Five IR-Blockers Tested for Visible Color Use on Full Spectrum Camera

[Stefano]

A VERY crude test, but even here you can see how bumpy that red is in the iMac screen. Interestingly, my iPad screen (not shown) has the same pattern with the bumpy reds.

The bumpy red reminds me of fluorescent lamps or CRT screens. If those are LCD screens and are backlit with LEDs, that's quite odd.

[Andy Perrin]

The bumpy red reminds me of fluorescent lamps or CRT screens. If those are LCD screens and are backlit with LEDs, that's quite odd.

The iMac screen is LCD. It doesn't seem to be odd based on my quick survey of locally available screens with the little spectroscope. I'm not sure what tech the iPhone is using but most likely OLED.

[Andrea B.]

I looked at the photos on the iPhone and on iPad (Retina screen). The conclusions remain the same except for the lavender-blue petals in the Flower Mix color-profiled Set 2. I’m not able to see much diff between them in that set on these two screens.

 

************

The background of the Black Hollyhock photos #4 and #5 are strange and I don’t know why.

 

************

[Andy Perrin]

Yeah, so I would trust the iPhone screen results the most. I think probably the computer screen was responsible for the differences on the lavender-blue then.

 

The iPhone screen is also capable of darker blacks than other tech if it is OLED. That probably means the dynamic range is higher on the iPhone and you can see more in the shadows than on the other screen.

--

Anyway, this is a reminder that the colors (and conclusions) depend on the entire chain from the original light to the light emitted by our displays, not just the filters and profiling.

[Cadmium]

Wasn't that also the camera that was particularly IR-sensitive? If that's the case then it would make sense that S8612 is needed to tone down the IR more aggressively.

 

Andy, That may have something to do with it. I have no experience with that camera myself. It is used as a forensic camera by the police and such. It has a spectrum of approximately 380nm – 1,000nm,

so not a camera I would get for UV depth, based on that, but they may be measuring those numbers differently than some would.

It is the only camera I know of that does visual best with S8612, but there may be others. All of my full spectrum converted cameras are Nikon DSLRs.

[Andrea B.]

David, you are correct.

 

IR-Blocker #1: Baader UV/IR-Cut

IR-Blocker #2: Schott BG38 x 2.00 mm

IR-Blocker #3: Schott BG40 x 2.00 mm

IR-Blocker #4: Schott BG39 x 2.00 mm

IR-Blocker #5: Schott S8612 x 2.00 mm

#6: Unfiltered

 

The best results are from the custom profiled BG38 and BG40. I found it difficult to see much difference between them for my choice of camera/lens/light. The custom profiled Baader UV/IR-Cut is next best and really quite close to the BG38/40.

 

Custom profiled BG38, BG40 and the Baader UV/IR-Cut are very similar to one another.

 

Custom profiled BG39 and S8612 are very similar to one another

but as a set they are rather different from the other 3 as a set.

 

Andy writes: Anyway, this is a reminder that the colors (and conclusions) depend on the entire chain from the original light to the light emitted by our displays, not just the filters and profiling.

 

Yes. And don't forget that professionals who need to do some serious color matching also take into account the ambient light in which the photos are viewed whether on-screen or in print.

 

But I'm simply trying to get the best Visible photos from a full-spectrum conversion most of which will be displayed only as smushed JPGs on a typical laptop, so I will stop here.

 

I enjoyed this experiment and hope that it is useful to other members and readers !!! :bee:

 

.

[Andy Perrin]

So, with your camera at least, I guess I really do prefer BG38 2mm. The Baader made a decent showing in my opinion, but my eyes have trouble with reds, so it’s possible I can’t see the discrepancies.

[Andrea B.]

I think that if someone does not want to make custom color profiles (for whatever reason, I'm not judging),

then we should definitely recommend the BG38 for correcting Visible color from a full-spectrum camera

because its WB-Only photos were most like its custom color profiled photos.

 

((Actually, over the years here on UVP, the BG38 *has* been recommended many, many times.

This current experiment simply reinforces that.))

 

If making custom color profiles with color checker cards is not a problem,

then any of the BG38, BG40 or Baader UV/IR-Cut would be good.

Of course, the BGs are quite a bit less expensive than the Buvircut, so there is that.

 

There are other IR-Blockers besides the ones I tested here.

Other tests and experiments with other IR-Blockers are very welcome to be posted on UVP.

 

Thanks to everyone for their comments.

[dabateman]

I used fall maple leaves as a subject, as it has enough of the red shifts to see the difference.

As you saw, the color checker is too limited.

I found BW 486 dichoic filter was best with my Olympus E510 full spectrum and Sigma SD14.

So a head of my Olympus Em1mk1 full spectrum conversion, bought a couple of BW 486 filters for my favorite lenses. Boy was I disappointed.

It works best with my 4mm BG38 filter, roughly 2mm BG40. So bought one of them.

 

Its really camera dependent. I wouldn't have guessed the Fuji XT1 works best with S8612.

 

People may just need to test on their own with the range, and see what looks best.

[Andy Perrin]

The BG38 2mm has always worked well on my Sony cameras. I got results like Andreas for the various filters.

[Andrea B.]

Yes, of course, custom color profiles are, or might be, camera dependent. I've tried to stress that the custom color profiles are unique for camera + lens + filter + lighting. I make them always on a per-session basis when shooting indoors. For my strong sunlight, outdoor photos -- not early, not late -- I have not found much difference in the profiles made for each camera.

 

Slightly off topic: That B+W 486 is dichroic?? I have one around here somewhere but no time to look at the mo.

 

 

Comment: A 4.00 mm thick BG 38 is a lot of thickness. What does that buy you over a 2.00 mm thickness? Well, more IR cut I suppose. But still...

.

[dabateman]

Yes, of course, custom color profiles are, or might be, camera dependent. I've tried to stress that the custom color profiles are unique for camera + lens + filter + lighting. I make them always on a per-session basis when shooting indoors. For my strong sunlight, outdoor photos -- not early, not late -- I have not found much difference in the profiles made for each camera.

 

Slightly off topic: That B+W 486 is dichroic?? I have one around here somewhere but no time to look at the mo.

 

 

Comment: A 4.00 mm thick BG 38 is a lot of thickness. What does that buy you over a 2.00 mm thickness? Well, more IR cut I suppose. But still...

.

 

Yes the BW 486 is a dichoic UV (well 380nm) IR (well leaks 708nm) blocking filter.

 

The 4mm BG38 came in a Leica microscope filter bundle. So wasn't the sole reason I bought it. But helpful. Not sure why so thixk as BG40 is easier. But it works. Just wish the seller didn't break the 0.5mm UG1 filter. That would have been fun. I also have broken the 1mm UG5 filter since, trying to clean it. But still kind of works taped together.

[Andrea B.]

FILTERS USED in THIS TEST

• IR-Blocker #1: Baader UV/IR-Cut
  
• IR-Blocker #2: Schott BG38 x 2.00 mm
  
• IR-Blocker #3: Schott BG40 x 2.00 mm
  
• IR-Blocker #4: Schott BG39 x 2.00 mm
  
• IR-Blocker #5: Schott S8612 x 2.00 mm
  
• #6: Unfiltered
  

Camera: Nikon D610 Full-Spectrum Conversion

Lens: UV-Nikkor 105 mm f/4.5

 

---

TRANSMISSION CHARTS

 

Here is a transmission chart for 4 of the 5 IR-Blockers I tested for use with full-spectrum conversions in Visible light. Thank you to Cadmium for providing this chart which can also be found in this section:

Filter Transmission Charts

• BG38 x 2.0 mm
  
• BG39 x 2.0 mm
  
• BG40 x 2.0 mm
  
• S8612 x 2.0 mm
  

(The chart also has a line for S8612 x 1.0 mm, not tested here.)

 

---

 

CONCLUSION: Best Visible Photos from CMOS Full-Spectrum Conversion = BG38

Read fully. I do try to make the appropriate disclaimers.

• For this experiment with the D610-conversion, the BG38 and the BG40 provided the best Visible photos. The primary advantage of the BG38 over the BG40 is slightly higher transmission. I find it interesting that the BG38 and BG40 gave the best Visible photos in spite of the fact that both of them pass some Infrared light.
  

• The Baader UV/IR-Cut held up well enough to be considered the 2nd-place (3rd?) winner. The lightness/darkness of Visible reds can be an issue. And I think the Visible bluish-purple or purplish-blue colors might be a problem. But I cannot say for sure if that is entirely the fault of the Baader UV/IR-Cut filter because it is well-known that there can be camera/sensor problems (minor) for Visible color near 400 nm.
  

• As with any such experiment, the results are valid only for the gear used in the test. However, for the typically Bayer-filtered, CMOS sensor, I think this result will hold. For any camera with different color filtering or different sensor, this experiment is not valid. For example, anyone using a Sigma 3-layer camera or any Fuji with an X-sensor should test what works best for those types of sensors.
  

• Remember my conclusions are based on custom color profiling. Custom color profiles should be used for the best Visible color reproduction from converted cameras. I think the experiment made that clear. If you cannot or do not want to use a custom color profile (I'm not judging), then the BG38 performed best with a white balance only correction.
  

• I did NOT investigate how the passage of Ultraviolet light from the BG filters might affect a Visible photo. (This would be a great test for someone to make.) My suspicion is that in long-distance landscapes you might see the scattering effect of UV in your Visible photos. Of course, all you need to do is stack a Longpass over your BG to cut the UV haze in a long-distance photo.
  

[bkg=yellow][/bkg]

 

 

Please let me know if I got anything wrong there. Comments, questions and additions always welcomed.

[Andy Perrin]

My particular go-to combination is BG38 2mm + MidOpt BP550, which has a square profile similar to the Baader UV/IR cut from 405-690nm:

https://midopt.com/filters/bp550/

 

The blocking on the MidOpt is not super great, but is adequate to cut the weak UV in sunlight, and any remaining IR from the BG38. I would not recommend this filter on its own for the same reason as the Baader: too much red.

[Andrea B.]

I could probably emulate that combo using the Buvircut + BG38 stack?

 

So here is a followup Visible experiment (alluded to above). The purpose would be to test how much UV-blocking plays a role in getting a good Visible photo from a converted camera.

• Buvircut only
  
• BG38 only
  
• Buvircut + Bg38 stack: this combo cuts both UV and IR.
  
• BG38 + some longpass. Maybe a 400nm or a 405nm longpass?
  
• Make 3 photos for each combo:
  
  • long distance, like a distant mountain range. Some scene with UV scatter.
    
  • close-up, 1-2 feet. Yet again, a flower shot would work.
    
  • medium range, maybe like 30 feet? Generic back-yard scene.
    

I would love to try this but I really really need to try to update the Index posts here on UVP. They are getting years out of date.

 

 

.

[photoni]

Good morning.

I see a basic problem, all files are in sRGB,

you should use a larger color space like AdobeRGB or if you have a good MAC or professional monitor the DCI-P3.

 

Below I put a file to check the quality of your monitor.

it is with ProPhoto profile (no monitor even pro sees all the colors)

with this monitor I can see all the notches of sRGB well.

With the calibrated NEC I see all AdobeRGB ... almost all P3 (I miss 2 red bars but I see it much more saturated than Adobe, and similar to ProPhoto)

 

For two days I have been trying to calibrate my "standard" Z7 with the new 105 macro.

Andrea ... if I don't confused your post I can put some sample files of Z7 develop with Caprure One, Photoshop, e colorchecker

[Andy Perrin]

I think the iPhone screen is good enough photini. I can see all Adobe except some of the blacks on it.

[Andrea B.]

Toni, excellent point you have raised.

 

I first have to do a little test to see which way my converter works on profiles because I have forgotten.

 

 

The first photo has embedded sRGB.

The second has embedded ProPhoto.

They look the same to me.

 

 

The first photo has applied sRGB.

The second has applied ProPhoto.

OK, now they look different.

[photoni]

Andrea all your photos have an embedded sRGB profile ... see the Exif

in the last most vivid I see: Andre G Blum

<aux: ImageNumber> 2268 </ aux: ImageNumber>

with your Nikon you made 2268 shots :)

Don't use ProPhoto, no monitors see those colors!

.

from the raw must come out a file with a large color space (AdobeRGB or P3) then you can cut the color table in sRGB

[photoni]

this is my best color table

Z7 - Raw developed with C1

gray balance on the second notch white 245, black 10 on 255 levels

Jpeg Q10 - color profile AdobeRGB

[Andrea B.]

I'll post a WB-Only set with Adobe RGB and its corresponding Custom Color Profiled set with Adobe RGB.

 

Here are the WB-Only mixed flower strips with Adobe RGB.

 

 

 

Here are the Custom Color Profiles mixed flower strips with Adobe RGB.

 

---

It seems to remain true that the last 3 photos in each set do not properly capture the colors of the actual flowers. Those were the BG39, S8612 and unfiltered photos.

 

In the second custom profiled set it remains true that the first 3 photos are the best. The differences are minor. It is still true that #2/BG38 and #3/BG40 are the best. I think I see a bit more difference between the custom profiled #2/BG38 and #3/BG40 here than I did in the sRGB custom profiled set above.

 

The BG38 still "wins". That is, it seems to provide the most closely matching colors.

 

((Did I do this correctly. I suddenly feel like I don't know what I'm doing.

[Andrea B.]

Compare custom color profiles for filter #2 with sRGB and with AdobeRGB.

The differences are minor. I think the change in the pinks is most noticible.

To review, #2 == BG38 x 2.00 mm.

 

[Andy Perrin]

I can see some differences but I agree they are minor. BG38 2mm is still the winner, with this camera anyhow. (And my Sonys.)

[Andrea B.]

I would like to ask a question.

 

When I am changing the ICC from sRGB to AdobeRGB, is it better to go all the way back to the raw NEF

and work forward through the TIF export and final resized JPG while using AdobeRGB from the start in the raw file?

 

OR, can I simply apply AdobeRGB to the final JPGs as a last step regardless of what ICCs I used along the way?

 

Thanks for any input on this.

[Andy Perrin]

Not an answer but some interesting reading here:

https://www.colourmanagement.net/advice/about-icc-colour-profiles/