wheel in which each colour component is recorded on separate (and subsequent) frames. This is how technicolour works.

Well sort of. The difference is that the image has changed from one frame to the next - however that can be managed using optical flow techniques that are a lot easier to do than laser interferometry because no special (expensive) hardware is required.

[carllooper]

But of course, the filter wheel/subsequent frame/optical flow approach, doesn't specifically require K40. So doing it on K40 doesn't exploit anything in particluar about K40.

Here's a question. What's wrong with the following proposal, (apart from shooting on video)?

You shoot on video and separate out the video into three channels that are recorded on K40 - shooting the computer screen through filters. Each channel is exposed in such a way that occupies a particular range of grey levels. For example, red occupys the range between 0 and 33%, green occupies the range from 33 to 66, and blue occupies the range from 66 to 100?

Carl

[S8 Booster]

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

actually i was thinking of the possibility of making 3 separate scans or processing digitally 3 different R-G-B film "strips" out of the 1 grey original scan adding 3 RGB filters 1 at the time onto each separate R-G-B strips and later "merging" the 3 digital strip into 1 "colour"...

just 1other idea


shoot....