LightField Forum

Light Field and Computational Imaging News

Forum

Welcome to the LightField Forum!
We believe that there should be a place for LightField photographers and enthusiasts to talk and exchange themselves, and present their favourite LightField pictures.
Please note: In order to prevent forum link spam, all external links are assigned the nofollow attribute!

Welcome Guest 

Show/Hide Header

Welcome Guest, posting in this forum requires registration.
Pages: 1
Author Topic: Analogue lightfield - possible?


dave.renda-
ll
newbie
Posts: 7
Analogue lightfield - possible?
on: 2013/12/18 13:14

Hi everyone.
I had an idea yesterday, but my understanding of lightfields is basic, so I have no idea if it will work or not! I thought it would be good to post it here so that people who know better could save me from wasting time on it if it wont work. It's a method of capturing and displaying a light-field using photographic film.
Please excuse the rough diagram:

Image

So on the left side we have a pinhole array in front of some photographic film. I shoot and develop 4x5 film, which has a resolution of over 100 Mpixels so a good place to start! You've probably noticed there isn't a main lens. My theory is that this setup will capture light intensity, colour (if I use colour reversal film) and direction. Please tell me if I am way off!

Now to the right side of the diagram. If I use the same pinhole array and the developed film in the same positions, with a diffuse light source, I should get the light coming out of the film in the same colour and direction (although probably a fraction of the intensity!). So if I was to look at the film from the position of the observer, and I was able to focus my eyes at the right distances, would the setup give me a 3D image with accomodation depth cue?

Feel free to tear my half-baked idea to shreds (but be gentle please :))
Thanks
Dave


markus
Administrator
Posts: 68
Re: Analogue lightfield - possible?
on: 2013/12/18 13:30

Hi Dave,
your setup should work if you make one modification: on the observer side, the pinhole array needs to be between the film and the observer, not behind the film. Nvidia used a similar setup (using a small display and microlens array instead of film and pinholes) for their near-eye light field display.


dave.renda-
ll
newbie
Posts: 7
Re: Analogue lightfield - possible?
on: 2013/12/18 13:49

Thanks Markus! I saw that nvidia display and was very excited - until I saw the price of each of the parts!
Wouldn't the pinhole array on the observer side of the film invert the lightfield, causing the light rays to converge instead of diverge (I feel another diagram coming!)? In my mind, all the pinhole images would have to be inverted for it to work with the pinholes in front. It would work that way with the correct lenslet array though, if I had more money than sense :)


dave.renda-
ll
newbie
Posts: 7
Re: Analogue lightfield - possible?
on: 2013/12/18 15:56

I can see the need for the pinholes on the observer side though, the photographic film will diffuse the light slightly. That would mean all light angles are visible from all directions (although more faint). I suppose I could invert the pinhole images by using the pinhole array with another sheet of film, that would work when viewed through the pinhole array (I think).


dave.renda-
ll
newbie
Posts: 7
Re: Analogue lightfield - possible?
on: 2013/12/18 17:44

My new diagram:

Image

In fact this isn't a bad thing for film negatives, as the colour needs to be inverted too.

Next question is how to make/buy a pinhole array. I think I could make a wonky pinhole array with 3 holes per mm, which might work for testing purposes. The linear resolution of the film could easily be 400 'dots' per mm, so perhaps a printed mask would give better resolution?


markus
Administrator
Posts: 68
Re: Analogue lightfield - possible?
on: 2013/12/20 13:20

Hi Dave,
when you record an image through a plenoptic setup (microlens or pinhole array), you basically encode the image and thus need to decode it using the same optical setup. Your sub-images will be upside-down and contain directional data when taking the picture. viewing the complete picture through another identical pinhole array will split these sub-images into individual lightrays again, so it should already be ready for viewing through another pinhole array (which restores the right-side up image). I don't see why you would need an "inverter" for that.
There is some research being done with printed masks, which is said to be both inexpensive and promising. Bear in mind, though, that lenses will collect much more light, allow for higher spatial resolution of your array (i think), and microlens arrays are already available for just over 100 usd.


dave.renda-
ll
newbie
Posts: 7
Re: Analogue lightfield - possible?
on: 2013/12/20 13:56

Thanks Markus,
I couldn't find a microlens array that would cover 4x5 film (95x120mm), and I assumed that getting one custom made would be much more expensive! Certainly for evaluation purposes a printed mask seems to be the best option. I might try printing a mask myself on a laser printer for proof of concept, but I'm not sure if 600dpi will be fine enough. If it could manage 60 pinholes per inch it might be enough for initial tests - I'll need to get the old microscope out to see what the print looks like!

As for the 'inverter', I was basing the need for this on the direction of the light rays. I may well be wrong! I was considering that the light rays should leave the viewer in the same direction as they enter the camera (as if the camera wasn't there). If the pinhole mask was used to view film1, wouldn't the light rays leave the viewer at the opposite angle?


dave.renda-
ll
newbie
Posts: 7
Re: Analogue lightfield - possible?
on: 2013/12/20 14:15

Here is another diagram with 'film 1' in the viewer.

Image

Perhaps this happens when you don't have a main lens in the setup? It might be worth me considering that, because I have a large format camera and lens already and it would save me film (if I used slide film instead of negative film).


dave.renda-
ll
newbie
Posts: 7
Re: Analogue lightfield - possible?
on: 2013/12/20 18:52

Another diagram for a proper plenoptic camera with a main lens:
Image

It would need the film to be closer to the lens than the focal point of the closest object for the viewer to work correctly (I think!). No problem with a large format camera. It would need a really slow shutter speed with a pinhole mask though. A rough calculation for the no-lens approach gives an aperture of f5, which is pretty reasonable. I've no idea what it would be with an f5 aperture both on the lens and the film - I assume there is a formula somewhere - but it would make a shutter speed in whole seconds I expect?
Pages: 1
Mingle Forum by cartpauj | ElegantPress by Theme4Press. Version: 1.0.33.2