Using Light Meter inside Camera Obscura

[Chan Tran]

If the diameter of the aperture is equal to the focal length (1:1), there is no attenuation of light and nearly 100% will make it through (minus absorption in the glass of course). That's effectively the same as having no aperture. To prove it; take a reading of a scene like normal, suppose the meter reads f/5.6 @ 1/250th. Now, put your lens @ f/5.6 and place the meter at the film plane. At f/1.0 the meter will tell you that you need an exposure of 1/250th.

Now, as for a f/0.95 or larger aperture actually magnifying or amplifying the light, think about this; the image formed by the lens is significantly smaller than the subject that you're photographing. With that in mind, if you can pack the same amount of light into a smaller area, you can indeed amplify the light at the film plane, relative to the subject.

No one seems to believe me on this point, but the reality is that physics proves it. The 2nd point you make about amplifying light is a little harder to wrap my head around, but I believe there's an equation out there somewhere that can show this as well. (I gotta be careful touting the "math", because if you asked me to do it, I'd have trouble...)

Here, the first two equations on the wikipedia exposure value page (http://en.wikipedia.org/wiki/Exposure_value) should be able to prove it if you feel like plugging in some numbers.

When you said take the meter reading at the scene like normal, how do you do that? Using incident or reflected light meter? Same thing when you said meter at the film plane, incident or reflected. And if measuring reflected light at the film plane how would you do it?

 

[holmburgers]

Johnielvis, now I'm really confused! But, it sounds like the boat's still holding water, right? Can you extrapolate on this a bit, it sounds interesting but I'm not sure I've wrapped my noggin around it yet.

Chan, the easiest way to do this would be to get a gray card and measure that with a reflective meter outside of the camera. However, an incident reading in the same position as the card should give the same reading. Then, measure the image of that gray card being projected at the film plane of your camera with an incident meter, or if you're really small (or have an enormous camera) you could take a reflective reading.

 

[Chan Tran]

Johnielvis, now I'm really confused! But, it sounds like the boat's still holding water, right? Can you extrapolate on this a bit, it sounds interesting but I'm not sure I've wrapped my noggin around it yet.

Chan, the easiest way to do this would be to get a gray card and measure that with a reflective meter outside of the camera. However, an incident reading in the same position as the card should give the same reading. Then, measure the image of that gray card being projected at the film plane of your camera with an incident meter, or if you're really small (or have an enormous camera) you could take a reflective reading.

If you use a reflected light meter and measuring the subject (gray card) and then with an f/1.0 lens and assuming there is no light loss due to flare, fall off, whatsoever and then you measure the illumination at the film plane with an incident meter the meter would indicdate 4.5 stops less light.

 

[johnielvis]

ok--that was early in the AM..I just noticed that .2 factor....where the devil did THAT come from? so I started up again with a diagram of object lens and image...then I thought...there's gotta be MORE out there...and there is...turns out that it's along thelines I was thinking--it DOES depend on object distance or magnification...however it simplifies for infinity focus to a parameter called "lens throughput" which is T=1/(4*F^2) where F is the fstop number of the lens. So that .2 factor is also wrong..it should be .25...however it can be lower for objects which are closer than infinity, so that's probably why they chose the .2 factor. the total energy passed through the lens non infinity is the throughput times 1/(1+M)^2 where M is the magnification--this is only good for a circular objects (corresponding circular images) and a circular aperture of course...

anyways...the 1/F^2 factor is dealt with using the f/1 setting trick...the rest you get by setting the iso by a factor of .25 (or .2 or whatever)....and these are of course gross approximations....but thats for an object omitting some energy...we can MEASURE what is hitting the lens with a light meter, in fact, that's what we're interested in...so I'm gonna continue looking...but short answer is use f/1 but compensate by moving 2 to 2.5 stops for the film speed......I'll see if I can come up with those other formulae myself...or I'll just keep looking for the derivation...it's much easier than doing it...and less prone to mistakes...and more fun.

 

[Chan Tran]

The 0.2 or 0.25 (assuming the lens has no loss due to focusing closer than infinity, object not at center, lens flare etc...) with the condition that the subject reflects 100% the light incident on it. If the subject is a gray card (or some kind of average scene) then it is 0.18 x 0.25 = 0.045 and that is -4.47 stops less.

 

[johnielvis]

nope--this is already based on light emanating from the object--so it's already reflected---this is light already coming towards the lens, so your 4.5 stops is not to be adjusted for since it's done already...the .2 or .25 has nothing to do with reflectivity since we start off assuming light already heading towards the lens from a source in front of the lens. in fact you can come up with different numbers besides .2 or .25 depending on how you assume the light strikes the lens coming in. the lens throughput is an exact energy transmission through the lens

the reason for the .2 or .25 or .3 or .7 (whatever it is) depends on the distance of the source and the collomation of the light coming from it. this assumes that the light has already been reflected or that the subject is a light source like a light bulb or something.

in other words the "perfect" lens would be f/0.5--this would give the same light meter reading outside the camera as inside the camera--so you meter FROM THE LENS pointed at the object with light coming towards the camera, take a reading--go inside the camera with an f/0.5 lens and meter and you get the same reading. This is assuming I think perfect collomation where the object is very huge to cover the entire view field...real objects are smaller than the view field and are closer than at infinity...so that's where the .25 starts getting towards .2 or something else.

see this here:

www.ieee.li/pdf/viewgraphs/2d_barcodes.pdf

 

[johnielvis]

there is a quick and dirty method to come up with the 1/(4F^2) number but even that involves the concept of the radation "view factor" or "shape factor" used in radiation transfer theory.....the concept here is covered in the wikipedia article on etendue and there is a specific diagram showing how it works for refraction through a lens...see this...it's nasty hairy--if you did any radiation heat transfer problems, then this should look familiar

http://en.wikipedia.org/wiki/Etendue#Conservation_of_etendue_in_free_space

SO...the upshot is: the light inside is basically attenuated by this 1/(4F^2) factor...So--set the meter to f/1 to take care of the 1/F^2 and set the iso to 4 times more sensitive film and then the exposure time will read correctly inside the camera obscura.

oh...and USE the REFLECTED light meter...not the incident!!!!! perfect "spot meter" for zone system here---I believe sinar had such a back that had a probe that went in by the ground glass an did just this for super accurate spot metering.