Advices and opinions for choosing a deeper IR filter

[FL_96]

Hello,

 

Firstly I apologize for being not a very active member here (photography outside the visible range has been a bit on hiatus)...

 

I've created this topic because I currently struggle about making a choice for a deeper IR Longpass filter. 

It is also fueled by the recent availability of more cut-off wavelength choices from Kolari Vision with addition of KV780, KV930 and KV1000nm filters.

 

I already own a B+W 093, but the diameter seems too small to cover most options, and it is also leaking some visible reds (can be an issue for long exposures).

 

By looking at the transmittance curves on their site and the forum, the Kolari 850 seems very similar, almost identical to Schott RG850 glass. 

Their 1000nm filter is also tempting because I believe some more pronounced effects can be achieved (like even more darker skies with unique contrasts and water starting to darken). 

However the transmittance curve is currently not available, with risk of leakage (unlike RG1000 glass), and I fear that the filter's lesser flexibility regarding exposure times will counteract the gain from more pronounced effects.

 

I am also tempted to cut the pear in half by choosing their 930nm, but again the transmittance curve is not available,and there is a possibility that the filter could be a bit leaky in the visible such as the Hoya RM90 or the B+W093(RG830). 

 

So far hesitation comes between :

- RG850/KV850, the former being difficult to find in my country

- RG1000 (maybe KV1000 ?)

- secondarily KV930

 

What are your experience with filters having similar cut-off wavelengths ? Which one of these would you recommend ?

 

Best Regards, 

 

 

 

 

 

  

[Andy Perrin]

Honestly the 1000nm filter (at least the Schott variant) cuts on verrry slowly, so it's more like "1000nm" in name only. I am not sure about Kolari's. My feeling is that you don't get much from going past 850nm with a longpass filter really, except less light. Any extra contrast can be made up for with adjusting the RAW in an editor.

 

If you want dark water, the best thing is to have a bandpass filter centered at the water absorption peak, which is 976nm. You can find an expensive one on Thorlabs site, or there was a cheap one (that needs additional stacking) from eBay. See Fandyus's thread.

[dabateman]

I have had some great success with the cheap Zomei IR filter sets. You can pick them up really cheap. 

 

I might have first heard of them from ulf or  Enrico. 

[Andrea B.]

IMHO, the best contrast IR photos are around 830 - 850 nm.  At 930 nm or 1000 nm, exposures are longer, and I've always thought the photos from my 1000 nm are a little too dark.

 

There is some reddish-pink false color with the B+W 093 (830 nm), but it goes away with a click of the white balance tool. I don't think this reddish-pink is visible red leak though.

 

[KaJashey]

20 hours ago, dabateman said:

I have had some great success with the cheap Zomei IR filter sets. You can pick them up really cheap. 

 

I've liked cheap filter sets too. Mine aren't made anymore so i guess Zomei it is.

[FL_96]

Hello, 

 

First thank you for your detailed replies, it helps a lot to elapse the decision.

 

So far it confirms what I was suspecting regarding the lack of versatility from a 1000nm filter (and also a bit lower image quality ?).

 

I think now a 850nm filter seems to be the best option. 

 

@Andrea B. The B+W 093 has a small but present visible red leak, when comparing it visually with a Hoya R72 by shining visible red leds through, they are more visible in the former. It needs additional filtration when used for long exposures. From what I've seen online, the RG850 for example does not seem to suffer from this issue. However, the pink-reddish cast seen with modified camera is definitely not related to the leakage :)

 

Best Regards,

 

[Andy Perrin]

The problem is that your test is flawed. You should never test a filter by shining LEDs through it — they are VERY intense and will punch through even high OD blocking that would be fine for photography. 
 

The other flaw is more about nomenclature. Usually on this board we consider “infrared” to start after 700nm. However the human eye can perceive significantly further than that at high intensity, well into the high 700nm’s. So what you are seeing may actually be mostly infrared according to our usual cutoff. 

[Andrea B.]

How can I test this for FL-96?

Perhaps stack an S8612 IR-blocker with the B+W 93 to see if any red gets thru?

 

I can get to this test tomorrow, so stay tuned!! 

 

[FL_96]

Hello, 

 

@Andy PerrinYes, you are right, the test is flawed and very subjective too. The led used are centered around 660nm. They are very bright, to the point of needing a strong lampshade.

 

For the red leak, I've done tests a few years ago between the 093 alone and stacked with the R72, on an unconverted D5100. The red leakage was corrected with the R72, but I realize now that the latter can have a similar transmittance in the deep red as well. 

 

@Andrea B. Thank you for your proposition, it can be very interesting to compare/cross the results, here is a group of images from a few years ago like mentioned above:

 

EDIT: added shot with R72 filter only, ISO 200, F/8, 15.0s of exposure

 

With B+W 093 only, ISO 200, F/8, 36.4s of exposure

 

With R72 stacked with B+W 093, ISO 200, F/9(if at F/8 it would be between 50 and 60s), 67.1s of exposure :

 

Best Regards,

 

[Andrea B.]

Brief review:

 

Schneider-Kreuznach says their 093 passes only IR light. Granted, a manufacturer can say anything, but S-K is very reputable.

https://schneiderkreuznach.com/en/photo-optics/b-w-filters/filtertypes/infrared

 

Here is a measurement of the B+W 093 filter by our member Shane Elen. http://www.beyondvisible.com/BV3-filter.html

No visible light.

 

Also note that the substrate of the B+W 093 is Schott glass.

So the only difference between the B+W 093 and its Schott RG830 equivalent is the coating on the B+W 093 which helps prevent flare.

 

But I will run that test tomorrow.

---

 

FL-96:  Just for the record, a well-known maxim from our founder Birna is that

almost any light can be forced through any filter if you make a long enough exposure.  

 

But to make sense of your presented photos, it would be nice to know what camera was used and what converter was used?

Also the thickness of the filters plays a role in all this.

 

A very important question is why the B+W 093 foto needed a 36 second exposure at f/8 and ISO-200. I would expect something much, much shorter!!

[FL_96]

Hello, 

 

The camera used was an unconverted Nikon D5100. The lens was a 18-55mm F/3.5-5.6 VR DX kit lens. This explains the long exposure time. It has not been observed with a converted camera. I do not know the thickness of the filter S-K use.

 

I may have been biased by this first use of the filter with an unconverted camera. Sorry for that.

 

Best Regards, 

[Andrea B.]

oh, no problem. Let's just see what a test shows. It's always a valid concern as to whether any of our filters has some leaks or not. Many do, but it's not enough to worry about. (I don't think the completely perfect filter has ever been made!)

 

I was able to get out and make a few test photos, so I might have something soon. My light was a bit variable from some moving clouds, so I'm not yet sure if I'm happy with the test photos.

[OlDoinyo]

On 3/4/2023 at 5:01 AM, Andrea B. said:

There is some reddish-pink false color with the B+W 093 (830 nm), but it goes away with a click of the white balance tool. I don't think this reddish-pink is visible red leak though.

 

That color is just differential sensitivities of the three channels in that general range, but the curves are all flat out there. There is no chromaticity associated with this tint.

[ulf]

With an unconverted camera you force the IR light through the camera's IR blocking filter.

The blocking of visual and shorter IR wavelengths done by an external IR-filter in combination with the internal filter is very difficult to predict.

The known behaviour of the RGB-sensor in the IR range in combination with the filter transmission characteristics do not apply in this situation.

The colours you get with the B+W 093 are unique for your camera model with it's weak IR blocking.

[Andrea B.]

As mentioned above, I did get outside to make a few IR photos this afternoon. But after looking through the photos, I think the light was too variable to make satisfactory test photos. Look at the shadows in the hills in the 4 attached photos. Those clouds were moving. And the sunlight (at my back left) was brightening/darkening from another set of moving clouds. 

 

However, just for my own amusement, I looked at the raw composites to get a feel for how the IR light was being recorded. I got the R, G and B average values from Raw Digger and scaled them to calculate an average color for each photo. I put a dot of that color on each photo. You can see that the RG 780 nm filter recorded red values, on average. The B+W 093 recorded almost cerise values ("cherry red") on average. Both the RG 850 and the RG 1000 recorded values closer to magenta, on average.

 

GEAR:  Nikon D610 + UV-Nikkor + Sunlight

All photos made at f/8 and ISO-200.

 

Please don't place too much faith in the following exposure times. While they do trend longer as the filter wavelengths get longer, I cannot be sure that the relative differences between filters would hold under stable lighting.

 

Focus was placed mid-range on the juniper shrubs just in front of the center shadowed hill in the first photo. It was not full infinity focus at that point.

 

Schott RG780, 2mm thick.

1/200 sec

 

 

B+W 093

1/125 sec

Cerise is at 330° on the color wheel.

 

 

Schott RG850

1/125 sec

Magenta is at 300° on the color wheel.

 

 

Schott RG1000

1/20 sec

[Andrea B.]

Here are the white balanced versions. I didn't process for full artistic value. Only black & white points were adjusted and those were very minor changes. Resizing was done in Photo Mechanic. No noise or sharpening tools were used.

 

Schott RG780

 

 

 

B+W 093

I think it is safe enough to conclude that there are not significant differences between the B+W 093 photo here and the Schott RG 850 used for the next photo. I yet haven't managed to find my Schott RG830 filter. Hopefully it will turn up soon so that I can include it for the next set of test photos.

 

 

 

Schott RG850

[Andrea B.]

Here's a weird result from the Hoya RM90. Look at those dark shrubs and grasses! What's going on with this filter? Compare this photo to the following Schott RG1000 photo which shows the usual brightness in shrubs and grasses. I must have done something wrong!! Did I accidentally stack two filters? I gotta try this RM90 again.

 

GEAR:  Nikon D610 + UV-Nikkor + Sunlight

Both  photos made at f/8 and ISO-200.

 

Hoya RM90 anomalous result

1/40 sec

 

 

 

Schott RG1000

1/20 sec

[photoni]

3 hours ago, Andrea B. said:

 

Hoya RM90 anomalous result

1/40 sec

looks like normal 590nm red filter, leaves are not clare

.... but 1/40" time is weird
did you use an IR filter + BG38?

.
do you have a VIS photo to understand the colors of the hills behind?

[lukaszgryglicki]

Because RM90 has 50% transmission in around 900nm but its curve is very flat/low angle, compared to those 720 ones, just google for RM90 transmission curve to see.

 

[ulf]

2 minutes ago, lukaszgryglicki said:

Because RM90 has 50% transmission in around 900nm but its curve is very flat/low angle, compared to those 720 ones, just google for RM90 transmission curve to see.

 

True.

Also the curve is from a base thickness of 2.5mm.

As always the thickness matters!

If the filter glass is thinner, the slope will be even less steep and more light from near VIS will be transmitted.

[Andrea B.]

Lukas, Ulf:  Thx for the feedback about that RM90.

 

Toni:  Yes I do have a visible. I'll get that posted soon.

[Andrea B.]

I looked at the RM90 transmission chart. But I'm still somewhat surprised. 

I need to make a photo of my standards with the RM90 and the RG1000 for comparison.

 

 

 

 

---

Here is the visible reference. This time of year everything is a bit dry looking.

The juniper shrubs in the foreground are a dark yellow-green. They do not reflect as much IR as do deciduous leaves or most grasses.

 

[Andrea B.]

To return to the question at hand from FL_96:

I want to test (re-test) the IR093, RG830 and RG850 for you. I think those three IR-pass filters all produce the same "look" after using a white balance tool. And those 3 filters have reasonably fast exposure times.

 

I don't have a channel swapper app currently, so I will upload the raw files for everyone to process as they please for different finishes.

 

[Andy Perrin]

Andrea, the dark shrubs are what I get when a cloud is shading them versus direct sun. Given the partly cloudy day you had, that’s my leading hypothesis, not filter shenanigans. 
 

See the difference between the shadow and direct sun on mountains far away in the same pic? I think that’s what’s happening.

[photoni]

2 hours ago, Andy Perrin said:

Andrea, the dark shrubs are what I get when a cloud is shading them versus direct sun. Given the partly cloudy day you had, that’s my leading hypothesis, not filter shenanigans. 
 

See the difference between the shadow and direct sun on mountains far away in the same pic? I think that’s what’s happening.

 

I think @Andy Perrin is right, the distant vegetation (where there is no cloud) has the same clear tone.