IR filters between 720 and 850

[lukaszgryglicki]

Hi, are there any infrared longpass filters sharply passing after say 750 nm (something between the most common 720 and 850 ones)?

Why asking?

- With 850 - images are basically monochrome.

- With 720 - red is so much dominating that in-camera white balance cannot be set (on Nikon) or can be set (with Fuji GFX) but it is still reddish and cannot be tweaked anymore to have less red.

 

I'm wondering if I can still get colors from IR phots but also be able to white balance in such a way that there is a room to still add/subtract at least a bit of any channel: red, green, blue.

I *think* this could be possible with something like (? I don't know) maybe 740, 750, 760, 780, 800? Hard to say...

 

Does anybody have such a filter to confirm that this is possible?

 

I only seen 760 nm option from Zomei (Chinese) and I wonder if that would work? Anybody used that filter?

 

I want filter in 82mm diameter, so I can put this on almost all my lenses with step-up rings...

 

[colinbm]

As you can see different sensors have different spectral responses in full spectrum IR....this list is not comprehensive......
https://maxmax.com/spectral_response.htm

PS. you could use a NIR diffraction grating to measure your cameras spectral response in the NIR with a broad light source like a incandescent or halogen lamp. Some brain food here.....
https://www.mdpi.com/2313-433X/6/8/79

[Stefano]

I didn't know some cameras have problems with white balance using a 720 nm filter.

[Andy Perrin]

Sony cameras can white balance pretty much anything. Nikkon seems to have more issues. Of course the best solution is to not white balance in-camera. Save the RAW and do it in a good editor. 
 

colin his sensor is not the issue. It’s just his camera body’s ability to set a white balance. 

[dabateman]

You're having issues with white balance on a Fujifilm 50R in IR?

I have a 720nm lp converted Fujifilm 50R that I bought used, the conversion was done by LifepixeI. I don't have issues with it.

It came with an interesting custom white balance 1 setting with -9 red. 

But I got a nice traditional white balance,  just off grass. 

Can you share an image?

[lukaszgryglicki]

I don't have issues with 50R - it sets WB just fine, but while doing so, it hits maximum setting for red channel, so I cannot subtract more red, still the resulting image is too red a bit. I was wondering if that will be better with some filter between 720 and 850, while still distinguishing some false colors. 850 is already monochromatic.

I know that I can set all of this is post-process from RAF or JPG - I was just wondering if for example Zomei 760nm registers a bit less red while the response is still different in channels so I can get some (false) colors.

[Andrea B.]

Just a quick Schott/Hoya reference. Our (absent) friend Cadmium put this together a few years back.

 

695nm longpass: Schott RG695 @3mm

700nm longpass: Hoya R-70 @2.5mm

 

715nm longpass: Schott RG715 @3mm

720nm longpass: Hoya R-72 @2.5mm

 

760nm longpass: Hoya IR-76 @2.5mm

780nm longpass: Schott RG780 @3mm

800nm longpass: Hoya IR-80 @2.5mm

830nm longpass: Schott RG830 @3mm, and Hoya IR-83 @2.5mm

850nm longpass: Schott RG850 @3mm, and Hoya IR-85 @2.5mm

 

960nm longpass: Hoya RM-90 @2.5mm

970nm longpass: Schott RG1000 @2mm (...interesting !!)

1000nm longpass: Schott RG1000 @3mm, and Hoya RM-100 @2.5mm

 

It seems like we are "missing" a 740 nm and a few between 850 nm & 960 nm.

 

Optical filter manufacturers can probably supply any "missing" IR longpass filters -- for a hefty price. I suspect that to attain an IR cut-in between 720 and 760 nm, or between 850 and 960 nm, coatings will be involved. Must be something about the typical chemicals used in glass filters that prevents these missing values.

 

Do note however, the variation in cut-in values based on the thickness of the same glass.

I had kinda forgotten that thinner Schott RG 1000 could cut-in earlier than thicker Schott RG 1000.

Does this mean that Hoya IR-76 at say 1.5mm might cut in at 740 nm or 750 nm???

(I need to go review this using the Schott curve generator.)

 

---

 

 

The thing is ---> IR pass filters used on a Bayer filtered conversion don't seem to have much raw or white-balanced color variation.

Look here: LINK or LINK

The other thing is ---> the IR-reflectivity of natural subjects seems to be somewhat stable between 700-1000 nm. 

Those two factors don't necessarily provide for a lot of false color variation in a triChrome stack.

 

That most def is *not* to say that you shouldn't try!! Try a mixture of natural and manmade subjects. Try boosting the colors. Try channel swaps. Try other stuff! 

 

 

[lukaszgryglicki]

I don't have anything between 720 and 850, just ordered Zomei 760 82mm - I wonder how it will work - 720 gives colors - but always too much red, while 850 is basically monochrome, 760 can be interesting then... will see when it arrives, price is around $35 with free shipping so worth trying.

 

[ulf]

I have all the typical Zomei IR-Pass filters and they work well.

The image result depends very much on the way the images are white balanced, but the general trend is that the higher the cutoff, the less chromaticity you get.

Beside buying pre-defined filter thicknesses sometimes it is possible to get a custom thickness from some of the Chinese filter manufacturers.

By changing thickness the cutoff wavelength will be shifted.

Earlier I could order such filters from Tangsinuo. 

They often have IR-filters with many different cutoff wavelengths.

That might also be possible for there QB39 that would work better at 2mm thickness than their 1.5mm variant.

The extra fee for custom thickness was not that high.

[Andrea B.]

760 nm will still have a raw red color. The blue doesn't kick in until about 780 nm. But only a tiny amount of blue at 780 nm.

These are Schott. I have no idea how well Zomei will track these color changes.

[Note that the intensity of the raw color is very, very, very low. From 780 onward, the raws can almost look like B&W. To make the most of it in a triChrome, you would probably want to push the saturation?]

[lukaszgryglicki]

OK so if 760 still has red - then perfect, I only want that red not to overwhelm other channels, if blue kicks-in at 780 - then also perfect - I hope that in-camera WB will be able to work with this 760 allowing all 3 channels to be manually tweaked in both directions (with 720 - white balance is set, but red is already at its minimum so cannot be adjusted anymore and image is still too red then - histogram of red is way more than the other two - I hope to balance them with 760 and still have enough variance between channels to register some false colors - with 850 - I can balance channels, but they all change in the same way - so image is basically monochorme, for example violet/purple - one without channel variations - so when you equalize all is monochrome then).

[Andy Perrin]

3 hours ago, Andrea B. said:

[Note that the intensity of the raw color is very, very, very low. From 780 onward, the raws can almost look like B&W. To make the most of it in a triChrome, you would probably want to push the saturation?]

 

You would want the opposite for a trichrome, Andrea, you want each channel of the trichrome to itself be monochrome, just like the R,G, and B channels of an ordinary photo. That's what we were talking about in the other thread, the best way to make them monochrome.

 

But Lukas's problem is different, he wants his built-in white balance to not hit the ceiling.

[lukaszgryglicki]

Exactly, and it seems to be between 720 and 780 - if ever.