Jump to content
UltravioletPhotography

CIR with #12 yellow & 590nm filters


Recommended Posts

The Tiffen yellow is my favourite filter for Aerochrome processing, and the 590nm a good balance of color and IR sensitivity.

 

post-189-0-99740700-1542832274.jpg

 

post-189-0-70495600-1542832291.jpg

 

post-189-0-44197100-1542832310.jpg

 

post-189-0-91895300-1542832332.jpg

 

post-189-0-02021200-1542832347.jpg

 

post-189-0-46247100-1542832364.jpg

Link to comment

Gary, Great Aerochrome processing shots.

That inspires me. :-)

What processing method are you using?

 

Much appreciated Cadmium !

CWB (Sony a7R) is set on neutral gray (daylight), with a #12 Y + Maxmax Xnite CC2 filter stack. In Sony Imaging Edge, I move the Hue slider to 100, which shifts to blue/purple palette to purple/red. In CS6 move the Hue/Saturation 'Hue' slider slightly to the right, until the reds look right. Everything else is done by eye, with CS6, NIK (Pro Contrast + Viveza), and sometimes Topaz Clarity. Basic processing takes about 5 minutes, per image.

Link to comment

While I am a fan of the EIR type shots, I really like 4 and 6 - love the golden sky.

 

Thanks Jonathan ! Golden yellow is what I was going for. The traditional IR processing method is to do a channel mixer red/blue swap, which results in a blue sky and yellow trees. For 590nm IR, I prefer to process the SOOC brown/blue images, without the swaps.

Link to comment

Hmm, #12 + CC2. I need to try that stack sometime.

CC1 CC2 is Schott KG1 (sorry I edited this, I meant to say CC2 = KG1. CC1 = BG38).

Going by Wratten #12, 50% is approximately 520nm, converted to Schott OG530 would probably be closest, but OG515 or OG550 might work about the same.

So here is what the stack looks like.

post-87-0-16939700-1542964968.jpg

Link to comment

Hi Cadmium,

The CC1 won't work, because it allows IR, well beyond 900nm.

 

My use of #12Y + CC2, was sparked by the Aerochrome-like images in this article, using a Kopp7093 filter, with a Sony a7RII, and Matlab.

https://www.nickspiker.com/blog/search-digital-kodak-aerochrome/

 

In the article, a cutoff of 900nm was recommended, using the Kopp7093 filter from the Kopp glass company, which has a sharp cutoff at 900nm.

http://www.koppglass...ilter-glass.php

 

The cutoff on the XNite CC2 is pretty close to the Kopp 7093, so I thought I'd try that. I honestly didn't think that the CWB setting would be any different, by stacking the CC2 to the #12Y, but it does ! I found that setting a CWB (on neutral gray) with the #12Y alone, results in purple/pink hues, whereas #12Y+CC2 results in blue/purple hues (better for EIR processing). As a side note, the #12Y alone, produces colors similar to Lifepixels 'Hyper Color' filter. https://www.lifepixel.com/hyper-color

 

Since I don't have any Matlab experience (as used in the article), I've just experimented with Hue settings in Sony Imaging Edge and CS6, until I approximated the colors of EIR. Aside from red/magenta foliage, Aerochrome skin tones in particular have a slightly greenish/yellow tone, as shown in this Aerochrome 35mm slide from my collection.

post-189-0-35243700-1542996085.jpg

Link to comment

The CC1 won't work, because it allows IR, well beyond 900nm.

 

It is the other way around.

CC1 blocks a lot more IR than does CC2.

Here is a graph showing the transmission or BG38 (CC1) and KG1 (CC2). The graph shows both 2mm and 3mm thicknesses.

post-87-0-92027500-1543019707.jpg

Link to comment

So you are filtering out longer-wavelength IR because Aerochrome was not sensitive to those wavelengths?

 

Does your processing method involve any channel swapping?

Link to comment

Theoretically the idea is simple, however because IR is being recorded in all three channels, and in this post processing scheme there is no subtraction from green and red (<blue and <green),

the resulting blue and green are not just from pure green and red, but instead they are a mix, thus blue and green are blue+IR and green+IR, and red?

In other words, it isn't calculated or subtracted, it is mostly just hue shifted, right? I sup[pose the partially attenuated higher IR helps top compensate for the non-subtracted, maybe that is the idea there, but...

Am I thinking straight here?

Not that it doesn't look good, because it looks good.

Link to comment

Nick’s method does do subtraction of channels. It is his step 5:

Load the spectral data into Microsoft Excel using LibRaw and Matlab, then solve based on the difference between the matrixed data and the Kodak data. Basically we add and subtract color channels until we get a curve that matches as best as possible our target (Kodak Aerochrome)
Link to comment
The hue shift, is the first step in my workflow. Once I get the reds to look somewhat like Aerochrome reds, I start with NIK Color Efex Pro > Pro Contrast, by adjusting the Correct Color Cast and Dynamic Contrast sliders to about halfway. The 'Correct Color Cast' adjustment is a subtractive color edit, which reduces the overall reddish cast, highlights the greens, and creates a variety of reds and magentas. After that, it's Viveza, spot edits and sometimes Topaz.
Link to comment

Perhaps a strong enough hue shift is somewhat similar to a channel swap...

 

Out of curiosity, how do vehicle taillights or human lips appear when images are processed this way?

Link to comment

Hi Clark,

Red/blue swap produces pink foliage and a cyan sky, so you still have to play around with hue and saturation. I've also used Ben Lincoln's 'The Mirror Surface Breaks' multispectral software, with excellent results, but it's a much longer process.

https://www.beneaththewaves.net/Software/The_Mirrors_Surface_Breaks.html

 

Vehicle tail lights with bulbs are orange, whereas red LED lights turn green. Fair skin tones are greenish with yellow lips.

 

post-189-0-51295600-1543175700.jpg

Link to comment

You can really tell which vehicles are using LEDs this way. Red LEDs are one of the few things that come up green in an IRG image--it's a rare color otherwise. Some of the vehicles have mixed lighting, with the overbar being LED while the taillights are incandescent--the Sienna at right midground is one of those. The traffic lights are LED here--but most are, nowadays.

 

Ben Lincoln's software is not user friendly, unfortunately--I have never tackled its learning curve. The approach taken by the Pixelbender plugin or Wavelength Pro has seemed more intuitive to me in terms of my being able to understand the details of what is happening. You are obviously getting an analogous result, but it is harder for me to wrap my head around how you are getting there.

Link to comment

Clark,

The green LEDs were an interesting surprise to me.

 

Unfortunately, Pixel Bender doesn't work in CS6, so I've never tried it. When I had Ben Lincoln's software up and running, it produced some very nice results, but each image took about 20 minutes to setup and process. My current processing is simple by comparison. Below are the steps.

 

Raw export from Sony a7R - CWB capture with #12Y + XNite CC2 filter stack

post-189-0-02617100-1543272293.jpg

 

Sony EDIT raw processing - move 'Hue' slider to 100

post-189-0-26031900-1543272364.jpg

 

CS6 Hue/Saturation edit - adjust 'Hue' slider for Aerochrome reds

post-189-0-18467000-1543272497.jpg

 

NIK Color Efex Pro > Pro Contrast - Move the 'Correct Color Cast' and 'Dynamic Contrast' sliders to reduce reddish cast and add definition.

post-189-0-04021400-1543272551.jpg

 

The results. The rest is just Viveza and my personal editing.

post-189-0-63395800-1543272682.jpg

Link to comment

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...