NAC 16mm Analysis Projector

[carllooper]

Will be doing the first transfer tests in a few days time using some 16mm film out of the archive. I haven't implemented the LED light source yet, or the shutter sensor. The test will be just for the optics and camera set up - to fine tune the mounting arrangement for that. And to get some quick wins out of it.

Carl

[timdrage]

Sounds like a pretty amazing machine and an interesting project, good luck!

[carllooper]

Sounds like a pretty amazing machine and an interesting project, good luck!

Thanks Tim.

I find it quite pleasurable working with mechanical systems and seeing what can be done with them. I'm not in any way mechanically inclined, but I'm learning. My day job is as a software developer where everything takes place inside my head. But with physical mechanical systems, such as the projector, they act as a kind of alternative reality. There are many things in common but also many things completely different. The weight of the thing. The energy required to move the thing from one room to another (physical exercise). The fact that it won't disappear were I to lose power on it. It acts as a counter-balance to my more cerebral activity.

It's also the history it holds - which is why I'd never make irreversible changes to it. Augment it, certainly. But not mangle it. It's too interesting as a thing in itself. History acts as a kind of memory which allows us to see more clearly where we might be today and why. Treating it as a recycling warehouse is the last thing on my mind.

Carl

[carllooper]

Quick Projector Test with 16mm from the archive

https://vimeo.com/44357577

The projector test was just a quick test, projecting onto a piece of a paper, ie. rather than directly onto the camera sensor.

The exposure time, per frame was 1 second, demonstrating that the projector shutter blades won't be an issue.

The camera was on a tripod and I was clicking the projector's frame advance, and camera trigger manually. As a result there was some minor camera shake per frame. I re-registered the result by motion tracking the bottom right sprocket, however due to the camera shake altering the perspective of the image the bottom right sprocket registration was not a complete remedy for the entire image.

I used the native halogen of the projector. The lighting isn't flat.

The LED lighting (to be implemented) won't necessarily be any more consistent than the hallogen. One trick is to photograph the illumination without any film and use that as a per-pixel template to correct subsequent captures. A similar technique is actually used by digital cameras to correct differences in sensor cell sensitivity. It's called the sensor's "dark image". The camera records an image while the shutter is closed and uses this "dark image" as a template to correct subsequent photographs.

The LED lighting's main purpose is to ensure each colour channel of the sensor is being used to their full extent. For example, in a halogen lamp the blue signal isn't as bright as the red and the green, so the blue signal occupies less of the available range in the blue sensor, so is more open to "staircasing" when subsequently corrected in the digital domain.

Another way around this is to use the halogen but to take multiple exposures of the same frame. This increases the dynamic range of the result, minimising staircasing in any subsequent colour correction.

Focus.

The image was focused on a piece of paper using the projector ens. And refocused by the camera lens. It will be a lot sharper when I use the enlarger lens to project the film directly onto the sensor. I also need to put the film in the projector the other way around so that the emulsion side is facing away from the lamp. This is not how film is normally projected but it actually makes the image sharper if it's done this way as one is not imaging through the acetate base.

[carllooper]

Testing the limits of film becomes rather simple using the analysis projector. Even just projecting it on a wall.

I projected the image up large and got right in close with a digital camera, capturing a small portion of the image. The capture resolution was the equivalent of scanning the whole 16mm frame at 24K (or Super8 at 12K, or 35mm at 48K).

In and of itself this didn't tell me much but when I differenced that against the equivalent of scanning 16mm at 12K (or Super8 at 6K, or 35mm at 24K) I was still getting a strong signal for the edges in the image. Plenty of noise everywhere else of course.

I did a heavy noise reduction on the difference signal, which looks weird. But the fact that there is still a perceivable image in the difference signal means that one still hasn't exhausted the signal component of film, even at these ultra high definitions.

So basically, if you are happy with the grain structure of film (ie. the way it looks to the eye when projected) then to reproduce the remaining signal component in digital can easily require scanning the film at definitions way higher than 4K. Easily.

So this "overkill" theory regarding scanning resolutions is really just complete and utter nonsense. The only real limit I've found (so far) is not technical at all, but one's budget!

Of course all of this is cheaper (in terms of hardware) to understand and forsee when you work through the theory (as Sheldon on The Big Bang Theory would appreciate). On the other hand, working through it experimentally (emperically), while more expensive (in terms of hardware) has it's own peculiar and profound rewards. There is that physical aspect. I wouldn't be seeing cows in paddock, if I were looking at the equations.

Where do those cows come from?

Carl

-------------

Regarding the cow question ...

In relation to the difference between theory and practice touched on above, there is a framework which retakes ownership of the term "theory" and gives it back to practice, in which practice becomes just as theoretical as 'theory'.

Phenomenology and empericism have never quite attained the respectability of rationalism. One of the criticisms of phenomenology/empericism is it's apparent focus on "presence". While this is historically true, one must be careful not to throw the baby out with the bathwater. Presence is certainly a problem, but a critique of presence is not quite a critique of phenomenology and empericism.

Deleuze once described himself as a transcendental empericist. But what does this mean? In simple terms it means extending traditional empericism (and phenomenology) beyond the immediate, the moment, or the present, into the domain of time and duration.

It is not quite right to say the cinematographic image is in the present. What is in the present is what the image 'represents' but not the image itself, which, in cinema, as in painting, is never to be confused with what it represents. (Preface to the English edition, Cinema2, Gilles Deleuze)

The image, in a Deleuzian universe, constitutes something which can be regarded as "transcending" the present. Rationalism is no different in this respect. Rationalism also transcends the present. But in rationalism (traceable back to Plato) an image was regarded as no more than what was immediately visible. In rationalism an image becomes a secondary thing - an effect for which some (rational ) cause was outside the image, beyond the image, in a reality that was not an image.

An alternative approach, and what Deleuze does, is to expand the definition of an image beyond what is immediately visible (or audible, tactile etc) into the invisible. Invisible images? Yes, but unlike rationalism, such images would not be fundamentally invisible. The concept of "hidden images" might be more appropriate here. Or "transcendental images". The image, not just as what is immediately visible but also what can be hidden (eg. a photo in a shoebox) or what could come to light in the future, neither of which are in the present. In a Deleuzian universe, these would remain part of what constituted an image.

The catalog of images in a Deleuzian universe includes not just those of perception, but of recollection and dreams as well. In this regard, can it be said that Deleuze goes beyond where the Stoics might have gone? Was it not the Stoics who put down "figments of the imagination"? No, what the Stoics argued is that what we imagine, as much as what we perceive has it's basis in what we experience or could experience. So even figments of the imagination would be part of the Stoic universe. What they rejected was anything which was fundamentally beyond experience. It would be the fundamentally invisible (as distinct from the hidden) which the Stoics would rhetorically call "figments of the imagination".

Now this sort of thinking is the complete opposite of classical philosophy of the platonic variety. It is also the complete opposite of what many theoretical physicists would propose. But it's something that experimental physicists and artists might otherwise intuitively understand.

I find this helps me heaps to understand the cows in the paddock.

[escher]

Hi Carl,

Stumbled onto this post trying to find some information on the NAC projector and was pleasantly surprised. I was planning on making some similar alterations to create an optical printing machine to be used with a Bolex. Originally I thought it might be possible to wire one of the controllers to a computer interface like the JK optical printing boxes (to make step printing, skip printing, etc. possible, or atleast easier). Have you made any progress with the controller?

Aside from optical printing, it would be nice to use the machine for digital transfer as well. I found a little mirrored box for telecine the other day at Goodwill, but think that removing the lens of the projector and trying to shoot the film itself with a macro setup would produce a better result. What's your recommendation for getting a clean transfer?