63 replies to this topic

[Christoph]

Today I had to break new ground (for me) to get a bit closer to my goal - to reproduce Kodak Aerochrome film authentically. My experiments with Sigma, especially with regard to details like the color change from red to yellow, were disappointing recently. So I grabbed my Nikon and went out. The vegetation is anything but super at the moment. But I was pleasantly surprised.

Here is the equipment for the experiment:
Nikon D7100 Full spectrum
Hoya X1 Green
Tiffen 16 Orange (or yellow)
Hot Mirror Kolari
B+W 830nm filter
Polarizer (Neewer)

First I tried the two-photo method with one image with visible light only, and one image with IR only. For this I used the Hot Mirror and the B+W filter. This was OK for me. Here is the result. Yes, yellow taillights, red vegetation, but the red car in the background was caramel colored (maybe it's the perspective though). These images were just processed with the color channel method, no additional reworking...









Then with only orange filter and polarizer - very disappointing. Again, no reworking, just the channels:




Finally I tried what was my greatest hope: the combination of green filter and orange filter (plus polarizer). I am quite happy with it. The red car is bright yellow, the vegetation is already good (for the season). The following are also straight out of the channel mixing, no reworking:








Those were reworked minimally via raw converter (red and orange hues both more red/magenta (-40) and vintage preset filter):






Probably a good method, because the X1 is known for differentiating vegetation very well, and it does a cutoff in the infrared spectrum...

Edited by Christoph, 08 January 2021 - 11:56.

[colinbm]

Very interesting Chris, thanks for sharing.

[dabateman]

I am confused about what you did.
Did you take an image with Hoya X1.
Then take a second image with tiffen 16 filter with a polarizer.
Then what from each image did you mix?
As in is X1 the red channel, tiffen 16 the green channel?

Or did you stack Hoya x1 with Tiffen 16 and a polarizer. Take one image.
Then switch the blue and red channels?

Thanks.

[Christoph]

@Da Bateman: So the best solution ist filter stack Orange 16 and Green X1, custom white balance on grey asphalt... then open PS, further optimize white balance, then do the channel stuff: Duplicate red and green channels, copy blue channel, paste into red channel, then copy the duplicated red channel, paste into green channel, then copy the duplicated green channel and paste it into the blue channel... that's it... It's all one image, one exposure!

Edited by Christoph, 08 January 2021 - 13:11.

[Christoph]

Essentially this: Digital replacement (emulation) of Ektachrome / Aerochrome… | Flickr But without subtracting IR from the channels...

[microbat52]

B -> R
R -> G
G -> B

Did I get it right? I need to try this.

[nfoto]

The "original" colour dye scheme of Infrared Ektachrome was,

IR -> R
R -> G
G -> B

when the addition you needed to keep blue (and UV) at bay by suitable filters.

[Christoph]

Here's another one -- some processing was done in the PS raw scheme, but just some hues and contrast (and grain added)... The faces of the people are yellow as well... the water pretty black

[Christoph]

microbat52, on 08 January 2021 - 13:44, said:

B -> R
R -> G
G -> B

Did I get it right? I need to try this.

Yes, because the blue serves as the IR channel. With Hoya and Orange you get blue/purple trees and green sky sooc

Edited by Christoph, 08 January 2021 - 14:13.

[nfoto]

Caucasians would usually appear pale yellow in Infrared Ektachrome.

[Christoph]

This is a starting point - I'm sure you can still optimize this... maybe add stronger cutoff for example. I didn't test with Red filter plus Green - that propably won't work, but I don't know...

[Stefano]

The sky appears a bit dark, but in IR (at least half IR) it is to be expected. Nice images.

[Andy Perrin]

Christoph, on 08 January 2021 - 13:22, said:

Essentially this: Digital replacement (emulation) of Ektachrome / Aerochrome… | Flickr But without subtracting IR from the channels...

If you don't subtract IR from the channels then you get a mix. I have no idea if that was true also of the original film. Depending on what you want to do with the images, you may not care if there is some IR contamination.

I have previously written up a method for doing IRG (infrared-red-green) using a dual-bandpass visible/IR filter from MidOpt that, as far as I can determine, fully separates the bands and is equivalent to photographing the visual and IR separately. These IRG images are not the same as a film emulation, however, nor were they intended to be.

[Christoph]

Andy, I assume that the green filter takes care of most of the IR contamination. If we had a lot of contamination, the image would look like the variant with only the orange filter - right? In the end, you can only compare and look, and if you compare the 2-photo method with my 1-photo method, I think the differences are negligible, respectively can be compensated by minimal readjustments. Important for me are that the main features of the Aerochrome are reproduced with a very simple workflow. Possibly an additional MidOpt dual band filter (by the way, have you also tried the triple band filters?) would help... Feel free to try :-) I would also love to see an unprocessed image (just after the channel mixing) from your method. Is it posted somewhere?

Edited by Christoph, 08 January 2021 - 21:36.

[Andy Perrin]

Christoph, I think we're not communicating properly. My images are Infrared-Red-Green images that are single photos but give the same end result as the 2-photo method. They are NOT Aerochrome/EIR reproductions, not are they meant to be.

Your images seem to be intended to reproduce the Aerochrome/EIR film look, which is a different goal. ("Important for me are that the main features of the Aerochrome are reproduced with a very simple workflow.")
You should not use the MidOpt dual band filter, I don't think, because that is not your intention.

Quote

Andy, I assume that the green filter takes care of most of the IR contamination. If we had a lot of contamination, the image would look like the variant with only the orange filter - right?

That makes no sense. The green filter passes plenty of IR, which is desirable here because if there were no infrared then you wouldn't have an Aerochrome. The contamination of the R and G CHANNELS with infrared is what I was referring to -- everything infrared past 850nm goes equally in to the R, G, and B channels, hence the need to subtract the B from the others. If you ignore this step then you have contamination. But it may not matter to you if you like the way the colors are turning out.

Hoya doesn't show what the X1 does past 750nm, but I strongly suspect it's letting through plenty of the 850nm range. I have one, I could check.


ETA: Incoming cool paper alert! Someone actually measured all the common green filters in IR! And it's for a paper on Aerochrome simulation with Foveons!

The black line is the Hoya X1.

Edited by Andy Perrin, 08 January 2021 - 21:58.

[Christoph]

Thanks for clarifying, Andy!

[Christoph]

I think I've got all the necessary color shifts with my method, as they are documented in this paper: https://core.ac.uk/d...df/72733525.pdf
Besides plants showing quite a wide variety of yellow to orange and red and magenta, as well as blue sky I got:
IR-reflecting red -> Yellow
Non-IR-reflecting green -> Blue (not sure about it since I only have photographed one green street sign in the distance that seems to be blue)
People having yellow skin tones
I'm unsure about the use of a polarizer. Maybe it casts the sky more magenta, but I have to do more testing.
Maybe vegetation will be more magenta with Yellow 12 or even Yellow K2... Stronger orange filter may produce a more uniform red coloration, but I don't think that Red 25A will work in the same way as it works with Aerochrome film.
Another shot from yesterday - again: minimal hue shifts after the channel mixture with making blue a little less magenta and red a little more red...

Edited by Christoph, 09 January 2021 - 16:06.

[Christoph]

Andy Perrin, on 08 January 2021 - 21:48, said:

The black line is the Hoya X1.

Apparently both the "acid green" and the black line show the Hoya x1 in different sizes. The black line is the 49mm, the green the 77mm. I do have the 77mm...

[UlfW]

Normally a proper Aerochrome image renders red like car tail lights as green.
That I think is true for both real film and fully successful digital simulation.
There was a Photoshop action called Tiffen12 that gives that effect with suitably filtered images.

[Christoph]

Shitty inside setup, but I think it's accurate: Shift from dark green in the visible spectrum to blue with the Orange/Green combi and channel switching...

Attached Images

•