View Full Version : Center centre filter distance

18-Jan-2009, 16:55

I'd like to ask a question about the distance a centre filter should be placed from its lens. I am assuming the filter made by the lens manufacturer.

I am imagining that if I were to stack filters between lens and centre filter it would give a different result to placing the filter on the lens directly.

Also, is the centre filter designed to be placed on the front of the lens only? Can I use the centre filter on the rear lens also?

I have a Rodenstock 90/f 6.8. If I were to use a filter at the rear I would need to invent some sort of push-on adapter to allow the filter to slip onto the rear of the lens. This could mean the filter would almost touch the lens! (less distance in this case)

I suspect in the end there may not be any noticeable difference with regard the distance or front/rear thing.

The reason I ask is because I want to know all the issues before buying. It is a rather expensive bit of kit!



Nathan Potter
18-Jan-2009, 17:32
Ideally center filters are designed to screw in directly to the front element. Additional filters are best placed on outside of the CF. In practice there can be a fair bit of distance variation from the front element without a lot of radial variation in density. With wide angle lenses the rear elements are small diameter compared to the filter diameter so the radial density correction is practically non functional. Don't use them on the rear elements.

Nate Potter, Austin TX.

18-Jan-2009, 18:37
Thanks Nate,

The extra filters would then be used on the rear element leaving only the centre on the front? I have a bunch of 77mm screw-on filters that I am not inclined to replace.

This is a pain for circular polarisers!

Thanks again,

Nathan Potter
18-Jan-2009, 18:53
I don't like to use any filters on the rear element. This is especially true with wide angle lenses where the exiting ray bundles are extremely off axis at the periphery of the image. The filter glass refraction path is longer at the edge of the filter than on axis at the center. The lens was not designed to accommodate such a situation. I admit I've not thought this through thoroughly so I may be over reacting but I'm sure someone here will clarify the issue more thoroughly.

Nate Potter, Austin TX.

Bob Salomon
18-Jan-2009, 19:05
No filter should be used behind the lens unless you are trying to degrade the image or unless the lens requires a filter behind it. Wide angles certainly do not!

David A. Goldfarb
18-Jan-2009, 19:24
Center filters have a larger front thread than the rear thread, so you can put filters on the larger thread usually without worrying about vignetting.

If the only option is to put a filter behind the lens, clean gel or polyester filters are the best alternative, because they are thinner than glass or resin and cause less image degradation. The problem with thicker filters, particularly with wide lenses, is that rays at the edge of the image are going through more glass than the rays going through the center of the image, so the focus shift caused by the filter will be greater in the corners than in the center. A thin filter like a gel causes less focus shift.

Peter K
19-Jan-2009, 00:18
No filter should be used behind the lens unless you are trying to degrade the image or unless the lens requires a filter behind it. Wide angles certainly do not!
A filter changes not only the focal-length of the lens but also over-corrects spherical abberation. And of course it adds two glass-air surfaces.

To minimize this effects the filter should always be mounted at the longer distance. For "normal" work this is the front-thread of a lens. But with macro work the filter should be mounted between lens and film.

So there is only one situation one can use front and rear thread: image scale 1:1. :cool:

Peter K

Emmanuel BIGLER
19-Jan-2009, 06:19
In addition to what has been said (and well said) I could add that the the centre filter is designed so that the filter should be located at a certain distance from the entrance pupil of the lens, in front.

To figure out what is going on, on one has to imagine that the lens is stopped down very much so that the only useful rays are those crossing the optical axis very close to the centre of iris inside the lens, this implies that rays crossing the optical axis (in fact : aiming at) very close to the centre of the entrance pupil on input.

So if you change the filter to another lens or if you stack the filter above another filter, the proper distance may not respected and the correction may not as good as it should be, since the correspondence between the angle of the rays and the position where they cross the filter plane depend on this distance.
Moreover modern wide-angle lens exhibit some amount of pupillar distorsion, so using the filter with another lens is at risk of getting two different reasons of mismatch, first a different distance to the entrance pupil, and second a different law of intensity distribution vs. angles as the result of different pupillar distorsion effects.

However the real problem IMHO is probably mechanical vignetting when stacking two filters one on top of the other with a wide-angle lens for which you demand to use all the available image circle ! If you not use the whole image circle, OK, but then you probably do not need a centre filter ;)
But this ordinary mechanical vignetting can be checked, in principle, by looking through the lens from the corners of the ground glass when the corner are cut or without ground glass at all.

Now for the academic side of the story, and just for the fun of discussing useless theoretical things ;) if we could neglect the additional spherical aberration generated by placing any filter at the rear (in the infinity->focus setting of course), we could argue that the focus shift is compensated by focusing on the ground glass.
Moreover, taking into account that wide-angle lenses are quasi-symmetrical, by placing the graduated filter at the same distance of the exit pupil, the filter would in principle make the same compensation job ; the fundamental reason is that in a quasi-symmetrical lens design, pupils are located very close to the nodal points, it means that the angles for the rays on exit are the same as on entry (by definition of the nodal points of a thick compound lens). And if the lens is symmetrical, well, without any knowledge of the puppillar distorsion effects, I would expect the same effects on both sides if the iris is located at the centre of the lens.
Symmetrical wide angle lenses are what makes lenses in LF so fascinating, people using SLRs do not know what they loose, not being able of mounting those lenses in front of their SLR body ;)

Bob Salomon
19-Jan-2009, 08:33
Let's make this very simple.

You take a picture with a lens. The lens makes the rays hit the image plane with all of the corrections built into the lens. Should you wish to modify the image you may use a filter. When the filter is in front of the lens, screwed into what the lens manufacturer calls the filter threads the light then passes through the filter and the lens can then make sure that the rays pass through the lens are still properly passed onto the film.

Now by putting a filter behind the lens you do a couple of things. First you shift the plane of focus by about 1/3rd the thickness of whatever you put back there. Sometimes this is desired and the manufacturer makes corrrector plates to do just that - Rodenstock Apo Sironar HR lenses for instances used with film. Then the rays, after passing through the lens itself, are then passed through more glass or plastic of a filter that was made to be used in front of the lens and not to become a part of the optical system.
Now that filter may have imperfections that don't matter in front of the lens, it may have fingerprints or smudges that will materially effect the rays passing through the lens, etc.
Then what type of filter? Coated, uncoated or muticoated? If multicoated or coated on one or both sides? Solid optical glass ground and polished parallel or two glass plates held together by colored cement? Molded glass or sheet acrylic?

Filters belong in front of a lens for best optical results. In your case extra filters fit the front thread of the Rodenstock center filter. Slim filter mounts minimize the possibility of mechanical vignetting and the lens manufacturers took mechanical vignetting into account when making their center filters by making the front rim of the filter much larger then the front of the lens. Lastly look at a lens manufactures specification sheet for their lenses. For instance the Rodenstock lens specifications give the following information:
Lens series/focal length and speed, Max. Recommended film size, smallest aperture, push-on mount diameter (lens cap size), filter thread, rear barrel diameter (lens cap size), flange focal length, overall length, weight.

Notice what is not there? REAR FILTER THREAD. That is because the manufacturer of the lens did not design the lens to use a filter behind it. In fact, the 90mm 6.8 Grandagon does not have rear threads exposed at all!

19-Jan-2009, 22:51
I so didn't want to buy any more filters!

Thanks to you all for your efforts.



12-Jan-2011, 19:51
Thank you Bob, you just saved me an experiment with my new 55 Grandagon...

Bob Salomon
13-Jan-2011, 00:20
I goes only in the front. It must be the correct density for the specific lens, other filter must be placed on top of the center filter. The lens must be stopped down at least two stops for the CF to work. No filter should ever go onto the back unless it is part of the optical formula. And a Center filter is not a part of the formula. It only belongs on front of the lens.