Easier Aerochrome: The Triple Bandpass TB550/660/850

[Christoph]

Well, here it is, the fabled filter from Midwest Optical that seems to be great for Aerochrome-style photos: The Triple Bandpass TB550/660/850. You can't buy it from any store or online store, unfortunately, but only directly from Midwest Optical. Or, and this is how I did it, from a dealer on Ebay. That also comes out a lot cheaper.

 

Originally, I inquired with Midwest Optical. They only seem to sell to companies, colleges, and so on. And the prices are very high. Filter size 77mm costs around 1100 Swiss francs (is about the same in US dollars right now). Even filter size 37.5mm - which I have now - still costs around 600 Swiss francs new. Unbelievable. I paid around 50 francs/dollar for a 37.5mm at the Ebay dealer (including delivery costs).

 

 

Actually, I wanted to test today, but the Sahara dust, which is currently blown over all of Europe, currently leads to heavy cloud formation. The sun glitters only in between from behind the gray ceiling. Nevertheless, I was outside and took some pictures. Here is one of them, which shows very well the Aerochrome color change. The car is actually red, as is the hydrant.

 

 

 

Preliminary assessment: The filter seems to work very well, possibly with better results than with the green-orange combination. It is greenish-yellow (thus reminiscent of my combi), but reflects strangely. What does it do exactly? It lets only green, red and near infrared (NIR) pass. By simply switching color channels (NIR is imaged in the blue channel, so you make blue -> red, red -> green, and green -> blue), you get a false-color IR image that strongly resembles Aerochrome. Because the process is so simple, I see great potential in it for video projects. The Orange-Green method is not very suitable for filming, because the processing is rather tedious and artifacts occur. I think and hope that this will be much easier with the triple bandpass filter. The blue vegetation (before the channel swap) is already there, the white balance also fits very well (with 2500 Kelvin).

 

Am curious how the results will look in sunshine.

[colinbm]

Praying for ol' sol to glow for you

[GaryR]

The red to yellow swap looks great. I like the reds I'm getting with my current MidOpt DB850, but I'd be interested in your results, when sunny days arrive.

 

For video, does your video app have a channel mixer? DaVinci Resolve 17, apparently has this feature, but I haven't tried it yet.

[Andy Perrin]

I have bought directly from MidOpt with no problems before. Just called them up and ordered.

 

NIR is imaged not JUST in the blue channel. It goes in ALL THREE CHANNELS EQUALLY. This means you should be removing some of the blue from the other two channels or it will partially desaturate them.

 

You can easily show this by stacking your filter with an R72 to block red and green, and checking that the R and G channels still have plenty of signal.

[GaryR]

MidOpt also sells filter stock on ebay, at a discounted price. I ordered a few different sizes of the DB850 at a very good price, from seller "discount-optics". Mine arrived brand new in MidOpt cases.

 

https://www.ebay.ca/usr/discount-optics?_trksid=p2047675.l2559

[Christoph]

I have bought directly from MidOpt with no problems before. Just called them up and ordered.

 

NIR is imaged not JUST in the blue channel. It goes in ALL THREE CHANNELS EQUALLY. This means you should be removing some of the blue from the other two channels or it will partially desaturate them.

 

You can easily show this by stacking your filter with an R72 to block red and green, and checking that the R and G channels still have plenty of signal.

 

The contamination issue is interesting, but it's kind of elusive to me. Maybe you can show me an example: Image with, image without contamination.

My primary goal is an authentic Aerochrome look, not a 100% scientifically flawless result. For me, the focus is on the following:

- All color changes present

- Tones in the vegetation are correct

- Sky is dark blue

[Andy Perrin]

The contamination issue is interesting, but it's kind of elusive to me. Maybe you can show me an example: Image with, image without contamination.

My primary goal is an authentic Aerochrome look, not a 100% scientifically flawless result. For me, the focus is on the following:

- All color changes present

- Tones in the vegetation are correct

- Sky is dark blue

I don't have a sky available to photograph at the moment, but I can at least show the effect on chlorophyl in the form of basil paste. Also shown are some highlighters for additional colors.

 

Setup is a converted Sony NEX-7 camera with a 55mm/2.8 Micro-Nikkor lens.

 

Here is the visible light photo, taken with BG-38 and white balanced off the paper in PhotoNinja.

 

Here is a photo with Tiffen#12 + Midopt DB850 dual band filter, with the following channel transformations performed in photoshop using Channel Mixer:

Rnew = B

Gnew = R

Bnew = G

followed by a white balance in PhotoNinja + contrast/brightness adjustment in Photoshop.

 

Here you can see the pale colors that you get when the blue has not been removed. Blue removal is trivial to accomplish in the Channel Mixer by simply setting the Blue slider to -100% in the Gnew and Bnew channels.

 

With a redo on the channel swapping, now with blue subtraction, you get:

Rnew = B

Gnew = R - B

Bnew = G - B

 

The colors are now far more saturated because the contaminating infrared has been removed from the other channels.

 

Apologies to all for the disgusting basil paste in IRG!

[colinbm]

Fantastic Andy, just wonderful

[Christoph]

Thanks Andy! The sun shone for a few hours, enough to take some photos. These are with standard setting in-camera, white balance in PS Raw converter, then swapped channels. Pretty pale colors, much paler than with my orange-green method (OG). I can see here now also the contamination in the single channels, different than with OG: You can see in the green as well as in the red channel that the vegetation is much brighter, so rather absorbs IR. What are you doing now, Andy? You go into the channel mixer and do what? I've attached a RAW for you so you can do that yourself.

 

 

 

 

_DSC0011.NEF

[Christoph]

That's how I find it aesthetically pleasing. The same routine as with OG method: In the raw converter with the blues about -30, then channel swap (you can see afterwards also in the red and green channels how the vegetation gets darker) and a little more contrast.

 

[colinbm]

Looks like a winner to me

[Andy Perrin]

What are you doing now, Andy? You go into the channel mixer and do what? I've attached a RAW for you so you can do that yourself.

 

Christof, I now think you were asking in bad faith, and didn’t really want to know the answer to your own question. I’m sorry I put in the work to answer it.

[nfoto]

The last one looks good, apart from the sky which seems to be washed out and almost devoid of (blue) colour.

[Christoph]

Christof, I now think you were asking in bad faith, and didn’t really want to know the answer to your own question. I’m sorry I put in the work to answer it.

 

Maybe there's a misunderstanding here, Andy.... What do you mean exactly? I am interested in what result you come to with your method. You don't have to do this. It's just a a request for help, so that I can understand your workflow. The raw file is also an offer to anyone who wants to experiment with it themselves. Sorry if my choice of words (I'm not English-native) pissed you off.

[Christoph]

The last one looks good, apart from the sky which seems to be washed out and almost devoid of (blue) colour.

 

Yes, the sky was still very cloudy, unfortunately. Here is one with more clear sky....

 

[Andy Perrin]

Christof, I described my workflow in great detail in the post which you seemed to ignore, followed by telling me to go play with the image myself. I don’t know what explanation you need. I was very precise about what operations I did.

[Christoph]

Christof, I described my workflow in great detail in the post which you seemed to ignore, followed by telling me to go play with the image myself. I don’t know what explanation you need. I was very precise about what operations I did.

 

Then I am sorry. Sorry that it came across that way. My intention was just to see how you handle this RAW file, respectively if you reach the same result as I did. Maybe we have a different workflow. When I subtract the blue, I get weird colors....

[Andy Perrin]

I’ll try it with your image then. I’m not sure I understand what result you want so I’m only going to try to separate the channels into infrared, red, and green. (unless you have a way to measure “film IR similarity” then we can’t really say what method is best)

[Christoph]

This is what I mean when I write about weird colors....

 

 

 

So after moving the channels around, the blue channel is the new red channel, the red channel is the new green channel and the green channel is the new blue channel. This should also have changed in the channel mixer, right?

 

So now you should subtract (in the channel mixer) the rest IR from (the new) green and blue channel, so subtract there 100 each in the blue range. But you have to compensate that too? I do it with green, so there 100 each additionally... Obviously wrong, right? So my question to you is, how do you do it?

[Andy Perrin]

No, the problem is that you are thinking of it as a two step process when it is actually a one step process. You have to start from the image BEFORE white balance though, which means exporting the RAW as TIFF using either PhotoNinja or some other software that can access pre-white balance colors. I don’t have time now but I will try to do it late tonight.

[Andy Perrin]

Ok, this is what I get. I decided to do 50% decontamination, with these settings, on the UN-WHITEBALANCED TIFF (important!):

 

I left half the infrared in there because you said you liked the paler red trees. You can also do -100% on the blue in the Gnew and Bnew channels to fully remove the IR.

 

--

The un-whitebalanced TIFF is obtained from PhotoNinja with these (lack of) settings as a preliminary step:

[Christoph]

Thank you for your work! The result looks good as well. It's interesting how the "unprocessed" original image looks like the raw image from the orange-green method. Unfortunately, I don't have the program - I'll have to see if it's worth buying. Do you think it could be done in PS? A test could be done. I think you've touched on something like that. With light source and Ir filter, then see to what extent the light can also be seen in the green and red channel. Anyway, thanks a lot! And sorry again for the misunderstandings. I really appreciate the exchange here!

[Andy Perrin]

Thank you for your work! The result looks good as well. It's interesting how the "unprocessed" original image looks like the raw image from the orange-green method. Unfortunately, I don't have the program - I'll have to see if it's worth buying. Do you think it could be done in PS? A test could be done. I think you've touched on something like that. With light source and Ir filter, then see to what extent the light can also be seen in the green and red channel. Anyway, thanks a lot! And sorry again for the misunderstandings. I really appreciate the exchange here!������

PhotoNinja is a good buy, but if you just want to test this method, just put an R72 stacked with your TB550/660/850 and set a white balance in-camera. Then it should come out "unwhitebalanced" in photoshop hopefully. (I haven't tested that method though.)

[Christoph]

hmm... in-camera custom on white card or grass?

[Andy Perrin]

On PTFE preferably, since you can't be sure a "white card" is white in IR. Or if you don't have PTFE yet, just salt will do, as long as it's not colored salt like that Himalayan stuff.