KMZ Horizon S3 Pro: A Case of a Wrong-Way IR Focus Shift?

Most lenses optimized for ordinary visible light start to experience focus shifts outside their design wavelength range, as the correction for chromatic aberration starts to break down. In the vast majority of cases, this shifts the focus toward the camera for UV and away from it for IR. But for certain lens designs it is not impossible for a wrong-way shift to occur (e.g. focus moving outward in UV or inward with IR,) although this is relatively uncommon. Recently, I decided to try to attempt IR photography with the Horizon (having already explored UV,) and so rigged a Hoya R-72 filter to the camera with the aid of some Blu-tack putty and an existing filter and shot a roll of Gevaert/Rollei IR400. This camera is fixed-focus, and the only control over focusing behavior one has is the aperture. It also has taken photos of reasonable sharpness with ordinary color film (see here for an example.) The camera shows some slight focus shift in the expected direction in the UV (see here for an example of that.) So I was expecting that the focus would shift away from the camera, making nearby objects more blurry, but leaving distant material untouched. This is not what happened. Instead, I got this (shot at f/5.6, one of the worst examples.)


Not only is this out of focus, the distant material seems to be the most out of of focus. Nearby subject matter (red arrow) seems less severely affected (needless to say, I was somewhat shocked and disgusted with this result.) There are others who have done IR with these cameras, but I have heard no discussion of this issue before. The film/filter combination would probably result in a peak sensitivity near 730-760 nanometers, which is not as far outside the visible range as one might have gotten with older IR films, so this degree of shift is a major surprise. In the future, as there is no focus adjustment, it would be wise to stick to f/16 and longer exposure times when attempting this.

I purposely use smaller apertures [higher f/numbers] to increase the depth of field and that has helped keep things in focus when using IR film with R23, R25, R29 and 720 filters and Rollei IR 400 film.

I've never experienced focus shift with this film (strongest filter used is a 720). It barely goes into the IR to worry about it. I only dealt with it when I was using HIE or Efke with an opaque IR filter.

That's weird. I can't really think of any strong reason for that to happen, and a few reasons why it shouldn't (rules of thumb you already know about focus shift). The Horizon lens is just a standard former-Soviet Industar/Tessar design, I think, nothing unusual. Tessars are likely not the most apochromatic design in the world, but also not likely to be a very special outlier.

I spent a little time looking for a Tessar design lens with an IR focus mark, and don't seem to have one myself, but there is this GN-Nikkor: http://www.photosynthesis.co.nz/nikon/p4528.jpg whose IR mark is in the usual sense, the lens has to be focused slightly further out, ie focal length increases at IR wavelengths. Same for a Retina-Xenar. Of course the Nikon and Xenar are different realizations of "Tessar" than the Horizon, but it suggests the design itself does not violate the usual rules.

It may be worth considering whether there was some issue with the filter (? not sure how unless it was curved by the mounting) or the film loading that caused a focus shift. Do you have the little snap in filters that came with the Horizon? I don't think there was a R25-like red filter, unfortunately, but if there were, you could try that.

All Tessar-type lenses of same aperture have practically same focus shift.

And no Tessar-type lens will show negative focus shift. (Though within our range of lenses there is one design that does.)

AgX said:

All Tessar-type lenses of same aperture have practically same focus shift.

And no Tessar-type lens will show negative focus shift. (Though within our range of lenses there is one design that does.)

Click to expand...

Now I'm curious, what lens design exhibits negative focus shift?

All optical glasses have a dispersion that decreases with wavelength, so any single element's focal length is longer at longer wavelengths (which of course causes chromatic aberration). The basic achromat corrects for this approximately by pairing a convex crown glass with strong power and low-ish dispersion with a concave flint of weaker power and higher dispersion; the net achromat has positive focal length and a positive focus shift to infrared (but much weaker than the individual elements).

That's basic stuff that you already know, but since many complex lenses are essentially elaborations of combinations of achromats, typical lens designs have positive focus shift and I'm curious what arrangement gives the opposite.

Angulon

Before I found this data from Schneider themselves, I too thought IR focus shift always needed for compensation added extention, not less extention.

reddesert said:

It may be worth considering whether there was some issue with the filter (? not sure how unless it was curved by the mounting) or the film loading that caused a focus shift. Do you have the little snap in filters that came with the Horizon? I don't think there was a R25-like red filter, unfortunately, but if there were, you could try that.

Click to expand...

The filter is a glass disk, purchased from Edmund. For the purpose of this exercise I stacked it atop one of the existing factory-supplied filters (the yellow-green one, as that did not seem to block too much IR when I tested it with a digital camera.) I secured it with Blu-Tack putty--not a very elegant solution, but it seemed to hold it in place OK without too much leakage. I then inserted the whole thing into the usual filter position and it seated normally.

Filters in front of a lens do not normally have much effect on focus if they are simple optical flats. Filters behind a lens are a different matter, but that does not apply here.

An R25 might have been some use with HIE, but it would be useless with any modern IR film. Even with the R72, the Wood effect is not as clean as one would like with the film I used. Its IR window is very narrow.

As far as I can tell, the film was loaded correctly.

Snowfire said:

The filter is a glass disk, purchased from Edmund. For the purpose of this exercise I stacked it atop one of the existing factory-supplied filters (the yellow-green one, as that did not seem to block too much IR when I tested it with a digital camera.) I secured it with Blu-Tack putty--not a very elegant solution, but it seemed to hold it in place OK without too much leakage. I then inserted the whole thing into the usual filter position and it seated normally.

Filters in front of a lens do not normally have much effect on focus if they are simple optical flats. Filters behind a lens are a different matter, but that does not apply here.

An R25 might have been some use with HIE, but it would be useless with any modern IR film. Even with the R72, the Wood effect is not as clean as one would like with the film I used. Its IR window is very narrow.

As far as I can tell, the film was loaded correctly.

Click to expand...

All these filters work with the modern IR films included Rollei IR 400 in increasing order of effectiveness: Orange, R23, R25. R29, R72.

AgX said:

Angulon

Before I found this data from Schneider themselves, I too thought IR focus shift always needed for compensation added extention, not less extention.

Click to expand...

Ah ha! Thank you. Angulons, the classic weirdo. The Angulon is sometimes called a "reverse Dagor" design - the usual Dagor is roughly symmetrical and each cell is a triplet of +,-,+ elements. The Angulon is also roughly symmetrical and each cell is a triplet of -,+,-. Both types were shown in the original Dagor patent at https://patents.google.com/patent/US528155A/en?oq=us+528155

The reverse aspect by itself doesn't prove the focus shift one way or the other - we would need to know the powers and dispersions of the elements to know that. It just makes it plausible that the Angulon would be a little different.

Another piece of lore that I've read is that Schneider quit making the Angulon because it used a type of glass that they couldn't get any more. Although that may just have meant there wasn't enough commercial demand for it at the time (probably 1960s or 70s?) to redesign the Angulon.

Snowfire said:

The filter is a glass disk, purchased from Edmund. For the purpose of this exercise I stacked it atop one of the existing factory-supplied filters (the yellow-green one, as that did not seem to block too much IR when I tested it with a digital camera.) I secured it with Blu-Tack putty--not a very elegant solution, but it seemed to hold it in place OK without too much leakage. I then inserted the whole thing into the usual filter position and it seated normally.

Filters in front of a lens do not normally have much effect on focus if they are simple optical flats. Filters behind a lens are a different matter, but that does not apply here.

An R25 might have been some use with HIE, but it would be useless with any modern IR film. Even with the R72, the Wood effect is not as clean as one would like with the film I used. Its IR window is very narrow.

As far as I can tell, the film was loaded correctly.

Click to expand...

I agree that flat glass in front shouldn't affect focus, especially if aligned with the lens axis. A flat but tilted glass in front in principle causes some astigmatism. Ordinarily I would guess that is a negligible effect for distant objects, but because swing lens cameras image each point through several different parts of the lens, I don't have an intuition for when small effects might become significant.

My intention in suggesting an R25 filter was not to get the Wood effect but to isolate the focus shift hypothesis from other oddities. If the lens has a focus shift with the R72 and Rollei IR400, which should have a relatively limited bandpass, the R25 should give a moderately lesser focus shift of the same sign.

reddesert said:

Another piece of lore that I've read is that Schneider quit making the Angulon because it used a type of glass that they couldn't get any more. Although that may just have meant there wasn't enough commercial demand for it at the time (probably 1960s or 70s?) to redesign the Angulon.

Click to expand...

Well, Schneider then made the Super-Angulon.

I haven't shot much Rollei IR 400, but I have shot some, and no focus adjustment was used. However, looking at the good example I've got posted, I shot at f/22 for 1/2 second (4x5).

IR films, like Rollei, and SFX, do not require any focus compensation, in my experience. They just don't reach far enough into the infrared to bother. I've shot both films at small apertures and wide open, with 72R filter.

reddesert said:

I agree that flat glass in front shouldn't affect focus, especially if aligned with the lens axis. A flat but tilted glass in front in principle causes some astigmatism. Ordinarily I would guess that is a negligible effect for distant objects, but because swing lens cameras image each point through several different parts of the lens, I don't have an intuition for when small effects might become significant.

My intention in suggesting an R25 filter was not to get the Wood effect but to isolate the focus shift hypothesis from other oddities. If the lens has a focus shift with the R72 and Rollei IR400, which should have a relatively limited bandpass, the R25 should give a moderately lesser focus shift of the same sign.

Click to expand...

Ok, now I see what your point is and that is well taken. To add to the discussion, remember that with shorter focal lengths the shift should be less.

i discussed this endlessly in a lot of forums one can go directly to rotatingpanoramiccameras.net. 1. IRs are sharp. 2. unsharpness in close region below or around 5m depending fstop is due to fixed focus which need altering lens-film-position or closeup lens.
when i have time i will also show again(testshots gone at pbase only text can be googled,. why the h... did they remove me? )the difference between sharp old horizon 202 lens and unsharp S3Pro(and other series) 28mm lens when stopped down to f11/16 and more. diffraction leading to unsharpness.
NB: I currently have found out that filmlenses made for ir-sensitive film cannot reach infinity without changing register(shorter adapter) or lens being adjusted on Digital-camera. some adapter are made shorter(Novoflex), K+F confort P67-adapter are not. not sharp at infinity. thats not 400m or 2.9km but 8 or 10km. zeiss, novoflex, leica all confirmed. all new zeiss lenses are corrected for this discrepancy.
Simplest is adding dymotape to filmrails. next is me improving this in a new S3Pro. filling the gap below the tape and shaving the ends so film cannot slip under the tape. NB: S3Pro in reality needs a transplanted 202-lens. to solve the diffraction problem. NB: forget there DOF table. also older one. they are not true. one can also change distance of lens to film. only experts can do that right. mine is Adira Kamera GmbH Baselwho also saved me a lot with widelux 1500 problems(see my forum at delphiforum) and roundshot-repair. and all horizont/Horizons. He found out why W1500 was never working properly.me finding out lens decentered!!!multiple cooks at that camera.

Europanorama, so, are you saying that lenses on 202 and S3 are totally different? I had issue on one of my S3 that front lens element was loose. Maybe just ¼ or turn. Now is much better. Replacing the lens and adjusting it's position is very serious job. Just to clarify, did you do it yourself or Adira Kamera GmbH in Basel did it for you?
I took you advise from years ago on Delphiforum for Dymo tape. It works very well. Thank you very much! It shifted focus in right direction and a just tad too close. It also helped me with major problem, roughness of film rails and bad interaction with different film emulsions. Now I switched to UHMW tape. Dymo tape was a much more durable and tougher, but I lost infinity focus. You can not win on both fronts. All the best to you!