filter behind the lens?

[mattmills]

So I was thinking it would be possible with the lens I use to just put a filter behing the lens in front of the shutter of my spankin' new M8. So I ordered an IR-cut filter from edmund optical, cut it to size, and popped it in for a test. What I found was that the filter for some unknown reason throws the plane of focus forward a bunch. So when the lens' distance scale indicates four and a half feet, I'm actually focused at 9 feet. I took the filter out, and repeated, and it focuses normally. The filter is fairly thick- specs say 3.3mm. Happens no matter which side of the glass, coated or uncoated, is closer to the sensor.

So, needless to say, I'm puzzled. But I was wondering if anyone here with more optical expertise had any ideas. The image will focus, and seems as sharp with or without the filter, but it's as if the sensor plane has been moved forward. I'm using a canon 50 1.2 for testing. Any thoughts?

 

[Avotius]

hm....I have to say thats a good idea to stick the filter inside the camera. The only thing that I can think of that is throwing you off is that the filter is rather thick and maybe is enough to alter the plane of focus acting almost like an extra lens element.

Other then that im stumped, its a neat idea though.

 

[PHOTOEIL]

Try to mount the filter on the back of the lens, the way it is done with foilfilters (like the Kodak-Wratten ones) in LF camera's.

Good luck,

 

[pfogle]

what you're finding is the 'optical thickness' of the filter - that's the change in the path-length of the light from the lens to the sensor when you put something in between with a refractive index greater than one (glass is around 1.4).

The effect is the same as when you look into a swimming pool, and it looks shallower than it really is because of the water.

As you say, it moves the focus point back a bit. That's why lenses which have rear filters either use gelatine (negligible thickness) or use a dummy clear glass when there's no filter inserted (like in long telephotos).

edit - this is also the reason why, when you get someone to remove the IR filter (like in the 20D special version for astro) you have to put a clear glass plate in its place.

 

[Riley]

You dont say whether you have an absorbative filter or reflective, as reflective could send light scattering all over the place inside the lens.

if you look carefully at M8 images without an IR cutoff filter you will see they are not focussed, try it both ways and you will see what I mean. I think this is because IR light (which without a filter you are letting past) has different properties. The different diffraction index angles through glass is one notable diversion.

Beyond this it gets murky for me, as with all lenses all forms of light pass through to be reduced at the IR cut off over the sensor. Whereas infront of the lens IR light doesnt pass through. But Im supposing the very short register of the M lenses exaserbates this problem, allowing more than usual angular light to pass through.

Certainly we know that an IR cutoff infront of the sensor wont work. Whether this is due to the angular differences in differing light wavelengths or just the fact that retrofocus lenses wont physically fit Im not sure. If it is case 1, your idea is lost. Or if it is just that adding a thickness of glass to a lens system shifts the focus poit ~ same.

Given the system works ok with an IR cut infront of the lens, but for the occassional flare, and about which would probably happen anyway with older lenses, generationaly poorer coatings, not designed for digital use. Why do you want to do this ?

pictured is Sigma's foveon equipped SD14 with IRcutoff filter in the mirrorbox bay.

 

[jaapv]

A filter does change the optical path, both on entry and on exit of the filter. The efect becomes more pronounced as the incidence angle gets lager. On the front of the lens the object distance is vast, from one metre to infinity, so the effect is minimal. At the back of the lens the register distance is less than 30 mm, so the effect is more pronounced. For instance Leica R lenses, designed with the filter at the back, like the apo telyt 280/4.0 for instance, may not be operated without a filter in the slit. They have been designed with this effect in mind and the back filter has been calculated in as an extra planparallel lens element.
Having said that, my Tegea takes a filter in the middle of the lens, next to the aperture and there the negative effects are nil. With IR filtering, no cyan vignetting - and the focal length is 9.8 mm!😱

 

[oftheherd]

I remember a Sigma 600mm cat I used to have required a clear filter in place (at the rear of the lens) if one of the colored filters wasn't being used. I guess because of what you noticed and what others have said about the filter, especially at the back of the lens, alering the light path noticibly.

 

[mattmills]

riley- http://www.edmundoptics.com/onlinecatalog/displayproduct.cfm?productID=1328this is the filter I'm using. The reason I wanted to do the filtering behind the lens is that Ir filters range from 100 to 200 bucks a pop, and on the outside, they're easily damaged. Filtering on the inside means that I don't have to worry about this, but it looks like it won't work, for reasons outlined above. Unless I can find another 1/2mm thick filter somewhere. But would the effect be different for every lens? Anyway , thanks for the responses, everyone.

 

[Xmas]

A filter is optical glass and may be about 1.5-1.6 ref index, it will have the same effect independent of focal length of lens.

The example is the blad polariod backs which have an optical plate to allow the focal plane of the film to be further back. These work for any lens.

If you use a lens hood it should protect the filter to a degree.

Noel

 

[Bryce]

It is all true about the 'optical thickness' of the filter, and that is one of the two reasons that filters are generally not used behind the lens.
In large format photography, you can get away with this because you're focusing on the ground glass and any change in back focus will be compensated. For this to work, focusing must happen with the filter in place.
The other reason is that any imperfection in the optical neutrality of the filter will result in poor definition at the film/ sensor plane.
Unfortunately, as also mentioned above, the difference in focus distance will vary with the particular lens in use.
So to make the scheme work, you'd have to alter the lens mount to move the optics forward by the effective optical thickness of the filter without disturbing the rangefinder cam... a major alteration.
As an aside, I bought a Russian made 20mm lens for my k-mount Pentaxes that uses specially made filters behind the lens. Each filter is actually a lens element, not flat on both sides, and threads into place. The lens doesn't perform well without one in place.