Very quick question - IR 1um+ filter recommendation

[lukaszgryglicki]

Hi, just a very quick (hopefully) question:

- Can you reccomend some IR longpass filter (not chinease no-name which can be anything) passing above 1um and blocking below?

Why? I realized I'm trying to get as low as possible with UV, even attempting UV-C, but never tried what can I do at the other end... I have Hoya R72 and some no-name IR850 one... nothing above 850, is there any reccommended, checdked & confirmed filter to pass only above 1 micron? Ideally 52mm size.

[lonesome_dave]

I would search ebay for used filters with trusted names.

 

One I bought recently still for sale:

Thorlabs 25mm FELH1250 Premium Longpass Hard-coated Filter $50

 

It has a cut-on wavelength of 1250nm. Not as large a diameter as you wanted but check around and you may find a 50mm version.

You may need to also get a filter holder for 25mm or 50mm to fit your desired lens filter size. Another $50 from Edmund if you can't find it on ebay.  You could also just try to mount the 25mm on the rear of your lens.

 

The deepest filter I've seen with standard photo filter thread sizes is 1000nm, all from China.

 

Andy has done a lot of work in deep IR, check his older posts for more info.

[Andy Perrin]

Yes, I use (and recommend) the Thorlabs premium line. But you won’t be happy until you get an ingaas sensor really. Also, stop trying to find 52mm. Needless expense and the standard for this kind of work is 25mm. 
 

If you get an InGaAs camera you can take photos easily rather than pushing a silicon sensor to tears. 

[lukaszgryglicki]

Hmmm then this is also not that easy. I don't have InGaAs camera and don't plan to buy one. I just wanted to see what I can record at the end of Si capabilities. Also I really wanted 52mm (I have holder from Edmund) - I will check Edmund for IR filters, but I already guess they will cost a lot, how about Hoya RM90 - anybody used that?

[photoni]

@lukaszgryglicki you could try a cheap 1000 LINK or Edmund LINK

or ... just for test there are many chinese filters of 950 nm diameter 52 at 20€

[ulf]

My experience with Chinese IR-pass filters is rather positive.

It might be worth a try to get and stack two or even three of the 950nm variants or a mix of 850 and 950nm, especially if you insist with 52mm threaded filters. 

Up there seams to be a rather uniform world besides the blackening of water.

 

 

 

[Lou Jost]

9 hours ago, lukaszgryglicki said:

Hmmm then this is also not that easy. I don't have InGaAs camera and don't plan to buy one. I just wanted to see what I can record at the end of Si capabilities. Also I really wanted 52mm (I have holder from Edmund) - I will check Edmund for IR filters, but I already guess they will cost a lot, how about Hoya RM90 - anybody used that?

I have the Hoya RM90, available in large sizes. It works well but, as with deep UV, typical lenses corrected for visible light will lose sharpness when used far from their design parameters. When I get home (I am traveling for a week or so) I can post some test shots for you if no one else does so first.

 

Here someone compares it to the 720nm filter: https://www.markcassino.com/b2evolution/index.php/digital_infrared_hoya_r72_vs_rm90?blog=2

[lukaszgryglicki]

OK, from what I read about UV-Nikkor - it is also good for IR up to 900nm (I mean it is designed to 900nm, in contrast to normal lenses which are disigned for visual - meaning ~700nm), so maybe UV-Nikkor will also be good with RM90?

 

[Lou Jost]

Yes, that makes sense, though you are still outside the specs of the UV-Nikkor, you are  alot closer with that lens than with a normal lens.

 

One advantage of the RM90 is that it will act as a narrowband filter, since normal sensors don't go much above 900nm. So your pictures could be sharper than those using the R72. On the other hand, longer wavelengths reach diffraction limits sooner than shorter wavelengths. I have no idea how these two competing effects will balance out.

 

I do recall that at least some of my normal lenses did make nice images with the RM90.

[ulf]

The Hoya RM90 is not sharp cut off to shorter wavelengths and combined with the lower sensitivity of the sensor's sensitivity the actual response in the image seams to be worse than a sharper cut 850nm filter.

 

The images in the link from Lou above with the RM90 are not monochrome!

Many lenses corrected for visual light are still reasonably sharp in IR.

 

I tried the Zuiko 135/3.5 and 100/2.8 and they are very sharp with a 850nm long pass filter.

 

[Lou Jost]

3 minutes ago, ulf said:

The Hoya RM90 is not sharp cut off to shorter wavelengths and combined with the lower sensitivity of the sensor's sensitivity the actual response in the image seams to be worse than a sharper cut 850nm filter.

 

The images in the link from Lou above with the RM90 are not monochrome!

Many lenses corrected for visual light are still reasonably sharp in IR.

 

Those are good points. However, in my experience it is very hard to find a wide-angle lens that works well on a FF sensor in IR. Longer lenses do tend to work fine.

 

The images in my link are not monochrome but certainly show the kinds of differences you get. Blacker skies,  lighter foliage.

[ulf]

The Sunex circular fish eye is very sharp  with 850nm filters. double click for a 100% view

[ulf]

6 minutes ago, Lou Jost said:

The images in my link are not monochrome but certainly show the kinds of differences you get. Blacker skies,  lighter foliage.

But with a sharper cut 850nm filter the images you get are very close to monochrome, indicating that you get less components in the image from the shorter wavelengths.

It is all about the combination of bayer response, sensor sensitivity and filter slopes.

[ulf]

My Irix 15/2.4 with a rear mounted 2mm thick 850nm filter was also quite sharp.

That is rather wide wide-angle lens.

I have no idea how it behaves with filters cutting at shorter wavelengths, but the combination of the 850nm filter and the reduced sensitivity of the sensor for longer wavelengths give a rather narrow wavelengths band, working well.

[photoni]

I'm intrigued, on FLICKR there are many images with 950 nm filter in the specialized groups.

I don't see big differences between 850 and 950 nm and the few photos taken with 1000 nm.

[ulf]

It is because at that upper range there is not much that changes because all three colour channels R, G and B have the same response, hence the monochrome images.

 

One interesting thing to catch in this upper NIR range is the darkening of water.

Andy showed  that with his special camera and also with a FS-modified camera using a narrow band BP-filter.

The absorption peak within range of a silicon sensor at around 970nm is not as strong as the one seen with Andy's special camera.

 

To get a good effect of this there cannot be much contribution of light outside the absorption peak.

 

Later member Fandyous found usable cheap small BP filters with the peak close enough, that also can do the trick if the lens opening is small enough:

https://www.ebay.com/sch/i.html?_from=R40&_trksid=p2047675.m570.l1313&_nkw=980nm+bandpass&_sacat=0

 

However these cheap filters are rather leaky in VIS and has to be used together with a 850nm long pass filter.

I too bought those cheap BP filters and rear mounted them on my 35mm Noflexar lens. I can confirm that they work.

 

 

[lukaszgryglicki]

The best from my GF lenses for IR (with 720nm filters) seem to be GF 250/4 and 45/2.8, 20-35 seems to be quite bad, 50/3.5 has hotspots at f=8 and more, other lenses also have hot spots at least from f=11, only GF 250/4 seems to hav eno hot spots and 45/2.8 is very good but still have some hot spots at f=16.

Generally right now, 45/2.8 wins - it is quite fast (for MF), light and works really well up to f=11 even maybe 16.

[dabateman]

2 hours ago, lukaszgryglicki said:

The best from my GF lenses for IR (with 720nm filters) seem to be GF 250/4 and 45/2.8, 20-35 seems to be quite bad, 50/3.5 has hotspots at f=8 and more, other lenses also have hot spots at least from f=11, only GF 250/4 seems to hav eno hot spots and 45/2.8 is very good but still have some hot spots at f=16.

Generally right now, 45/2.8 wins - it is quite fast (for MF), light and works really well up to f=11 even maybe 16.

That seems to match Jims findings: 

https://blog.kasson.com/gfx-50s/fujifilm-g-lenses-hotspotting-at-720-nm/

 

His 110mm and 80mm were also good.

[Lou Jost]

Note that there are many meanings to the term "good lens" in IR. Most internet sites focus on the hotspot issue and call a lens "good" if there is no hotspot. But this is a very low bar to cross. Image quality in the IR has to be considered, not just presence or absence of a hotspot. Even more demanding criteria are needed when we talk about good lenses for full-spectrum (vis + IR). For that we have to demand that visible and IR light are focused in the same plane. and that requires exceptional lenses such as the Hasselblad Superachromats and Leitz Apo-Telyts. Though you can stop down regular lenses to about f/11-22 and you will get approximately overlapping planes at the cost of diffraction blur.

[Lou Jost]

7 hours ago, ulf said:

My Irix 15/2.4 with a rear mounted 2mm thick 850nm filter was also quite sharp.

That is rather wide wide-angle lens.

I have no idea how it behaves with filters cutting at shorter wavelengths, but the combination of the 850nm filter and the reduced sensitivity of the sensor for longer wavelengths give a rather narrow wavelengths band, working well.

That's really good to know!

[Lou Jost]

Ah, the Irex requires a 95mm front filter or a tiny one in back. I think this is not suitable for those of us who already have a large investment in normal (49-82mm) filters.

[ulf]

4 hours ago, Lou Jost said:

Ah, the Irex requires a 95mm front filter or a tiny one in back. I think this is not suitable for those of us who already have a large investment in normal (49-82mm) filters.

 

Well I tried to be very specific about my findings. The Irix 15/2.4 worked well for me with a rear putty mounted 850nm 2mm thick filter on a Sony A7III FS modified by removing internal filters without replacing them with an internal window.

The result might differ with other types of FS-modifications with an added compensation window replacing the internal filters.

 

I define worked well as good sharpness all over when stopped down enough to be in focus at infinity as well as close by when using the lens standing up.

Then I had no problem with hotspots.

 

It is not immune to strong light that are not shaded with it's lens hood.

That I think is true for most lenses.

 

When placing a filter as thick as 2mm at the rear of a WA-lens the focus scale gets really off but that is not a problem as you have to focus from the image seen by the sensor anyhow. 

 

It is completely unknown if other filters with shorter cutoff wavelength also works well. 

 

To make the rear filter fit I had to remove the filter glass itself from the mounting ring and use the glass round itself.

I think the filter was a 43mm threaded filter. TBD if someone else want to do the same thing.

 

The filter I used was a cheap Chinese IR-pass filter.

Adding a set of a few of those for this lens only is not expensive. 

 

If it works well with other cutoffs I think three or four would be enough.

1. 850nm  for deep NIR monochrome style. OK
2. 720nm or 680nm for coloured NIR style. OK?
3. 590nm for Goldie NIR style. OK?
4. 550nm for Aerochrome style. OK?

 

[lukaszgryglicki]

So, I general, GF lenses suffer in IR:

- hotspots.

- flares, a lot of them.

- generally lower sharpness, especially away from image center.

[Lou Jost]

Ulf, I don't want t9 give you the wrong impression, I appreciated the recommendation very much. It's just that I use a very large number of filters for my asto/IR/UV work (R, G, B, Ha, L, IR, UV, IRChrome) and it would be very expensive to buy a completely new set for a single lens.  And I'd have to remove the lens for each filter change. Wouldn't work well for me. But a very nice lens.

 

[ulf]

58 minutes ago, Lou Jost said:

Ulf, I don't want t9 give you the wrong impression, I appreciated the recommendation very much. It's just that I use a very large number of filters for my asto/IR/UV work (R, G, B, Ha, L, IR, UV, IRChrome) and it would be very expensive to buy a completely new set for a single lens.  And I'd have to remove the lens for each filter change. Wouldn't work well for me. But a very nice lens.

 

No problem, no offence taken. 

 

I just wanted to clarify what I know that so far, that is working well with that lens, and give a possible idea of a cheap way of using it with a limited number of other IR-filters.

 

Do you really need to get a completely new set for such a lens?

 

I do not like to remove a lens for filter changes either, especially as there are increased risk of getting dust on the sensor.

However, the rear mounting method worked well for me with the Sunex Fisheye and my EF 8-15mm FE where that was the only way to add filters.  

Really wide lenses cannot be used with front filters anyhow, or only with crazy big ones, like the 15mm Irix.

For me that is not an option at all, but I still want to be able to use that lens for IR.

 

I too have invested in very many different filters that I cannot use with the Irix lens. 

My standard filter dimensions are 52mm and 77mm and I too have very many Xume-rings for many of them.

It is a shame that Manfrotto seams to phase out the Xume rings now.

 

It is interesting to try the same motif with many filters and in my macro setup it was not unusual that I went through 10 or 20 of them.

After a while though, a few of the filter types have become more essential and used more. 

My idea of what filter set I need for the Irix is much more limited than the full set, just as I described above. 

 

I think there will be an accentuated vignetting if a big front filter is used due to the longer optical light path for the beams in the corners than in the center of the image