Strange infrared behavior

[jk4u59]

Hello all,

I have a question that puzzles me, because I wasn't able to find an explanation: I recently decided to have my previous mirrorless (a Lumix GH1) converted to infrared, to give it a "second life" instead of leaving it collecting dust on a shelf. So, its IRcut filter that avoids the sensor being hit by infrared radiation has been replaced with a 720 nm filter that, on the contrary, allows (almost) only IR to reach the sensor.

I had the converted camera back from the repairing lab just some days ago, and I'm now testing this new configuration, trying all my lenses (I read that some of them might produce hot spots when working in IR). Well, instead of any hot spot, I discovered an unexpected behavior: my (fully manual) Samyang 7.5mm fisheye does not focus properly! At least, it cannot focus at infinitive, resulting almost useless. So my question is, why? I made several other tests:

• I tried to mount again the fisheye on my current camera (OM-D E-M5 Mk.II): it stil works perfectly (note that, before the IR conversion, I used many times this lens on the GH1 to make several 360° panoramas, without any problem)
• All the AF lenses I have work well also on the IR-converted body
• I mounted the old Zuiko OM 50mm with its M4:3 adapter on the GH1, focusing an object at about 1 m distance, then I passed it on the OM-D without moving at all the focusing helicoid and checked the focus from the same position as before: the object was slightly out of focus!

This did not surprised me, I know that focus changes a bit from IR to visible. Only, I cannot understand why, among all the lenses I tested (Lumix Vario G 14-42mm, Lumix Vario G 45-200mm, Zuiko 9-18mm, Zuiko OM 50mm, then this Samyang 7.5mm), only the fisheye cannot focus to infinitive. Does anyone has some explanation for that?

Please check the (downsized) attachments:

• "Fisheye_7.5" is clearly blurred
• "VarioG_14-42" is fine (I also did some postprocessing on it)

Thanks in advance,
Ivan (Milan, Italy)

 

[fortwodriver]

If you focus very slowly, zoomed in, does it seem to snap in and then out of focus? Many of my lenses are marked such that you have to focus in front of "Infinity" on the markings to make them sharp with IR photography.

The lenses that have IR focus marks have them positioned to the right of the regular focus mark (for lenses that turn to the left to get to infinity,) so you have to turn them back from infinity slightly to line them up with the IR mark.

Also, it's very possible that the infinity mark on that Samyang is not quite accurate.

 

[bargainguy]

The way I understand it, if you have a favorite lens you want to use with your new IR converted camera, you can send it along and have the IR calibration done for that lens. If the lab you sent it to doesn't do this, be aware that some offer that service.

So why would your widest lens have a problem? Just a guess - there's enough phase shift from normal spectrum to IR to draw the image out of focus completely from its (normal spectrum) original design parameters. The problem could also be the thickness of whatever filter they used in the conversion. Too-thick filters result in issues with some wideangles.

Don't forget, virtually every lens made this days (rare exceptions from Zeiss, Coastal Optics etc. $$$$$$) is not calibrated for use with infrared, they're designed for use with normal spectrum. If it doesn't work with IR, isn't the designer's fault. When you go off-label by using lenses outside their design parameters, anything can happen.

FWIW, I have the Panny 7-14/4. With my 470nm IR-converted EM-1.1, it won't autofocus at 7mm at all. 8-14mm no problem, but 7mm, forget it. Not a fault of the lens! Another off-label use. Stuff like this happens with IR.

 

[Roboticspro]

Hi,

I just tried my Bower(Samyang) 7.5mm FE with my IR 720nm converted E-PM1; no issues with focus. I set the lens to infinity and the f-stop to f5.6.

I think it is a filter thickness/back-focus issue, as if your camera works with other lenses and this one does not. It does not take very much to be out of the plane of focus at that F.L. (7.5mm)

Regards,

Edd

Subscribe to see EXIF info for this image (if available)

Subscribe to see EXIF info for this image (if available)

Subscribe to see EXIF info for this image (if available)

 

[jk4u59]

Well,

plenty of stuff! please check the attached picture: my old Zuiko OM has the shifted red mark for IR pictures, while Samyang has not.

I tried to focus using several aperture values, from full opening (f/3.5) down to f/5.6 to f/22: nothing changes, even moving the focusing ring extremely slowly.
Moreover, it seems even worser: the fisheye simply does not focus at all! For shorter distances, it simply approaches the correct focus, but it never reaches an acceptable sharpness.

At this point, I agree that it could be the fisheye design that is not compatible with IR: after all, lenses are designed for visible light, not IR, and 7.5mm is an extreme focal lenght!

By the way: using my Zuiko 9-18mm (AF) it works fine and focuses precisely, but... I noticed that at the lower focal lenght (9mm) it always needs two attempts to reach focus! At the first one, a red square appears on GH1's LCD (meaning it's out of focus): pressing halfway the shutter button a second time, the square turns into green (focus lock) and the picture goes fine. I didn't tested the longest lenght. I'll do soon and report the results.

Thanks for the suggestions!

 

[jk4u59]

Hi,

I just tried my Bower(Samyang) 7.5mm FE with my IR 720nm converted E-PM1; no issues with focus. I set the lens to infinity and the f-stop to f5.6.

I think it is a filter thickness/back-focus issue, as if your camera works with other lenses and this one does not. It does take very much to be out of the plane of focus at that F.L. (7.5mm)

I read that the filter in front of the sensor MUST be of the same thickness of the one replaced, just because this parameter is used by the AF algorithms.
Might be that mine is not EXACTLY the same, but has some slight difference from the original, that is not relevant for all the lenses except for the lowest one (7.5 mm), where it is instead capable to completely spoil the focusing.

 

[PakkyT]

Random thoughts which are not really solution, just my thoughts on some comments.

There is a difference shooting IR on a mirrorless camera vs. on a SLR camera. The old OM lenses that had an IR mark on them was for shooting IR film but still used two light paths so what you would see in the viewfinder after going through the mirror or prism to be in focus may not match what IR light does when going through the lenses so the mark was placed to show the adjustment of what you needed vs. what you were seeing.

Likewise the calibration some IR conversion companies offered was for SLR cameras where again the AF sensor not using the same light path as the image sensor had to be adjusted for the two separate paths.

With mirrorless, most of that goes away because the image sensor is also the AF sensor and therefore they will be the same light used for focusing and imaging.

As to moving a lens focused on an object using an IR camera to a non-IR camera, having the focus be a bit different is not surprising because you are using different wavelengths of light so the focus might be different for each and would need to be adjusted slightly differently for visible light.

As to why you can not achieve infinity focus at 7.5mm I am not sure but replacing one filter (IR cut) with another (720nm) since they may not be the same thickness nor have the same refractive index can result in giving you a closer working distance but not be able to achieve infinity focus. Likely the difference is small enough that most lenses can still work especially if they have the ability to focus past infinity to begin with. But it may be that at 7.5mm and at 720nm+ wavelengths (index of refraction is wavelength dependent), it is simply a little more too extreme of a change to get there anymore. You might try getting a hold of an 8mm fisheye and see if the 0.5mm makes the difference or not or maybe the Oly 9mm BCL.

 

[Mack]

For the OP, are you saying your Samyang 7.5mm is hitting the focus stop prior to it achieving focus? In situations where the subject is closer, does it focus okay?

Aside, I bought a "lightweight" version of the Laowa 7.5mm lens for use on my drone. It refused to achieve focus as the focus ring hit the stop prior to actually focusing at infinity. I returned the lens for another copy and it did the same thing; it would hit the stop prior to achieving focus pointed at infinity objects.

The dpreview.com forum had issues of that Laowa lens not being able to focus well at infinity on some other cameras/drones and their solution was to remove the four screws holding the rear element group and putting a shim under the stack so the group was moved closer to the sensor and gave one a bit more focusing throw. So I cut out a piece of 0.002" brass shim stock and put it under the stack. Of course the focusing ring's scale isn't right in agreeing with the actual focusing distance, but least it works now and is quite sharp still. Another matter is the Laowa isn't a "chipped lens" so the camera gets confused as to what lens is mounted and some features do not work. On the drone, I have to do what is called a "hot-swap" where you put on a normally chipped or OEM lens and then swap it for the Laowa while camera is still running.

 

[jk4u59]

Hi, just to clarify:

• my Samyang 7.5mm, apart from not being "chipped" but only a traditional full manual lens, has always worked perfectly with both my cameras, GH1 (before its conversion to IR) and OM-D E-M5 Mk.II. It still focuses at infinity when mounted on the Olympus (hence in visible light).
• with the converted GH1, actually, it's not just like I wrote (I did some other tests meanwhile): it's not matter of not focusing at infinity: the fisheye is now simply is unable to focus at all! I wrote "unable to focus at infinity" because it's quite evident, but even if I try to focus on near objects they never went sharp (a friend of mine told me that, rather than "unable to focus" he would define it "a general lack of sharpness")
• my feeling, while I rotate the focusing helicoid, is that "it doesn't work": the subject's sharpness, apart from never becoming ecceptable, varies very slowly, for sure very differently with respect to the fisheye's standard usage (I mean in visible light), where the sharpness change rate, moving the helicoid, is like any other lens.

 

[RAH]

I don't know anything about the IR aspect, but I have a Samyang 7.5 m43 FE lens and it hits infinity right at the stopping point. So there is absolutely no leaway, and if you back off a little, it is immediately OOF. So even though it appears that I can achieve infinity focus, I have never liked this aspect of it (I kind of always think that maybe if I could go a little further, I might get even better infinity focus).

There is a way to change the focusing ring on the lens, as described here:

https://www.focusonnewfoundland.com/section488880_174000.html

This is for the 8mm version, so it may not be the same as on the m43 version, but it is worth trying. I should think you could adjust it to allow you to go further beyond the infinity stopping point and get it to work, even with the IR filtering.

Just a suggestion.

 

[PakkyT]

There is a way to change the focusing ring on the lens, as described here:

http://www.focusonnewfoundland.com/section488880.html

I think you meant https://www.focusonnewfoundland.com/section488880_174000.html

 

[RAH]

Fixed! Glad you noticed.

 

[jk4u59]

I don't know anything about the IR aspect, but I have a Samyang 7.5 m43 FE lens and it hits infinity right at the stopping point. So there is absolutely no leaway, and if you back off a little, it is immediately OOF. So even though it appears that I can achieve infinity focus, I have never liked this aspect of it (I kind of always think that maybe if I could go a little further, I might get even better infinity focus).

There is a way to change the focusing ring on the lens, as described here:

https://www.focusonnewfoundland.com/section488880_174000.html

This is for the 8mm version, so it may not be the same as on the m43 version, but it is worth trying. I should think you could adjust it to allow you to go further beyond the infinity stopping point and get it to work, even with the IR filtering.

Just a suggestion.

Hello RAH,

thanks for your suggestions, I think that I'm now much closer to a correct understanding of the "strange behaviour" of my fisheye in IR light!
Here below my comments:

1) Fisheye focusing problem in IR

After reading your post, I set up a small test bed for the Samyang fisheye on converted GH1 body (see picture). I placed 3 small objects at increasingly (small) distances from the camera, on my kitchen's table:

• schoolbus at 10" (25 cm)
• wooden swan at 28" (70 cm)
• grey rabbit at 42" (110 cm)

I set the lens' maximum aperture, so to avoid evaluation errors due to the depth of field effect.

A) Tests performed with the IR converted GH1

1. at first I (manually) focused the schoolbus at 10", managing to get it sharp, then I looked at the distance scale on the lens' helicoid: it was already close to infinity!
2. then I tried in turn with the swan at 28": it wasn't yet really on focus but I'd already reached the end of scale (infinity)
3. eventually, nothing to do with the rabbit: it was always out of focus.

Of course, repeating the tests at f/22 gave an apparently better situation, due to the big depth of field available also beyond the focused point.

B) Tests performed with the NOT converted OM-D E-M5 Mk.II camera

I then repeated the same tests mounting the fisheye on my standard OM-D E-M5 II camera: as expected, I managed to focus all the small objects, and the distance scale on the lens' barrel was more or less the real one.

C) My conclusion

So, my assumption is now this one:

The Samyang 7.5 mm fisheye lens, when in IR, shows an heavily uncalibrated focusing ring, that would require to continue several degrees (let's say almost 90°) beyond the stop at infinity to be able to focus correctly at infinity!​

​

I didn't realize this fact up to now, simply because I didn't suspect such a big difference from focusing in visible light, that is way bigger than the small distance of the red IR focusing mark from infinity that can be seen on lots of the old analogue lenses.


2) Fisheye focusing fixing suggestions

Well, it would be nice, but I fear that modifying the focusing of the fisheye is not possible (or, at least, not simple at all). My lens assembly is not like the description you passed: there's no rubber on the focusing ring (it's made of very solid plastics), and there aren't any screws on it (the only visible screws I found in the whole lens are under its mounting flange: 3 small Phillips screw, plus a tiny slotted screw close to the rear lens.
But, above all, even if they were in the same position described in the fixing procedure, the required recalibration looks so heavy that I fear the fisheye wouldn't work anymore in visible light!

Anyway, thanks a lot for your contribution!
Ivan (Milan, Italy)

 

[RAH]

Yeah, I figured the much smaller m43 version of the lens might be different in build than the APS-C version. Too bad. Maybe you could do a web search on resetting the m43 scale? I agree that if you have to make that big a change (if you could find out how), it might not work on a non-IR camera. It almost seems that you need a special IR lens, just as you have a special IR camera body.

In fact, having said that - one option might be to buy an APS-C version of the lens (maybe they are FF too; just a Canon-mount or whatever) and alter it for your IR camera. You can probably find one cheap and an adapter should also be inexpensive since you don't have AF to think about at all. Glad we figured out the problem!

 

[PakkyT]

BTW I just remembered my own issue with my 8mm fisheye years ago. Shorter version is on my E-M1 the lens was blurry through out the focus range. The lens is a M42 screw mount & unscrewing the adapter from the M42 mount a tiny bit to extend the distance put the lens at the right distance to give me good focus through the focus range. I made a shim ring to put that distance into the mount. It was only 0.96mm. Less than 1mm off and the entire focus range was useless. So it doesn't take much variance in lens to sensor distance for a ultrawide fish eye lens to be out of the focus range. Your lens is even wider and you only are missing on the infinity end, so it would be an even tinier variance in the mount depth coupled with a ever so slight focus distance shift with IR to have this effect.

 

[jk4u59]

Yeah, I figured the much smaller m43 version of the lens might be different in build than the APS-C version. Too bad. Maybe you could do a web search on resetting the m43 scale? I agree that if you have to make that big a change (if you could find out how), it might not work on a non-IR camera. It almost seems that you need a special IR lens, just as you have a special IR camera body.

In fact, having said that - one option might be to buy an APS-C version of the lens (maybe they are FF too; just a Canon-mount or whatever) and alter it for your IR camera. You can probably find one cheap and an adapter should also be inexpensive since you don't have AF to think about at all. Glad we figured out the problem!

Well, all other lenses I have do work fine on both visible and IR lights. I think that, by now, I'll live with the Zuiko 9-18 mm zoom, as an ultra-wide lens. Anyway, I'll also think to that Olympus' small fisheye additional lens (I don't recall the neme, by now) as a cheap alternative.
I'd like to try to do a 360° browsable panorama in IR... I always used the fisheye for this task, just to reduce the number of shots required (16, usually...). I have to figure out how to manage the shots sequence if I use the 9-18mm, instead.... well, next activity found!

 

[jk4u59]

BTW I just remembered my own issue with my 8mm fisheye years ago. Shorter version is on my E-M1 the lens was blurry through out the focus range. The lens is a M42 screw mount & unscrewing the adapter from the M42 mount a tiny bit to extend the distance put the lens at the right distance to give me good focus through the focus range. I made a shim ring to put that distance into the mount. It was only 0.96mm. Less than 1mm off and the entire focus range was useless. So it doesn't take much variance in lens to sensor distance for a ultrawide fish eye lens to be out of the focus range. Your lens is even wider and you only are missing on the infinity end, so it would be an even tinier variance in the mount depth coupled with a ever so slight focus distance shift with IR to have this effect.

Hi,

so, if I understood well, that fisheye is SO sensible to even minimal modifications in its position wrt the sensor that even an ultra thin additional ring in the mount could benefit a lot?
I'll try to do a search about my particular mount, anyway I have a question: provided that such an additional thin ring is available, inserting it in the mount bayonet wouldn't cause any problem? I mean forcing the coupling mechanical parts, that for sure weren't designed to allow any extra thickness, even very little.
...after all, I prefer keeping the fisheye well working in the visible range.

Bye

 

[PakkyT]

that even an ultra thin additional ring in the mount could benefit a lot?

At least for me on my fisheye that was the case. A bit more information and a photo of the lens with the red spacer ring installed can be seen here...
https://www.mu-43.com/threads/adapter-accuracy-or-not.96796/#post-1120488

 

[bargainguy]

I'm not convinced you can bring your 7.5mm back into focus with your IR conversion with modifications.

If it were just a simple matter of IR phase shift, I shouldn't be able to focus my 7-14 Panny at the 7mm end. Although my 470nm EM-1.1 refuses to autofocus at 7mm (8-14 is fine), it'll still manually focus and capture wonderful images there. The 7-14 works fine on all my other bodies at all focal lengths, both IR and normal spectrum.

So I'm thinking your problem is not just phase shift, it's flange distance +/- filter thickness +/- phase shift.

You can mess with the flange distance via spacers, but I wouldn't do this unless you never plan on using this lens on any other camera.

It's a lot of trouble to bring one lens into compliance. Not sure it's worth the effort.

 

[jk4u59]

Perhaps better: following your advice, I searched the net for something like "Samyang MFT fixing" and landed to this site:

https://www.diyphotography.net/a-diy-guide-for-fixing-the-rokinonsamyang-mft-fisheye-focus-issue/

It seems that sometimes this lens (that's exactly same as mine) is not well calibrated, and sometimes focuses infinity too early, sometimes can't reach the focus position because it would fall beyond the stop. It depends on the rear element of the lens, that might be too screwed in or out (probably because it's not properly tightened in its correct position). The site reports the trial-and-error procedure to fix the issue, ans seems not disruptive. Goot to know!